Keywords: Human-Robot Collaboration, Human-Robot Teaming
Abstract: Human and robot partners increasingly need to work together to perform tasks as a team. Robots designed for such collaboration must reason about how their task-completion strategies interplay with the behavior and skills of their human team members as they coordinate on achieving joint goals. Our goal in this work is to develop a computational framework for robot adaptation to human partners in human-robot team collaborations. We first present an algorithm for autonomously recognizing available task-completion strategies by observing human-human teams performing a collaborative task. By transforming team actions into low dimensional representations using hidden Markov models, we can identify strategies without prior knowledge. Robot policies are learned on each of the identified strategies to construct a Mixture-of-Experts model that adapts to the task strategies of unseen human partners. We evaluate our model on a collaborative cooking task using an Overcooked simulator. Results of an online user study with 125 participants demonstrate that our framework improves the task performance and collaborative fluency of human-agent teams, as compared to state of the art reinforcement learning methods.

Keywords: Physical Human-Robot Interaction, Safety in HRI, Human-Aware Motion Planning
Abstract: Ergonomics and human comfort are essential concerns in physical human-robot interaction applications. Defining an accurate and easy-to-use ergonomic assessment model stands as an important step in providing feedback for postural correction to improve operator health and comfort. Common practical methods in the area suffer from inaccurate ergonomics models in performing postural optimization. In order to retain assessment quality, while improving computational considerations, we propose a novel framework for postural assessment and optimization for ergonomically intelligent physical human-robot interaction. We introduce DULA and DEBA, differentiable and continuous ergonomics models learned to replicate the popular and scientifically validated RULA and REBA assessments with more than 99% accuracy. We show that DULA and DEBA provide assessment comparable to RULA and REBA while providing computational benefits when being used in postural optimization. We evaluate our framework through human and simulation experiments. We highlight DULA and DEBA's strength in a demonstration of postural optimization for a simulated pHRI task.

Keywords: Gesture, Posture and Facial Expressions, Social HRI, Simulation and Animation
Abstract: Robots and artificial agents that interact with humans should be able to do so without bias and inequity, but facial perception systems have notoriously been found to work more poorly for certain groups of people than others. In our work, we aim to build a system that can perceive humans in a more transparent and inclusive manner. Specifically, we focus on dynamic expressions on the human face, which are difficult to collect for a broad set of people due to privacy concerns and the fact that faces are inherently identifiable. Furthermore, datasets collected from the Internet are not necessarily representative of the general population. We address this problem by offering a Sim2Real approach in which we use a suite of 3D simulated human models that enables us to create an auditable synthetic dataset covering 1) underrepresented facial expressions, outside of the six basic emotions, such as confusion; 2) ethnic or gender minority groups; and 3) a wide range of viewing angles that a robot may encounter a human in the real world. By augmenting a small dynamic emotional expression dataset containing 123 samples with a synthetic dataset containing 4536 samples, we achieved an improvement in accuracy of 15% on our own dataset and 11% on an external benchmark dataset, compared to the performance of the same model architecture without synthetic training data. We also show that this additional step improves accuracy specifically for racial minorities when the architecture's feature extraction weights are trained from scratch.

Keywords: Human-Robot Collaboration
Abstract: To collaborate well with robots, we must be able to understand their decision making. Humans naturally infer other agents' beliefs and desires by reasoning about their observable behavior in a way that resembles inverse reinforcement learning (IRL). Thus, robots can convey their beliefs and desires by providing demonstrations that are informative for a human learner's IRL. An informative demonstration is one that differs strongly from the learner's expectations of what the robot will do given their current understanding of the robot's decision making. However, standard IRL does not model the learner's existing expectations, and thus cannot do this counterfactual reasoning. We propose to incorporate the learner's current understanding of the robot's decision making into our model of human IRL, so that a robot can select demonstrations that maximize the human's understanding. We also propose a novel measure for estimating the difficulty for a human to predict instances of a robot's behavior in unseen environments. A user study finds that our test difficulty measure correlates well with human performance and confidence. Interestingly, considering human beliefs and counterfactuals when selecting demonstrations decreases human performance on easy tests, but increases performance on difficult tests, providing insight on how to best utilize such models.

Keywords: AI-Based Methods, AI-Enabled Robotics, Human-Centered Robotics
Abstract: Advanced social cognitive skills enhance the effectiveness of human-robot interactions. Research shows that an important precursor to the development of these abilities in humans is Theory of Mind (ToM) -- the ability to attribute mental states to oneself and to others. In this work, we endow robots with ToM abilities and propose a ToM-based approach to proactive robotic assistance by appealing to epistemic planning techniques. Our evaluation shows that robots implementing our approach and demonstrating ToM are measurably more helpful and perceived by humans as more socially intelligent compared to robots with a deficit in ToM.

Keywords: Multi-Modal Perception for HRI, Vision-Based Navigation
Abstract: Goal-based navigation in public places is critical for independent mobility and for breaking the boundaries that exist for blind or visually impaired (BVI) people in our sight-centric society. Through this work, we present a proof-of-concept system that can autonomously find socially preferred seats and safely guide its user towards them in unknown indoor environments. The robotic system includes a camera, an IMU, vibrational motors, and a white cane, powered via a backpack-mounted laptop. The system combines techniques from computer vision, robotics, and motion planning with insights from psychology to perform 1) SLAM and object detection, 2) goal disambiguation and scoring, and 3) path planning and guidance. We introduce a novel 2-motor haptic feedback system on the cane’s grip for navigation assistance. Through a pilot user study, we show that the system is successful in autonomously classifying and providing haptic navigation guidance to socially preferred seats, while optimizing for users’ convenience, privacy, and intimacy in addition to increasing their confidence in independent navigation. The implications are encouraging as this technology, with careful design guided by the BVI community, can be adopted and further developed to be used with medical devices enabling the BVI population to better independently engage in socially dynamic situations like seat choice.

Keywords: Imitation Learning, Human-Aware Motion Planning, Learning from Experience
Abstract: In this paper, we present the Sample Efficient Social Navigation from Observation (SESNO) algorithm that efficiently learns socially-compliant navigation policies from observations of human trajectories. SESNO is an inverse reinforcement learning (IRL)-based algorithm that learns from human trajectory observations without knowledge of their actions. We improve the sample-efficiency over previous IRL-based methods by introducing a shared experience replay buffer that allows reuse of past trajectory experiences to estimate the policy and the reward. We evaluate SESNO using publicly available pedestrian motion data sets and compare its performance to related baseline methods in the literature. We show that SESNO yields performance superior to existing baselines while dramatically improving the sample complexity by using as few as a hundredth of the samples required by existing baselines.

Keywords: Human-Centered Robotics, Learning from Demonstration
Abstract: Humans use audio signals in the form of spoken language or verbal reactions effectively when teaching new skills or tasks to other humans. While demonstrations allow humans to teach robots in a natural way, learning from trajectories alone does not leverage other available modalities including audio from human teachers. To effectively utilize audio cues accompanying human demonstrations, first it is important to understand what kind of information is present and conveyed by such cues. This work characterizes audio from human teachers demonstrating multi-step manipulation tasks to a situated Sawyer robot along three dimensions: (1) duration of speech used, (2) expressiveness in speech or prosody, and (3) semantic content of speech. We analyze these features for four different independent variables and find that teachers convey similar semantic content via spoken words for different conditions of (1) demonstration types, (2) audio usage instructions, (3) subtasks, and (4) errors during demonstrations. However, differentiating properties of speech in terms of duration and expressiveness are present for the four independent variables, highlighting that human audio carries rich information, potentially beneficial for technological advancement of robot learning from demonstration methods.

Keywords: Gesture, Posture and Facial Expressions, Social HRI, Representation Learning
Abstract: Co-speech gestures are a principal component in conveying messages and enhancing interaction experiences between humans and critical ingredients in human-agent interaction, including virtual agents and robots. Existing machine learning approaches have yielded only marginal success in learning speech-to-motion at the frame level. Current methods generate repetitive gesture sequences that lack appropriateness with respect to the speech context. To tackle this challenge, we take inspiration from successes in natural language processing on context and long-term dependencies, and propose a new framework that views text-to-gesture as machine translation, where gestures are words in another (non-verbal) language. We propose a vector-quantized variational autoencoder structure as well as training techniques to learn a rigorous representation of gesture sequences. We then translate input text into a discrete sequence of associated gesture chunks in the learned gesture space. Ultimately, we use translated gesture tokens from the input text as an input to the autoencoder’s decoder to produce gesture sequences. Subjective and objective evaluations confirm the success of our approach in terms of appropriateness, human-likeness, and diversity. We also introduce new objective metrics using the quantized gesture representation.

Keywords: Robotics and Automation in Agriculture and Forestry, Object Detection, Segmentation and Categorization
Abstract: Machine vision is a critical subsystem for enabling robots to be able to perform a variety of tasks in orchard environments. However, orchards are highly visually complex environments, and computer vision algorithms operating in them must be able to contend with variable lighting conditions and background noise. Past work on enabling deep learning algorithms to operate in these environments has typically required large amounts of hand-labeled data to train a deep neural network or physically controlling the conditions under which the environment is perceived. In this paper, we train a neural network system in simulation only using simulated RGB data and optical flow. This resulting neural network is able to perform foreground segmentation of branches in a busy orchard environment without additional real-world training or using any special setup or equipment beyond a standard camera. Our results show that our system is highly accurate and, when compared to a network using manually labeled RGBD data, achieves significantly more consistent and robust performance across environments that differ from the training set.

Keywords: Robotics and Automation in Agriculture and Forestry, Environment Monitoring and Management, Agricultural Automation
Abstract: The global grape and wine industry has been considerably impacted by diseases such as downy mildew (DM). Agricultural robots have demonstrated great potential to accurately and rapidly map DM infection for precision applications. Although the robots can autonomously acquire high-resolution images in the vineyard, data processing is mostly performed offline because of network infrastructure and onboard computing power constraints, limiting the use of agricultural robots for field operations. To address this issue, we developed a semantic segmentation model based on the modified DeepLabv3 network for near real time DM segmentation in high resolution images. Compared with state-of-the-art real time semantic segmentation models, the developed one achieved the best efficiency-accuracy balance on the DM dataset using embedded computing devices that can be easily integrated with commercial robotic platforms. DM severity estimation pipeline based on the model also showed a comparable measurement accuracy and statistical power in differentiation of fungicide treatments as the one based on offline semantic segmentation models. This enables the use of robotic perception systems for field operations.

Keywords: Robotics and Automation in Agriculture and Forestry, Agricultural Automation, Computer Vision for Automation
Abstract: In view planning, the position and orientation of the cameras have been a major contributing factor to the quality of the resulting 3D model. In applications such as precision agriculture, a dense and accurate reconstruction must be obtained quickly while the data is still actionable. Instead of using an arbitrarily large number of images taken from every possible position and orientation in order to cover the desired area of study, a more optimal approach is required. We present an efficient and realistic pipeline, which aims to optimize the positioning of cameras and hence the quality of the 3D reconstruction of a field of row crops. This is achieved with four steps; an initial flight to obtain a sparse point cloud, the fitting of a simple mesh model, the planning of images via a discrete optimization process, and a second flight to obtain the final reconstruction. We demonstrate the effectiveness of our method by comparing it with baseline methods commonly used for agricultural data collection and processing.

Keywords: Robotics and Automation in Agriculture and Forestry, Agricultural Automation, Field Robots
Abstract: Cultivation and weeding are two of the primary tasks performed by farmers today. A recent challenge for weed- ing is the desire to reduce herbicide and pesticide treatments while maintaining crop quality and quantity. In this paper we introduce BonnBot-I a precise weed management platform which can also performs field monitoring. Driven by crop monitoring approaches which can accurately locate and classify plants (weed and crop) we further improve their performance by fusing the platform available GNSS and wheel odometry. This improves tracking accuracy of our crop monitoring approach from a normalized average error of 8.3% to 3.5%, evaluated on a new publicly available corn dataset. We also present a novel arrangement of weeding tools mounted on linear actuators evaluated in simulated environments. We replicate weed distributions from a real field, using the results from our monitoring approach, and show the validity of our work-space division techniques which require significantly less movement (a 50% reduction) to achieve similar results. Overall, BonnBot-I is a significant step forward in precise weed management with a novel method of selectively spraying and controlling weeds in an arable field. Keywords — Robotics and Automation in Agriculture and Forestry; Agricultural Automation; Field Robotics.

Keywords: Agricultural Automation, Software-Hardware Integration for Robot Systems, Robotics and Automation in Agriculture and Forestry
Abstract: Contemporary robots in precision agriculture focus primarily on automated harvesting or remote sensing to monitor crop health. Comparatively less work has been performed with respect to collecting physical leaf samples in the field and retaining them for further analysis. Typically, orchard growers manually collect sample leaves and utilize them for stem water potential measurements to analyze tree health and determine irrigation routines. While this technique benefits orchard management, the process of collecting, assessing, and interpreting measurements requires significant human labor and often leads to infrequent sampling. Automated sampling can provide highly accurate and timely information to growers. The first step in such automated in-situ leaf analysis is identifying and cutting a leaf from a tree. This retrieval process requires new methods for actuation and perception. We present a technique for detecting and localizing candidate leaves using point cloud data from a depth camera. This technique is tested on both indoor and outdoor point clouds from avocado trees. We then use a custom-built leaf-cutting end-effector on a 6-DOF robotic arm to test the proposed detection and localization technique by cutting leaves from an avocado tree. Experimental testing with a real avocado tree demonstrates our proposed approach can enable our mobile manipulator and custom end-effector system to successfully detect, localize, and cut leaves.

Keywords: Robotics and Automation in Agriculture and Forestry, Agricultural Automation, Software-Hardware Integration for Robot Systems
Abstract: Due to labor shortage and rising labor cost for the apple industry, there is an urgent need for the development of robotic systems to efficiently and autonomously harvest apples. In this paper, we present a system overview and algorithm design of our recently developed robotic apple harvester prototype. Our robotic system is enabled by the close integration of several core modules, including visual perception, planning, and control. This paper covers the main methods and advancements in deep learning-based multi-view fruit detection and localization, unified picking and dropping planning, and dexterous manipulation control. Indoor and field experiments were conducted to evaluate the performance of the developed system, which achieved an average picking rate of 3.6 seconds per apple. This is a significant improvement over other reported apple harvesting robots with a picking rate in the range of 7-10 seconds per apple. The current prototype shows promising performance towards further development of efficient and automated apple harvesting technology. Finally, limitations of the current system and future work are discussed.

Keywords: Robotics and Automation in Agriculture and Forestry, Grasping, Data Sets for Robot Learning
Abstract: Apple picking is a challenging manipulation task, but it is difficult to test solutions due to the limited window of time that apples are in season. Previous methods have built simulations of apple trees, but simulations rarely capture soft contact and deformation well, both of which are common in fruit picking. In this paper we present and validate a physical proxy that replicates the mechanics of a real world apple pick. This proxy, in conjunction with a novel hand with multiple sensors, enables large-scale capture of sensor data for data collection and testing. To validate our approach, we train a Long Short-Term Memory network to classify a pick as successful or failed based on sensor feedback from the robot hand. We show that a network trained on the proxy performs as well or even better than a network trained solely on real apple trees, with accuracies up to 90 %. We determine which sensors are most important for pick classification and also demonstrate that our proxy preserves the most important sensor feature data for pick classification. Specifically for the implemented hand, the most informative feature group was the finger's servomotor effort.

Keywords: Robotics and Automation in Agriculture and Forestry, Deep Learning for Visual Perception, RGB-D Perception

Keywords: Agricultural Automation, Computer Vision for Automation, Object Detection, Segmentation and Categorization
Abstract: The evolution of smaller and more powerful GPUs over the last 2 decades has vastly increased the opportunity to apply robust deep learning-based machine vision approaches to real-time use cases in practical environments. One ex- citing application domain for such technologies is precision agriculture, where the ability to integrate on-board machine vision with data-driven actuation means that farmers can make decisions about crop care and harvesting at the level of the individual plant rather than the whole field. This makes sense both economically and environmentally. This paper assesses the feasibility of precision spraying weeds via a comprehensive evaluation of weed detection accuracy and speed using two separate datasets, two types of GPU, and several state-of-the- art object detection algorithms. A simplified model of precision spraying is used to determine whether the weed detection accuracy achieved could result in a sufficiently high weed hit rate combined with a significant reduction in herbicide usage. The paper introduces two metrics to capture these aspects of the real-world deployment of precision weeding and demonstrates their utility through experimental results.

Keywords: Semantic Scene Understanding, Vision-Based Navigation, Audio-Visual SLAM
Abstract: Reflective and textureless surfaces such as windows, mirrors, and walls can be a challenge for scene reconstruction, due to depth discontinuities and holes. We propose an audio-visual method that uses the reflections of sound to aid in depth estimation and material classification for 3D scene reconstruction in robot navigation and AR/VR applications. The mobile phone prototype emits pulsed audio, while recording video for audio-visual classification for 3D scene reconstruction. Reflected sound and images from the video are input into our audio (EchoCNN-A) and audio-visual (EchoCNN-AV) convolutional neural networks for surface and sound source detection, depth estimation, and material classification. The inferences from these classifications enhance 3D scene reconstructions containing open spaces and reflective surfaces by depth filtering, inpainting, and placement of unmixed sound sources in the scene. Our prototype, demos, and experimental results from real-world with challenging surfaces and sound, also validated with virtual scenes, indicate high success rates on classification of material, depth estimation, and closed/open surfaces, leading to considerable improvement in 3D scene reconstruction for robot navigation.

Keywords: Robot Audition
Abstract: The field of bioacoustics is concerned with monitoring wild animals based on their vocalisations. Passive acoustic recorders are now commonly used to collect data of the soundscapes of our wild places. While the data they collect is extremely useful, the majority of the recorders use a single omnidirectional microphone, and thus cannot independently perform localisation of a calling animal. Localisation can be useful to differentiate between multiple calling animals, to improve statistical estimates of abundance, and to locate calling posts, which may be close to nests. In this paper, we consider the design of a low-cost, practical, passive directional acoustic recorder that will facilitate animal localisation, and present and evaluate a prototype system for this purpose.

Keywords: Robot Audition
Abstract: Sound source separation is a method to extract a target sound source from a mixture of various sound sources and noises. One of the typical sound source separation methods is beamforming, which can separate sound sources by direction based on the phase difference between channels from the recorded signal of a microphone array, a multi-channel recording system. However, beamforming is a direction-based method and cannot separate multiple sources in the same direction. In this paper, we propose a method for separating sources in the same direction using multiple microphone arrays. The proposed method performs beamforming using multiple microphone arrays and extracts only the target sound source from the separated sound by the Non-negative Matrix Factorization (NMF), thus reducing the influence of other sources in the same direction. In this paper, to investigate the effectiveness of the proposed method, experiments were conducted assuming the presence of another sound source in the same direction from an arbitrary microphone array. The results show that the proposed method outperforms the delay-sum method in a simulation environment. In addition, experiments were conducted in a real environment to verify the effect of reverberation.

Keywords: Reinforcement Learning, Robot Audition
Abstract: Humans integrate multiple sensory modalities e.g., visual and audio) to build a causal understanding of the physical world. In this work, we propose a novel type of intrinsic motivation for Reinforcement Learning (RL) that encourages the agent to understand the causal effect of its actions through auditory event prediction. First, we allow the agent to collect a small amount of acoustic data and use K-means to discover underlying auditory event clusters. We then train a neural network to predict the auditory events and use the prediction errors as intrinsic rewards to guide RL exploration. We first conduct proof-of-concept experiments using a set of Atari games for an in-depth analysis of our module. We then apply our model to embodied audio-visual exploration using the Habitat simulator and active exploration with a rolling robot using the ThreeDWorld (TDW) simulator. Experimental results demonstrate the advantages of using audio signals over vision-based models as intrinsic rewards to guide RL explorations.

Keywords: Robot Audition, Deep Learning Methods, Human Performance Augmentation
Abstract: This paper describes the practical response- and performance-aware development of online speech enhancement for an augmented reality (AR) headset that helps a user understand conversations made in real noisy echoic environments (e.g., cocktail party). One may use a state-of-the-art blind source separation method called fast multichannel nonnegative matrix factorization (FastMNMF) that works well in various environments thanks to its unsupervised nature. Its heavy computational cost, however, prevents its application to real-time processing. In contrast, a supervised beamforming method that uses a deep neural network (DNN) for estimating spatial information of speech and noise readily fits real-time processing, but suffers from drastic performance degradation in mismatched conditions. Given such complementary characteristics, we propose a dual-layer robust online speech enhancement method based on DNN-based beamforming with FastMNMF-guided adaptation. FastMNMF (back end) is performed in a mini-batch style and the noisy and enhanced speech pairs are used together with the original parallel training data for updating the direction-aware DNN (front end) with backpropagation at a computationally-allowable interval. This method is used with a blind dereverberation method called weighted prediction error (WPE) for transcribing the noisy reverberant speech of a speaker, which can be detected from video or selected by a user's hand gesture or eye gaze, in a streaming manner and spatially showing the transcriptions with an AR technique. Our experiment showed that the word error rate was improved by more than 10 points with the run-time adaptation using only twelve minutes observation.

Keywords: Robot Audition, Range Sensing, Object Detection, Segmentation and Categorization
Abstract: This paper proposes a microphone array with a speaker to recognize the shape of the surface of the target object by using the standing wave between the transmitted and the reflected acoustic signals. Because the profile of the distance spectrum encodes both the distance to the target and the distance to the edges of the target’s surface, this paper proposes to fuse distance spectra using a microphone array to estimate the three-dimensional structure of the target surface. The proposed approach was verified through numerical simulations and outdoor field experiments. Results showed the effectiveness of the method as it could extract the shape of the board located 2m in front of the microphone array by using a chirp tone with 20kHz bandwidth.

Keywords: Perception for Grasping and Manipulation, Recognition, Robot Audition
Abstract: We investigated the use of impact sounds generated during exploratory behaviors in a robotic manipulation setup as cues for predicting object surface material and for recognizing individual objects. We collected and make available the YCB-impact sounds dataset which includes over 3,500 impact sounds for the YCB set of everyday objects lying on a table. Impact sounds were generated in three modes: (i) human holding a gripper and hitting, scratching, or dropping the object; (ii) gripper attached to a teleoperated robot hitting the object from the top; (iii) autonomously operated robot hitting the objects from the side with two different speeds. A convolutional neural network (ResNet34) is trained from scratch to recognize the object material (steel, aluminium, hard plastic, soft plastic, other plastic, ceramic, wood, paper/cardboard, foam, glass, rubber) from a single impact sound. On the manually collected dataset with more variability in the action, nearly 60% accuracy for the test set (unseen objects) was achieved. On a robot setup and a stereotypical poking action from top, accuracy of 85% was achieved. This performance drops to 79% if multiple exploratory actions are combined. Individual objects from the set of 75 objects can be recognized with a 79% accuracy. This work demonstrates promising results regarding the possibility of using sound for recognition in tasks like single-stream recycling where objects have to be sorted based on their material composition.

Keywords: Gesture, Posture and Facial Expressions, Social HRI, Emotional Robotics
Abstract: As a type of body language, gestures can largely affect the impressions of human-like robots perceived by users. Recent data-driven approaches to the generation of co-speech gestures have successfully promoted the naturalness of produced gestures. These approaches also possess greater generalizability to work under various contexts than rule-based methods. However, most have no direct control over the human impressions of robots. The main obstacle is that creating a dataset that covers various impression labels is not trivial. In this study, based on previous findings in cognitive science on robot impressions, we present a heuristic method to control them without manual labeling, and demonstrate its effectiveness on a virtual agent and partially on a humanoid robot through subjective experiments with 50 participants.

Keywords: Localization, Search and Rescue Robots, Aerial Systems: Applications
Abstract: For robot and drone auditions, microphone arrays have been used for estimating sound source directions and sound source locations. By using sound source localization techniques, for example, drones can detect people calling for help even if the target person is not visible. Most sound source localization methods are based on estimated sound source directions and triangulation. However, when it comes to situations using drones, severe drone noise distorts direction estimation results which could worsen the localization results badly due to the discreteness of direction estimation. In this perspective, the authors have proposed a sound source localization method that can omit outlying triangulation points, which could improve its localization performance. In this paper, an outdoor experiment has been held, and the proposed method is evaluated whether it can localize a sound source even if real drone noise is added to the recordings. Experiment results show that the proposed method can localize with 4.15 m of estimation error for a sound source up to 50 m away, suppress the impact of outliers, and use only plausible triangulation points.

Keywords: Machine Learning for Robot Control, Reinforcement Learning, Humanoid Robot Systems
Abstract: State-of-the-art reinforcement learning is now able to learn versatile locomotion, balancing and push-recovery capabilities for bipedal robots in simulation. Yet, the reality gap has mostly been overlooked and the simulated results hardly transfer to real hardware. Either it is unsuccessful in practice because the physics is over-simplified and hardware limitations are ignored, or regularity is not guaranteed, and unexpected hazardous motions can occur. This paper presents a reinforcement learning framework capable of learning robust standing push recovery for bipedal robots that smoothly transfer to reality, providing only instantaneous proprioceptive observations. By combining original termination conditions and policy smoothness conditioning, we achieve stable learning, sim-to-real transfer and safety using a policy without memory nor explicit history. Reward engineering is then used to give insights into how to keep balance. We demonstrate its performance in reality on the lower-limb medical exoskeleton Atalante.

Keywords: Machine Learning for Robot Control
Abstract: cooperative adaptive cruise control (CACC) is a smart transportation solution that can mitigate traffic jams and improve road safety. CACC performance is heavily impacted by communication time delay; moreover, control theory solutions generally compromise control performance by tuning control gains in order to maintain plant stability. We propose a control-machine learning hybrid approach called deep time delay filter (DTDF). DTDF predicts the present (un-delayed) car states given time delayed versions. We successfully train a neural network for the DTDF method and use a physical testbed to show that DTDF can mitigate the effects of constant time delays as large as 5s while maintaining superior control performance compared to that of a baseline control algorithm.

Keywords: Machine Learning for Robot Control, Behavior-Based Systems, Swarm Robotics
Abstract: Automatic control design for robotic systems is becoming more and more popular. However, this usually involves a significant computational cost, due to the expensive and noisy evaluation of candidate solutions through high-fidelity simulation or even real hardware. This work aims at reducing the computational cost of automatic design of behavioral arbitrators through the introduction of a two-step approach. In the first step, the structure of the finite state machine governing the behavioral arbitrator is optimized. To this purpose, a more abstracted model of the robotic system is leveraged in order to significantly reduce the computational cost. In the second step, the close-to-hardware, behavioral parameters are fine-tuned using a high-fidelity model. We show that, for a scenario involving a single robot and multiple tasks to be solved sequentially, using the proposed method results in a significant decrease of the computational cost while reaching the same controller performance both in simulation and reality.

Keywords: Machine Learning for Robot Control, Service Robotics, AI-Based Methods
Abstract: Learning from Demonstration (LfD) is a very appealing approach to empower robots with autonomy. Given some demonstrations provided by a human teacher, the robot can learn a policy to solve the task without explicit programming. A promising use case is to endow smart robotic wheelchairs with active assistance to navigation. By using LfD, it is possible to learn to infer short-term destinations anywhere, without the need of building a map of the environment beforehand. Nevertheless, it is difficult to generalize robot behaviors to environments other than those used for training. We believe that one possible solution is learning from crowds, involving a broad number of teachers (the end users themselves) who perform demonstrations in diverse and real environments. To this end, in this work we consider Federated Learning from Demonstration (FLfD), a distributed approach based on a Federated Learning architecture. Our proposal allows the training of a global deep neural network using sensitive local data (images and laser readings) with privacy guarantees. In our experiments we pose a scenario involving different clients working in heterogeneous domains. We show that the federated model is able to generalize and deal with non Independent and Identically Distributed (non-IID) data.

Keywords: Machine Learning for Robot Control, AI-Enabled Robotics, Manipulation Planning
Abstract: Pouring liquids accurately into containers is one of the most challenging tasks for robots as they are unaware of the complex fluid dynamics and the behavior of liquids when pouring. Therefore, it is not possible to formulate a generic pouring policy for real-time applications. In this paper, we propose PourNet, as a generalized solution to pouring different liquids into containers. PourNet is a hybrid planner that uses deep reinforcement learning, for end-effector planning, and Nonlinear Model Predictive Control, for joint planning. In this work, we introduce a novel simulation environment using Unity3D and NVIDIA-Flex to train our agents. By effective choice of the state space, action space, and the reward functions, we allow for a direct sim-to-real transfer of the learned skills without additional training. In the simulation, PourNet outperforms state-of-the-art by an average of 4.9g deviation for water-like, and 9.2g deviation for honey-like liquids. In the real-world scenario using Kinova Movo Platform, PourNet achieves an average pouring deviation of 2.3g for dish soap when using a novel pouring container. The average pouring deviation measured for water was 5.5g.

Keywords: Machine Learning for Robot Control, Industrial Robots, Reinforcement Learning
Abstract: Increasingly volatile markets challenge companies and demand flexible production systems that can be quickly adapted to new conditions. Machine Learning has proven to show significant potential in supporting the human operator during the time-consuming and complex task of robot programming by identifying relevant parameters of the underlying robot control program. We present a solution to learn these parameters for contact-rich, force-controlled assembly tasks from a simulation using hardware-independent robot skills. We show that successful learning and real-world execution are possible even under process deviation and tolerances utilizing the designed learning system. We present learning skill parameters as high-level robot control, evaluation and comparison of extensive simulations, and preliminary experiments on a physical robot test-bed. The developed solution approach is evaluated and discussed using the Peg-in-Hole process, a typical benchmark process in force-controlled assembly.

Keywords: Machine Learning for Robot Control, Reinforcement Learning, Humanoid Robot Systems
Abstract: Control of wheeled humanoid locomotion is a challenging problem due to the nonlinear dynamics and underactuated characteristics of these robots. Traditionally, feedback controllers have been utilized for stabilization and locomotion. However, these methods are often limited by the fidelity of the underlying model used, choice of controller, and environmental variables considered (surface type, ground inclination, etc). Recent advances in reinforcement learning (RL) offer promising methods to tackle some of these conventional feedback controller issues but require large amounts of interaction data to learn. Here, we propose a hybrid learning and modelbased controller Hybrid LMC that combines the strengths of a classical linear quadratic regulator (LQR) and ensemble deep reinforcement learning. Ensemble deep reinforcement learning is composed of multiple Soft Actor-Critic (SAC) and is utilized in reducing the variance of RL networks. By using a feedback controller in tandem the network exhibits stable performance in the early stages of training. As a preliminary step, we explore the viability of Hybrid LMC in controlling wheeled locomotion of a humanoid robot over a set of different physical parameters in MuJoCo simulator. Our results show that Hybrid LMC achieves better performance compared to other existing techniques and has increased sample efficiency.

Keywords: Machine Learning for Robot Control, Reinforcement Learning, Probabilistic Inference
Abstract: Robotic manipulation stands as a largely unsolved problem despite significant advances in robotics and machine learning in recent years. One of the key challenges in manipulation is the exploration of the dynamics of the environment when there is continuous contact between the objects being manipulated. This paper proposes a model-based active exploration approach that enables efficient learning in sparse-reward robotic manipulation tasks. The proposed method estimates an information gain objective using an ensemble of probabilistic models and deploys model predictive control (MPC) to plan actions online that maximize the expected reward while also performing directed exploration. We evaluate our proposed algorithm in simulation and on a real robot, trained from scratch with our method, on a challenging ball pushing task on tilted tables, where the target ball position is not known to the agent a-priori. Our real-world robot experiment serves as a fundamental application of active exploration in model-based reinforcement learning of complex robotic manipulation tasks. Project page https://sites.google.com/view/aerm.

Keywords: Control Architectures and Programming, Reinforcement Learning, Transfer Learning
Abstract: Deep reinforcement learning (DRL) is a promising approach to solve complex control tasks by learning policies through interactions with the environment. However, the training of DRL policies requires large amounts of training experiences, making it impractical to learn the policy directly on physical systems. Sim-to-real approaches leverage simulations to pretrain DRL policies and then deploy them in the real world. Unfortunately, the direct real-world deployment of pretrained policies usually suffers from performance deterioration due to the different dynamics, known as the reality gap. Recent sim-to-real methods, such as domain randomization and domain adaptation, focus on improving the robustness of the pretrained agents. Nevertheless, the simulation-trained policies often need to be tuned with real-world data to reach optimal performance, which is challenging due to the high cost of real-world samples.

This work proposes a distributed cloud-edge architecture to train DRL agents in the real world in real-time. In the architecture, the inference and training are assigned to the edge and cloud, separating the real-time control loop from the computationally expensive training loop. To overcome the reality gap, our architecture exploits sim-to-real transfer strategies to continue the training of simulation-pretrained agents on a physical system. We demonstrate its applicability on a physical inverted-pendulum control system, analyzing critical parameters. The real-world experiments show that our architecture can adapt the pretrained DRL agents to unseen dynamics consistently and efficiently.

Keywords: Modeling, Control, and Learning for Soft Robots, Hydraulic/Pneumatic Actuators, Soft Sensors and Actuators
Abstract: Compared to their rigid counterparts, soft material robotic systems offer great advantages when it comes to flexibility and adaptability. Despite their advantages, modeling of soft systems is still a challenging task, due to the continuous and often highly nonlinear nature of deformation these systems exhibit. Tasks like motion planning or design optimization of soft robots require computationally cheap models of the system’s behavior. In this paper we address this need by deriving operational point dependent Cosserat rod models from detailed volume finite element models (FEM). While the latter offer detailed simulations, they generally come with high computational burden that hinders them from being used in time critical model-based methods like motion planning or control. Basic Cosserat rod models promise to provide computationally efficient mechanical models of soft continuum robots. By using a detailed FE model in an offline stage to identify operational point dependent Cosserat rod models, we bring together the accuracy of volumetric FEM with the efficiency of Cosserat rod models. We apply the approach to a fiber reinforced soft pneumatic bending actuator module (SPA module) and evaluate the model’s predictive capabilities for a single module as well as a two-module robot.

Keywords: Modeling, Control, and Learning for Soft Robots, Kinematics, Sensor Fusion
Abstract: Proprioception, or the perception of the configuration of one's body, is challenging to achieve with soft robots due to their infinite degrees of freedom and incompatibility with most off-the-shelf sensors. This work explores the use of inertial measurement units (IMUs), sensors that output orientation with respect to the direction of gravity, to achieve soft robot proprioception. A simple method for estimating the shape of a soft continuum robot arm from IMUs mounted along the arm is presented. The approach approximates a soft arm as a serial chain of rigid links, where the orientation of each link is given by the output of an IMU or by spherical linear interpolation of the output of adjacent IMUs. In experiments conducted on a 660mm long real-world soft arm, this approach provided estimates of its end effector position with a median error of less than 10% of the arm's length. This demonstrates the potential of IMUs to serve as inexpensive off-the-shelf sensors for soft robot proprioception.

Keywords: Modeling, Control, and Learning for Soft Robots, Motion Control
Abstract: Soft robots have intrinsic advantages in interaction with humans or complex environments for actual applications, during which various external disturbances (e.g., external contact or collision) are inevitable. They show remarkable abilities in complicated tasks due to their easily deformable bodies and compliance characteristic, while also bringing challenges to the modeling, control, and trajectory planning in precise tasks. Thereby, perception and reaction to external disturbances are quite critical. In this paper, we focus on the slowly varying external contact and propose a contact detection method for the fiber-reinforced soft bending actuator (FRSBA), which is based on the system dynamical behavior. When contact is detected, a parameter extension method is introduced to modify the dynamic model. Then, a backstepping-based integrated direct/indirect adaptive robust controller with contact detection and accommodation strategy (CDA-DIARC) is designed to deal with system nonlinearities, uncertainties, and parametric variations caused by the external contact. Theoretical proof and physical experiments validate the convergence and high trajectory tracking performance of the proposed methods under different contact environments.

Keywords: Modeling, Control, and Learning for Soft Robots, Optimization and Optimal Control, Robust/Adaptive Control
Abstract: Fluidically actuated soft robots have promising capabilities such as inherent compliance and user safety. The control of soft robots needs to properly handle nonlinear actuation dynamics, motion constraints, workspace limitations, and variable shape stiffness, so having a unique algorithm for all these issues would be extremely beneficial. In this work, we adapt Model Predictive Control (MPC), popular for rigid robots, to a soft robotic arm called SoPrA. We address the challenges that current control methods are facing, by proposing a framework that handles these in a modular manner. While previous work focused on Joint-Space formulations, we show through simulation and experimental results that Task-Space MPC can be successfully implemented for dynamic soft robotic control. We provide a way to couple the Piece-wise Constant Curvature and Augmented Rigid Body Model assumptions with internal and external constraints and actuation dynamics, delivering an algorithm that unites these aspects and optimizes over them. We believe that a MPC implementation based on our approach could be the way to address most of model-based soft robotics control issues within a unified and modular framework, while allowing to include improvements that usually belong to other control domains such as machine learning techniques.

Keywords: Modeling, Control, and Learning for Soft Robots, Soft Robot Applications
Abstract: As robots begin to move from structured industrial environments to the real world, they must be equipped to not only safely interact with the environment, but also reason about how to leverage contact to perform tasks. In this work, we develop a modeling and motion planning framework for continuum robots that accounts for contact anywhere along the robot. We first present an analytical model for continuum manipulators under contact and discuss the ideal choice of generalized coordinates given properties of the manipulator and task specifications. We then demonstrate the utility of our model by developing a motion planning framework that can solve a diverse set of tasks. We apply our framework to end effector path planning for a soft arm in an obstacle-rich environment, and grasp planning for soft robotic grippers, where contact can happen anywhere on the arm or gripper. Finally, we verify the utility of our model and planning framework by planning a grasp with a desired contact force for a soft antipodal gripper and testing this grasp in a hardware demonstration. Overall, our model and planning approach further enhance soft and continuum robots where they already excel: utilizing contact with the world to achieve their goals with a gentle touch.

Keywords: Modeling, Control, and Learning for Soft Robots, Soft Robot Applications, Contact Modeling
Abstract: Finite deformation is the principal actuation basis of elastomer-based pneumatic soft actuators. Desired deformation behavior is the key design requirement for such actuators. The objective of current study is to optimize the design of a flat shell gripper and to investigate its interaction with a cylindrical object. Herein, we propose an analytical model for a membrane-based flat shell gripper. The model is based on finite strain membrane theory and neo-Hookean material. The proposed model considers the contact interaction of the actuator with flat and cylindrical rigid substrates. The model is developed for three different states of the actuator: (a) free-space; (b) contact with a flat substrate; and (c) contact with a cylindrical substrate. In application, the model was used to predict the relative position and air pressure required to grasp a cylindrical object by a parallel two-fingered shell gripper. Additionally, the frictional behavior of the actuator in contact with a cylindrical substrate is investigated. The model involves only solving nonlinear algebraic equations and is computationally efficient. The theoretically predicted deformation behavior of the actuator is experimentally validated via free-space deformation, force measurement, and grasping tests.

Keywords: Modeling, Control, and Learning for Soft Robots, Soft Robot Materials and Design, Soft Robot Applications
Abstract: Soft robotics has the potential to revolutionize robotic locomotion, in particular, soft robotic swimmers offer a minimally invasive and adaptive solution to explore and preserve our oceans. Unfortunately, current soft robotic swimmers are vastly inferior to evolved biological swimmers, especially in terms of controllability, efficiency, maneuverability, and longevity. Additionally, the tedious iterative fabrication and empirical testing required to design soft robots has hindered their optimization. In this work, we tackle this challenge by providing an efficient and straightforward pipeline for designing and fabricating soft robotic swimmers equipped with electrostatic actuation. We streamline the process to allow for rapid additive manufacturing, and show how a differentiable simulation can be used to match a simplified model to the real deformation of a robotic swimmer. We perform several experiments with the fabricated swimmer by varying the voltage and actuation frequency of the swimmer's antagonistic muscles. We show how the voltage and frequency vary the locomotion speed of the swimmer while moving in liquid oil and observe a clear optimum in forward swimming speed. The differentiable simulation model we propose has various downstream applications, such as control and shape optimization of the swimmer; optimization results can be directly mapped back to the real robot through our sim-to-real matching.

Keywords: Modeling, Control, and Learning for Soft Robots, Soft Sensors and Actuators
Abstract: A soft robotic arm should ideally be working efficiently, robustly, and safely in human-centered environments to provide assistance in real-world situations. For this goal, soft robots need to be able to estimate their state and external interactions based on (proprioceptive) sensors. Estimating disturbances allows a soft robot to perform desirable force control. Even in the case of rigid manipulators, force estimation at the end-effector is seen as a non-trivial problem. And indeed, other current approaches to address this challenge have shortcomings that prevent their general application. They are often based on simplified soft dynamic models, such as the ones relying on a piece-wise constant curvature approximation or matched rigid-body models that do not represent enough details of the problem. Thus, the applications needed for complex human-robot interaction can not be built. Finite element method (FEM) based modelings allow for predictions of soft robot dynamics in a more generic fashion. Here, using the soft robot modeling capabilities of the framework SOFA, we build a detailed FEM model of a multi-segment soft continuum robotic arm composed of compliant deformable materials and fiber-reinforced pressurized actuation chambers. In addition, a model for sensors that provide orientation output is presented. This model is used to establish a state observer for the manipulator. The sensor model is adequate for representing the output of flexible bend sensors as well as orientations provided by IMUs or coming from tracking systems, all of which are popular choices in soft robotics. Model parameters were calibrated to match imperfections of the manual fabrication process using physical experiments. We then solve a quadratic programming inverse dynamics problem to compute the components of external force that explain the pose error. Our experiments show an average force estimation error of around 1.2%. As the methods proposed are generic, these results are encouraging for the task of building soft robots exhibiting complex, reactive, sensor-based behavior that can be deployed in human-centered environments.

Keywords: Modeling, Control, and Learning for Soft Robots, Deep Learning Methods
Abstract: From a medical standpoint, detecting the size and shape of hard inclusions hidden in soft three-dimensional objects is of great significance for early detection of cancer through palpation. Soft robots, especially soft grippers, substantially broaden robots' palpation capabilities from soft to hard materials without the assistance of a camera. We have recently introduced a CNN-Bayes approach which added a Naive Bayes classifier to a convolutional neural network (CNN) architecture called SoftTactNet for variable stiffness object recognition on a three-finger FinRay soft gripper. SoftTactNet itself lacks uncertainty estimations though it can reach a certain level of recognition accuracy. In this paper, We further improve the framework by merging Bayes method directly into CNN architectures and build a new Bayes-SoftTactNet for object recognition. The new approach, using a prior distribution instead of point estimation, allows the network to present results with uncertainty estimates. We conduct new experiments using the same soft gripper with tactile sensor arrays to grasp different variable stiffness objects surrounded by non-different soft material and generate tactile images as dataset. The results show that our new algorithm is more efficient than the previous approach and still able to achieve higher recognition accuracy than general deterministic CNNs.

Keywords: Visual-Inertial SLAM, Localization, Mapping
Abstract: Simultaneous Localization and Mapping (SLAM) algorithms perform visual-inertial estimation via filtering or batch optimization methods. Empirical evidence suggests that filtering algorithms are computationally faster, while optimization methods are more accurate. This work presents an optimization-based framework that unifies these approaches, and allows users to flexibly implement different design choices, e.g., the number and types of variables maintained in the algorithm at each time. We prove that filtering methods correspond to specific design choices in our generalized framework. We then reformulate the Multi-State Constrained Kalman Filter (MSCKF) and contrast its performance with that of sliding-window based filters. Our approach modularizes state-of-the-art SLAM algorithms to allow for adaptation to various scenarios. Experiments on the EuRoC MAV dataset verify that our implementations of these algorithms are competitive with the performance of off-the-shelf implementations in the literature. Using these results, we explain the relative performance characteristics of filtering and batch-optimization based algorithms in the context of our framework. We illustrate that under different design choices, our empirical performance interpolates between those of state-of-the-art approaches.

Keywords: Perception-Action Coupling, Motion and Path Planning, Field Robots
Abstract: Autonomous exploration in unknown environments is a fundamental task for robots. Existing approaches mostly were concentrated on the efficiency of the exploration with the assumption of perfect state estimation, but the drift of pose estimation in visual SLAM occurs frequently and is detrimental to robot's localization safety and exploration performance. In this paper, a perception-aware exploration(PAE) method is proposed for rapidly and safely autonomous exploration in outdoor environments. The adaptive semantic perception is proposed to improve the robustness of perceptual ability, and based on the perception module, both the selection of exploration goal on a novel weighted information gain and the path planning can avoid the areas with high localization uncertainty. In addition, thanks to the proposed pipeline, including the scan-based frontier detection, kd-tree based map prediction and suboptimal frontier buffer strategy, the PAE planner can explore the environment with high success rate and high efficiency. Several simulations are performed to verify the effectiveness of our methods.

Keywords: Visual-Inertial SLAM, Omnidirectional Vision
Abstract: We present a Visual-Inertial Odometry (VIO) algorithm with multiple non-overlapping monocular cameras aiming at improving the robustness of the VIO algorithm. An initialization scheme and tightly-coupled bundle adjustment for multiple non-overlapping monocular cameras are proposed. With more stable features captured by multiple cameras, VIO can maintain stable state estimation, especially when one of the cameras tracked unstable or limited features. We also address the high CPU usage rate brought by multiple cameras by proposing a GPU-accelerated frontend. Finally, we use our pedestrian carried system to evaluate the robustness of the VIO algorithm in several challenging environments. The results show that the multi-camera setup yields significantly higher estimation robustness than a monocular system while not increasing the CPU usage rate (reducing the CPU resource usage rate and computational latency by 40.4% and 50.6% on each camera). A demo video can be found at https://youtu.be/r7QvPth1m10.

Keywords: Robotics in Hazardous Fields, Environment Monitoring and Management, Probabilistic Inference
Abstract: This paper advocates the Gaussian belief propagation solver for factor graphs in the case of gas distribution mapping to support an olfactory sensing robot. The local message passing of belief propagation moves away from the standard Cholesky decomposition technique, which avoids solving the entire factor graph at once and allows for only areas of interest to be updated more effectively. Implementing a local solver means that iterative updates to the distribution map can be achieved orders of magnitude quicker than conventional direct solvers which scale computationally to the size of the map. After defining the belief propagation algorithm for gas mapping, several state of the art message scheduling algorithms are tested in simulation against the standard Cholesky solver for their ability to converge to the exact solution. Testing shows that under the wildfire scheduling method for a large urban scenario, that distribution maps can be iterated at least 10 times faster whilst still maintaining exact solutions. This move to an efficient local framework allows future works to consider 3D mapping, predictive utility and multi-robot distributed mapping.

Keywords: Search and Rescue Robots, Data Sets for Robotic Vision, Object Detection, Segmentation and Categorization
Abstract: Sensor-equipped unoccupied aerial vehicles (UAVs) have the potential to help reduce search times and alleviate safety risks for first responders carrying out Wilderness Search and Rescue (WiSAR) operations, the process of finding and rescuing person(s) lost in wilderness areas. Unfortunately, visual sensors alone do not address the need for robustness across all the possible terrains, weather, and lighting conditions that WiSAR operations can be conducted in. The use of multi-modal sensors, specifically visual-thermal cameras, is critical in enabling WiSAR UAVs to perform in diverse operating conditions. However, due to the unique challenges posed by the wilderness context, existing dataset benchmarks are inadequate for developing vision-based algorithms for autonomous WiSAR UAVs. To this end, we present WiSARD, a dataset with roughly 56,000 labeled visual and thermal images collected from UAV flights in various terrains, seasons, weather, and lighting conditions. To the best of our knowledge, WiSARD is the first large-scale dataset collected with multi-modal sensors for autonomous WiSAR operations. We envision that our dataset will provide researchers with a diverse and challenging benchmark that can test the robustness of their algorithms when applied to real-world (life-saving) applications.

Link to dataset: https://sites.google.com/uw.edu/wisard/

Keywords: Visual-Inertial SLAM, Vision-Based Navigation
Abstract: In this paper, we introduce an event-based visual odometry and mapping framework that relies on decaying event-based corners. Event cameras, unlike conventional cameras, can provide sensor data during high-speed motions or in scenes with high dynamic ranges. Rather than providing intensity information at a global shutter rate, events are triggered asynchronously depending on whether there is a change in brightness at the pixel location. This novel sensing paradigm calls for unconventional ego-motion estimation techniques to address these new challenges. The key aspect of our framework is the use of a continuous representation of inertial measurements to characterise the system's motion which accommodates the asynchronous nature of the event data while estimating a discrete state in an optimisation-based approach. The proposed method relies on corners extracted from events-only data and associates them with a spatio-temporal locality scheme based on exponential decay. Event tracks are then tightly coupled with temporally accurate preintegrated inertial measurements, allowing for the estimation of ego-motion and a sparse map. The proposed method is evaluated on the Event Camera Dataset showing performance against the state-of-art in event-based visual-inertial odometry.

Keywords: Visual-Inertial SLAM, Vision-Based Navigation, Sensor Fusion
Abstract: In this article, we propose a visual-inertial navigation system that directly minimizes a photometric error without an explicit data-association. We focus on the photometric error parametrized by pose and structure parameters that is highly nonconvex due to the nonlinearity of image intensity. The key idea is to introduce an optimal intensity gradient that accounts for a projective uncertainty of a pixel. Ensembles sampled from the state uncertainty contribute to the proposed gradient and yield a correct update direction even in a bad initialization point. We present two sets of experiments to demonstrate the strengths of our framework. First, a thorough Monte Carlo simulation in a virtual trajectory is designed to reveal robustness to large initial uncertainty. Second, we show that the proposed framework can achieve superior estimation accuracy with efficient computation time over state-of-the-art visual-inertial fusion methods in a real-world UAV flight test, where most scenes are composed of a featureless floor.

Keywords: SLAM, Mapping, Autonomous Vehicle Navigation
Abstract: This article presents FAST-LIO2: a fast, robust, and versatile LiDAR-inertial odometry framework. Building on a highly efficient tightly coupled iterated Kalman filter, FAST-LIO2 has two key novelties that allow fast, robust, and accurate LiDAR navigation (and mapping). The first one is directly registering raw points to the map (and subsequently update the map, i.e., mapping) without extracting features. This enables the exploitation of subtle features in the environment and, hence, increases the accuracy. The elimination of a hand-engineered feature extraction module also makes it naturally adaptable to emerging LiDARs of different scanning patterns; the second main novelty is maintaining a map by an incremental k-dimensional (k-d) tree data structure, incremental k-d tree ( ikd-Tree ), that enables incremental updates (i.e., point insertion and delete) and dynamic rebalancing. Compared with existing dynamic data structures (octree, R -tree, and nanoflann k-d tree), ikd-Tree achieves superior overall performance while naturally supports downsampling on the tree. We conduct an exhaustive benchmark comparison in 19 sequences from a variety of open LiDAR datasets. FAST-LIO2 achieves consistently higher accuracy at a much lower computation load than other state-of-the-art LiDAR-inertial navigation systems. Various real-world experiments on solid-state LiDARs with small field of view are also conducted. Overall, FAST-LIO2 is computationally efficient (e.g., up to 100 Hz odometry and mapping in large outdoor environments), robust (e.g., reliable pose estimation in cluttered indoor environments with rotation up to 1000 deg/s), versatile (i.e., applicable to both multiline spinning and solid-state LiDARs, unmanned aerial vehicle (UAV) and handheld platforms, and Intel- and ARM-based processors), while still achieving a higher accuracy than existing methods. Our implementation of the system FAST-LIO2 and the data structure ikd-Tree are both open-sourced on Github.

Keywords: Field Robots, Intelligent Transportation Systems, SLAM
Abstract: In this paper, we present a global navigation satellite system (GNSS) aided LiDAR-visual-inertial scheme, RailLoMer-V, for accurate and robust rail vehicle localization and mapping. RailLoMer-V is formulated atop a factor graph and consists of two subsystems: an odometer assisted LiDAR-inertial system (OLIS) and an odometer integrated Visual-inertial system (OVIS). Both the subsystem exploits the typical geometry structure on the railroads. The plane constraints from extracted rail tracks are used to complement the rotation and vertical errors in OLIS. Besides, the line features and vanishing points are leveraged to constrain rotation drifts in OVIS. The proposed framework is extensively evaluated on datasets over 800 km, gathered for more than a year on both general-speed and high-speed railways, day and night. Taking advantage of the tightly-coupled integration of all measurements from individual sensors, our framework is accurate to long-during tasks and robust enough to grievously degenerated scenarios (railway tunnels). In addition, the real-time performance can be achieved with an onboard computer.

Keywords: Surgical Robotics: Steerable Catheters/Needles, Surgical Robotics: Planning, Medical Robots and Systems
Abstract: Steerable needles are medical devices with the ability to follow curvilinear paths to reach targets while circumventing obstacles. In the deployment process, a human operator typically places the steerable needle at its start position on a tissue surface and then hands off control to the automation that steers the needle to the target. Due to uncertainty in the placement of the needle by the human operator, choosing a start position that is robust to deviations is crucial since some start positions may make it impossible for the steerable needle to safely reach the target. We introduce a method to efficiently evaluate steerable needle motion plans such that they are safe to variation in the start position. This method can be applied to many steerable needle planners and requires that the needle’s orientation angle at insertion can be robotically controlled. Specifically, we introduce a method that builds a funnel around a given plan to determine a safe insertion surface corresponding to insertion points from which it is guaranteed that a collision-free motion plan to the goal can be computed. We use this technique to evaluate multiple feasible plans and select the one that maximizes the size of the safe insertion surface. We evaluate our method through simulation in a lung biopsy scenario and show that the method is able to quickly find needle plans with a large safe insertion surface.

Keywords: Surgical Robotics: Steerable Catheters/Needles
Abstract: The introduction of neuroendoscopy, microneurosurgery, neuronavigation, and intraoperative imaging for surgical operations has made significant improvements over other traditionally invasive surgical techniques. The integration of magnetic resonance imaging (MRI)-driven surgical devices with intraoperative imaging and endoscopy can enable further advancements in surgical treatments and outcomes. This work proposes the design and development of an MRI-driven endoscope leveraging the high (3-7 T), external magnetic field of an MR scanner for heat-mitigated steering within the ventricular system of the brain. It also demonstrates the effectiveness of a Lorentz force-based grasper for diseased tissue manipulation and ablation. Feasibility studies show the neuroendoscope can be steered precisely within the lateral ventricle to locate a tumor using both MRI and endoscopic guidance. Results also indicate grasping forces as high as 31 mN are possible and power inputs as low as 0.69 mW can cause cancerous tissue ablation. These findings enable further developments of steerable devices using MR imaging integrated with endoscopic guidance for improved outcomes.

Keywords: Medical Robots and Systems, Simulation and Animation, Visual Servoing
Abstract: Robot-assisted gastrointestinal endoscopic surgery (GES) as a kind of natural orifice transluminal endoscopic surgery (NOTES) is the next-generation minimally invasive surgery (MIS). Besides, rendering certain autonomy to a Gastrointestinal Endoscopic Surgical Robot (GESR) is currently promising but highly challenging. Therefore, to accelerate the development and augment the autonomy of GESR, we use CoppeliaSim to develop the first robotic simulator for the GESR system (GESRsim) based on our previous design. The GESRsim provides several 3D models and kinematics of our designed manipulators and endoscopic snake bone. Additionally, we build several scenes for robotic GES training and then utilize different programming interfaces to perform teleoperation. Furthermore, several advanced control algorithms, including visual servoing (VS) and deep reinforcement learning (DRL), are implemented to verify the performance of the GESRsim.

Keywords: Surgical Robotics: Steerable Catheters/Needles, Data Sets for Robot Learning, Flexible Robotics
Abstract: Establishing a physics-based model capturing the kinetostatic behavior of concentric tube continuum robots is challenging as elastic interactions between the flexible tubes constituting the robot result in a highly non-linear problem. The Goldstandard physics-based model using the Cosserat theory of elastic rods achieves reasonable approximations with 1.5-3% with respect to the robot's length, if well-calibrated. Learning-based models of concentric tube continuum robots have been shown to outperform the Goldstandard model with approximation errors below 1%. Yet, the merits of learning-based models remain largely unexplored as no common dataset and benchmark exist.

In this paper, we present a dataset captured from a three-tube concentric tube continuum robot for use in learning-based kinematics research. The dataset consists of 100,000 joint configurations and the corresponding four 6 dof sensors in SE(3) measured with an electromagnetic tracking system (github.com/ContinuumRoboticsLab/CRL-Dataset-CTCR-Pose). With our dataset, we empower the continuum robotics and machine learning community to advance the field. We share our insights and lessons learned on joint space representation, shape representation in task space, and sampling strategies. Furthermore, we provide benchmark results for learning the forward kinematics using a simple, shallow feedforward neural network. The benchmark results for the tip error are 0.74 mm w.r.t. position (0.4 % of total robot length) and 6.49 grad w.r.t. orientation.

Keywords: Soft Robot Applications, Medical Robots and Systems, Soft Sensors and Actuators
Abstract: Determining the externally-induced motion of a soft robot in minimally-invasive procedures is highly challenging and commonly demands specific tools and dedicated sensors. Intrinsic force sensing paired with a model describing the robot's compliance offers an alternative pathway which relies heavily on knowledge of the characteristic mechanical behaviour of the investigated system. In this work, we apply quasi-static intrinsic force sensing to a miniature, parallel soft robot designed for endoluminal ear interventions. We characterize the soft robot's nonlinear mechanical behaviour and devise methods for inferring forces applied to the actuators of the robot from fluid pressure and volume information of the working fluid. We demonstrate that it is possible to detect the presence of an external contact acting on the soft robot's actuators, infer the applied reaction force with an accuracy of 28.1mN and extrapolate from individual actuator force sensing to determining forces acting on the combined parallel soft robot when it is deployed in a lumen, which can be achieved with an accuracy of 75.45mN for external forces and 0.47Nmm for external torques. The intrinsically-sensed external forces can be employed to estimate the induced motion of the soft robot in response to these forces with an accuracy of 0.11mm in translation and 2.47deg in rotational deflection. The derived methodologies could enable designs for more perceptive endoscopic systems and pave the way for developing sensing and control strategies in endoluminal and transluminal soft robots.

Keywords: Surgical Robotics: Steerable Catheters/Needles, Medical Robots and Systems
Abstract: Continuum robots such as robotic catheters are increasingly being used in minimally invasive surgery. Compliance contributes to enhanced safety during e.g. catheter insertion, however, estimation of contact force and location may help clinicians avoiding exerting excessive force. Ultimately this could lead to faster and safer interventions. Researchers proposed force sensors integrated in the catheter tip in the past. However, such sensors add extra complexity to the catheter design. Also, tip force sensors do not provide insights on forces that act along the catheter length. This paper proposes a data-driven approach for localizing contact forces that appear over the length of the catheter. The proposed approach consists of a collision detection method and a contact localization method. The framework only requires the measurement of the catheter's shape which can be done by an embedded multi-core Fiber Bragg Grating fiber. The method was validated experimentally with a 3D-printed continuum robot with an integrated multi-core fiber. A second contact localization method which is based on identifying the discontinuity in the measured curvature, is also implemented and compared with the proposed method. The static and dynamic experiments show a mean average localization error of 2.3 mm and 4.3 mm which correspond to respectively 3.3% and 6.1% of a 70 mm long flexible robot. These findings demonstrate that the proposed framework outperforms the previous methods and yields promising results. The contact state estimation algorithm can detect collisions in at most approximately 1.08s.

Keywords: Surgical Robotics: Steerable Catheters/Needles, Medical Robots and Systems
Abstract: In cardiovascular interventions, when steering catheters and especially robotic catheters, great care should be paid to prevent applying too large forces on the vessel walls as this could dislodge calcifications, induce scars or even cause perforation. To address this challenge, this paper presents a novel compliant motion control algorithm that relies solely on position sensing of the catheter tip and knowledge of the catheter's behavior. The proposed algorithm features a data-driven tip position controller. The controller is trained based on a so-called control Long Short-Term Memory Network (control-LSTM). Trajectory following experiments on four different trajectories are conducted to validate the quality of the proposed control-LSTM. The performance was compared with the performance of a controller that makes use of an analytical hysteresis model, i.e. the inverse Deadband Rate-Dependent Prandtl-Ishlinskii (IDRDPI) model. Results demonstrated superior positioning capability with sub-degree precision of the new approach in the presence of severe rate-dependent hysteresis. Experiments both in a simplified setup as well as in an aortic phantom further show that the proposed approach allows reducing the interaction forces with the environment by around 70%. This work shows how deep learning can be exploited advantageously to avoid tedious modeling that would be needed to precisely steer continuum robots in constrained environments such as the patient's vasculature.

Keywords: Surgical Robotics: Planning, Medical Robots and Systems, Vision-Based Navigation
Abstract: Flexible Endoscopes (FEs) for colonoscopy present several limitations due to their inherent complexity, resulting in patient discomfort and lack of intuitiveness for clinicians. Robotic FEs with autonomous control represent a viable solution to reduce the workload of endoscopists and the training time while improving the procedure outcome. Prior works on autonomous endoscope FE control use heuristic policies that limit their generalisation to the unstructured and highly deformable colon environment and require frequent human intervention. This work proposes an image-based FE control using Deep Reinforcement Learning, called Deep Visuomotor Control (DVC), to exhibit adaptive behaviour in convoluted sections of the colon. DVC learns a mapping between the images and the FE control signal. A first user study of 20 expert gastrointestinal endoscopists was carried out to compare their navigation performance with DVC using a realistic virtual simulator. The results indicate that DVC shows equivalent performance on several assessment parameters, being more safer. Moreover, a second user study with 20 novice users was performed to demonstrate easier human supervision compared to a state-of-the-art heuristic control policy. Seamless supervision of colonoscopy procedures would enable endoscopists to focus on the medical decision rather than on the control of FE.

Keywords: Surgical Robotics: Steerable Catheters/Needles, Soft Robot Materials and Design, Medical Robots and Systems
Abstract: Variable stiffness catheters typically rely on thermally induced stiffness transitions with a transition temperature above body temperature. This imposes considerable safety limitations for medical applications. In this work, we present a variable stiffness catheter using a hybrid control strategy capable of actively heating and actively cooling the catheter material. The proposed catheter is made of a single biocompatible shape memory polymer, which significantly increases its manufacturability and scalability compared to existing designs. Increased safety is obtained by ensuring a lower-risk compliant state at body temperature while maintaining higher stiffness ranges in actively controlled states. Additionally, the combined use of variable stiffness and magnetic actuation increases the dexterity and steerability of the device compared to existing robotic tools.

Keywords: Compliance and Impedance Control, Aerial Systems: Applications, Machine Learning for Robot Control
Abstract: The recent development of novel aerial vehicles capable of physically interacting with the environment leads to new applications such as contact-based inspection. These tasks require the robotic system to exchange forces with partially-known environments, which may contain uncertainties including unknown spatially-varying friction properties and discontinuous variations of the surface geometry. Finding a solution that senses, adapts, and remains robust against these environmental uncertainties remains an open challenge. This paper presents a learning-based adaptive control strategy for aerial sliding tasks. In particular, the gains of a standard impedance controller are adjusted in real-time by a neural network policy based on proprioceptive and tactile sensing. This policy is trained in simulation with simplified actuator dynamics in a student-teacher learning setup. The real-world performance of the proposed approach is verified using a tilt-arm omnidirectional flying vehicle. The proposed controller structure combines data-driven and model-based control methods, enabling our approach to successfully transfer directly and without adaptation from simulation to the real platform. We attribute the success of the sim-to-real transfer to the inclusion of feedback control in the training and deployment. We achieved tracking performance and disturbance rejection that cannot be achieved using fine-tuned state of the art interaction control method.

Keywords: Compliance and Impedance Control, Energy and Environment-Aware Automation, Human-Robot Collaboration
Abstract: Tactile robots shall be deployed for dynamic task execution in production lines with small batch sizes. Therefore, these robots should have the ability to respond to changing conditions and be easy to (re-)program. Operating under uncertain environments requires unifying subsystems such as robot motion and force policy into one framework, referred to as tactile skills. In this paper, we propose the enhancement of these skills for passivity-based skill motion learning in stiffness-adaptive unified force-impedance control. To achieve the increased level of adaptability, we represent all tactile skills by three basic primitives: contact initiation, manipulation, and contact termination. To ensure passivity and stability, we develop an energy-based approach for unified force-impedance control that allows humans to teach the robot motion through physical interaction during the execution of a tactile task. We incorporate our proposed framework into a tactile robot to experimentally validate the motion adaptation by interaction performance and stability of the control. While the polishing task is presented as our use case through the paper, the experiments can also be carried out with various tactile skills. Finally, the results show the novel controller's stability and passivity to contact-loss and stiffness adaptation, leading to successful programming by interaction.

Keywords: Compliance and Impedance Control, Learning from Demonstration, Probabilistic Inference
Abstract: One of the major challenges in learning from demonstration is to teach the robot a wider set of task features than the plain trajectories to be followed. In this sense, one key parameter is stiffness, i.e., the rigidity that the manipulator should exhibit when performing a task. The required robot stiffness is often not known a priori and varies along the execution of the task, thus its profile needs to be inferred from the demonstrations. In this work, we propose a novel, force-based algorithm for inferring time-varying stiffness profiles, leveraging the relationship between stiffness and tracking error, and involving human-robot interaction. We begin by gathering a set of demonstrations with kinesthetic teaching. Then, the robot executes a perturbed reference, obtained from these demonstrations by means of Gaussian process regression, and the human intervenes if the perturbation makes the manipulator deviate from its expected behaviour. Human intervention is measured and used to infer the desired control stiffness. In the experiments section, we show that our algorithm can be combined with different types of force sensors, and provide suitable processing algorithms. Our approach correctly infers the stiffness profiles from the force and electromyography sensors, their combination permitting also to comply with the physical constraints imposed by the environment. This is demonstrated in three experiments of increasing complexity: a motion in free Cartesian space, a rigid assembly task, and bed-making.

Keywords: Compliance and Impedance Control, Legged Robots, Whole-Body Motion Planning and Control
Abstract: Quadrupedal manipulators require to be compliant when dealing with external forces during autonomous manipulation, tele-operation or physical human-robot interaction. This paper presents a whole-body controller that allows for the implementation of a Cartesian impedance control to coordinate tracking performance and desired compliance for the robot base and manipulator arm. The controller is formulated through an optimization problem using Quadratic Programming (QP) to impose a desired behavior for the system while satisfying friction cone constraints, unilateral force constraints, joint and torque limits. The presented strategy decouples the arm and the base of the platform, enforcing the behavior of a linear double-mass spring damper system, and allows to independently tune their inertia, stiffness and damping properties. The control architecture is validated through a set of simulations using the 90kg HyQ robot equipped with a 7-DoF manipulator arm. Simulation results show the impedance rendering performance when external forces are applied at the arm's end-effector. The paper presents results for full stance condition (all legs on the ground) and, for the first time, also shows how the impedance rendering is affected by the contact conditions during a dynamic gait.

Keywords: Compliance and Impedance Control, Mechanism Design, Soft Robot Applications
Abstract: Electro-adhesive clutches have become effective tools for variable stiffness functions in many robotic systems due to their light weight, high speed and strong brake force. In this paper, we present a novel, tubular design of an electro-adhesive clutch. Our clutch consists of flexible electrode sheets rolled into a tubular structure. This design allows encapsulating large electrode areas in a compact size for strong brake force. Additionally, the tubular structure acts as a guide for directional sliding without external guides. The structure also ensures that the electrode surfaces are encapsulated, preventing the accumulation of dust and thus leading to reliable performance. This structure is therefore an improvement over the commonly used planar designs. The characterization of the electro-adhesive tubular clutch shows that the frictional force increases with the increase of the electrode contact area, the decrease of the roll diameter and the dielectric layer thickness. A retractable tubular clutch is made by fixing an elastic cable along the clutch axis and achieves a stiffness change factor up to 260. Applications of this retractable clutch in robotics to achieve variable stiffness are demonstrated in two systems: a tensegrity structure and a wing skeleton. Changes in stiffness by 13.2 and 30.2 times are achieved for the two systems, respectively. The proposed tubular clutch is an effective means of achieving variable stiffness, particularly in the case of robotic systems that transmit forces through tensioned cables.

Keywords: Compliance and Impedance Control, Medical Robots and Systems, Physically Assistive Devices
Abstract: This paper proposes a variable impedance control architecture to facilitate eye surgery training in a dual-user haptic system. In this system, an expert surgeon (the trainer) and a novice surgeon (the trainee) collaborate on a surgical procedure using their own haptic devices. The mechanical impedance parameters of the trainer's haptic device remain constant during the operation, whereas those of the trainee vary with his/her proficiency level. The trainee's relative proficiency might be objectively quantified in real-time based on position error between the trainer and the trainee. The proposed architecture enables the trainer to intervene in the training process as needed to ensure the trainee is following the right course of action and to avoid the trainee's from potential tissue injuries. The stability of the overall nonlinear closed-loop system has been investigated using the input-to-state stability (ISS) criterion. High-gain observer with unknown inputs is considered in this work to estimate the interaction forces. Simulation and experimental results under different scenarios confirm the effectiveness of the proposed control methods.

Keywords: Compliance and Impedance Control, Compliant Joints and Mechanisms, Learning from Demonstration
Abstract: A robot designed to coexist and work with humans in the same workspace should be able to work at the same speed as humans and have safe contact with humans and with the environment. However, when a robot arm has been given flexibility through mechanisms and controls for the purpose of coexistence, it is difficult for it to perform tasks at the speed and accuracy desired by humans if it is moved simply by using conventional position-based controls. With such an arm, we consider that the use of learning-based control is necessary to achieve both safety and speed. Therefore, we prototyped a low-inertia, high-backdrivability arm as a platform for studying learning-based control and tested two types of learning-based control. This paper describes our design process, in which hardware suitable for learning-based control was developed according to the requirements of the specific task. It also presents the results of our evaluation experiments, in which tasks involving quick movements and motion requiring physical contact with an object were performed using learning-based control.

Keywords: Prosthetics and Exoskeletons, Rehabilitation Robotics, Medical Robots and Systems
Abstract: One typical application of a robotic exoskeleton is to automate rehabilitation, where the robot is worn by a stroke patient and provides assistance to help perform repetitive motions and regain motor functions. The deployment of exoskeletons can alleviate the shortage of experienced therapists and would also play a vital role in countries with aging populations. However, the intelligence level of existing exoskeletons is relatively low, wherein a robot cannot adapt to either the online change of a subject's motion (e.g., the gait pattern) or the variation of his/her body parameters (i.e., a new subject who is going to wear the robot), potentially resulting in conflict between human and robot and possibly even leading to physical damage. As such, the application of a robotic exoskeleton in clinical studies is limited. This article introduces a new bilateral exoskeletal assistive robot for rehabilitation (i.e., BEAR-H) in which the main novelty is the integration of multiple intelligent features, such as gait recognition and synchronization, cloud-computing diagnosis, and individualized gait generation. Such an integration helps the robot to better understand the patient’s condition and hence provide effective assistance, by the end achieving smart rehabilitation. BEAR-H is a successfully commercialized product, and its performance has been validated in actual clinical studies with 30 patients, producing experimental results from different aspects that are analyzed and presented.

Keywords: Prosthetics and Exoskeletons, Compliance and Impedance Control, Optimization and Optimal Control
Abstract: Although the average healthy adult transitions from sit to stand over 60 times per day, the majority of robotic prosthesis control research has focused on walking. In this paper, we present a data-driven controller that enables sitting, standing, and walking with minimal tuning. Our controller comprises two high level modes of sit/stand and walking, and we develop heuristic biomechanical rules to control transitions between the two. We use a phase variable based on the user's thigh angle to parameterize both walking and sit/stand motions, where variable impedance control is used during ground contact and position control is used during the swing phase of walking. We extend previous work on data-driven optimization of continuous impedance parameter functions to design the sit/stand control mode using able-bodied data. We test our controller in experiments with a participant with an above-knee amputation, comparing clinical measures including loading symmetry and sit/stand transition time. The participant completed the sit/stand task 20% faster with approximately half of the loading asymmetry relative to his everyday passive prosthesis. The controller also facilitated a timed up and go test involving sitting, standing, walking, and turning, with only a mild (10%) decrease in speed compared to the everyday prosthesis. Our sit/stand/walk controller enables multiple activities of daily life with minimal tuning and mode switching.

Keywords: Software Tools for Robot Programming, Autonomous Vehicle Navigation, Field Robots
Abstract: Occupancy grid maps (OGMs) are fundamental to most systems for autonomous robotic navigation. However, CPU-based implementations struggle to keep up with data rates from modern 3D lidar sensors, and provide little capacity for modern extensions which maintain richer voxel representations. This paper presents OHM, our open source, GPU-based OGM framework. We show how the algorithms can be mapped to GPU resources, resolving difficulties with contention to obtain a successful implementation. The implementation supports many modern OGM algorithms including NDT-OM, NDT-TM, decay-rate and TSDF. A thorough performance evaluation is presented based on tracked and quadruped UGV platforms and UAVs, and data sets from both outdoor and subterranean environments. The results demonstrate excellent performance improvements both offline, and for online processing in embedded platforms. Finally, we describe how OHM was a key enabler for the UGV navigation solution for our entry in the DARPA Subterranean Challenge, which placed second at the Final Event.

Keywords: Software Tools for Robot Programming, Deep Learning Methods, Redundant Robots
Abstract: Inverse kinematics—finding joint poses that reach a given Cartesian-space end-effector pose—is a fundamental operation in robotics, since goals and waypoints are typically defined in Cartesian space, but robots must be controlled in joint space. However, existing inverse kinematics solvers return a single solution, in contrast, systems with more than 6 degrees of freedom support infinitely many such solutions, which can be useful in the presence of constraints, pose preferences or obstacles. We introduce a method that uses a deep neural network to learn to generate a diverse set of samples from the solution space of such kinematic chains. The resulting samples can be generated quickly (2000 solutions in under 10ms) and accurately (to within 10 millimeters and 2 degrees of an exact solution) and can be rapidly refined by classical methods if necessary.

Keywords: Software Tools for Robot Programming, Distributed Robot Systems, Performance Evaluation and Benchmarking
Abstract: Testing and debugging have become major obstacles for robot software development, because of high system complexity and dynamic environments. Standard, middleware-based data recording does not provide sufficient information on internal computation and performance bottlenecks. Other existing methods also target very specific problems and thus cannot be used for multipurpose analysis. Moreover, they are not suitable for real-time applications. In this paper, we present ros2_tracing, a collection of flexible tracing tools and multipurpose instrumentation for ROS 2. It allows collecting runtime execution information on real-time distributed systems, using the low-overhead LTTng tracer. Tools also integrate tracing into the invaluable ROS 2 orchestration system and other usability tools. A message latency experiment shows that the end-to-end message latency overhead, when enabling all ROS 2 instrumentation, is on average 0.0033 ms, which we believe is suitable for production real-time systems. ROS 2 execution information obtained using ros2_tracing can be combined with trace data from the operating system, enabling a wider range of precise analyses, that help understand an application execution, to find the cause of performance bottlenecks and other issues. The source code is available at: https://github.com/ros2/ros2_tracing.

Keywords: Software, Middleware and Programming Environments, Computer Architecture for Robotic and Automation, Hardware-Software Integration in Robotics
Abstract: Hardware acceleration can revolutionize robotics, enabling new applications by speeding up robot response times while remaining power-efficient. However, the diversity of acceleration options makes it difficult for roboticists to easily deploy accelerated systems without expertise in each specific hardware platform. In this work, we address this challenge with RobotCore, an architecture to integrate hardware acceleration in the widely-used ROS 2 robotics software framework. This architecture is target-agnostic (supports edge, workstation, data center, or cloud targets) and accelerator-agnostic (supports both FPGAs and GPUs). It builds on top of the common ROS 2 build system and tools and is easily portable across different research and commercial solutions through a new firmware layer. We also leverage the Linux Tracing Toolkit next generation (LTTng) to enable low-overhead real-time tracing and benchmarking of accelerated ROS 2 systems. To demonstrate the acceleration enabled by this architecture, we use it to deploy a ROS 2 perception computational graph on a CPU and FPGA. We also employ our integrated tracing and benchmarking to analyze bottlenecks, uncovering insights that guide us to improve FPGA communication efficiency. In particular, we design an intra-FPGA ROS 2 node communication queue template and use it in conjunction with FPGA-accelerated nodes to achieve a 24.42% speedup over a CPU.

Keywords: Software Tools for Robot Programming, Engineering for Robotic Systems, Optimization and Optimal Control
Abstract: We present Tasho (Task specification for receding horizon control), an open-source Python toolbox that facilitates systematic programming of optimal control problem (OCP)-based robot motion skills. Separation-of-concerns is followed while designing the components of a motion skill, which promotes their modularity and reusability. This allows us to program complex motion tasks by configuring and composing simpler tasks. We provide templates for several basic tasks like point-to-point and end-effector path-following tasks to speed up prototyping. Internally, the task's symbolic expressions are computed using CasADi and the resulting OCP is transcribed using Rockit. A wide and growing range of mature open-source optimization solvers are supported for solving the OCP. Monitor functions can be easily specified and are automatically deployed with the motion skill, so that the generated motion skills can be easily embedded in a larger control architecture involving higher-level discrete controllers. The motion skills thus programmed can be directly deployed on robot platforms using the C-code generation capabilities of CasADi. The toolbox has been validated through several experiments both in simulation and on physical robot systems. The open-source toolbox can be accessed at: https://gitlab.kuleuven.be/meco-software/tasho

Keywords: Software, Middleware and Programming Environments, Control Architectures and Programming, Autonomous Agents
Abstract: Containerization promises to strengthen platform-independent development, better resource utilization, and secure deployment of software. As these benefits come with negligible overhead in CPU and memory utilization, containerization is increasingly being adopted in mobile robotic applications. An open challenge is supporting software tasks that have mixed-criticality requirements. Even more challenging is the combination of real-time containers with orchestration, which is an emerging paradigm to automate the deployment, networking, scaling, and availability of containerized workloads and services. This paper addresses this challenge by presenting a framework that extends the de-facto reference standard for container orchestration, Kubernetes, to schedule tasks with mixed-criticality requirements. Quantitative experimental results on the software implementing the mission of a Robotnik RB-Kairos mobile robot demonstrate the effectiveness of the proposed approach. The source code is publicly available on GitHub.

Keywords: Software Tools for Robot Programming, Mapping, Motion and Path Planning
Abstract: —In this letter, we present a volumetric mapping system that effectively calculates Occupancy Grid Maps (OGMs) and Euclidean Distance Transforms (EDTs) with parallel computing. Unlike these mappers for high-precision structural reconstruction, our system incrementally constructs global EDT and outputs high-frequency local distance information for online robot motion planning. The proposed system constructs OGM with a massive amount of data from multiple types of sensor inputs. Using GPU programming techniques, the system quickly computes EDT in parallel within a local volume. The new observation is continuously integrated into the global EDT using the parallel wavefront algorithm while preserving the historical observations. Experiments with datasets have shown that our proposed approach outperforms existing state-of-the-art robot mapping systems and is particularly suitable for mapping unexplored areas. In its actual implementations on aerial and ground vehicles, the proposed system achieves realtime performance with limited onboard computational resources.

Keywords: Software Tools for Robot Programming, Software, Middleware and Programming Environments, Software Architecture for Robotic and Automation
Abstract: In the Robot Operating System (ROS), a major middleware for robots, the Transform Library (TF) is a mandatory package that manages transformation information between coordinate systems by using a single-rooted directed tree and providing methods for registering and computing the information. However, the tree has two fundamental problems. The first is its poor scalability: since it accepts only a single thread at a time due to using a single giant lock for mutual exclusion, the access to the tree is sequential. Second, there is a lack of data freshness: it retrieves non-latest synthetic data when computing coordinate transformations because it prioritizes temporal consistency over data freshness. In this paper, we propose methods to solve these problems. First, we decentralize the giant lock to provide performance scalability and show that this results in a throughput 243 times higher than conventional TF on a read-only workload. Second, we design transactional methods based on serializable protocols that prevent anomalies, thus retrieving the freshest data. These transactional methods show a freshness up to 1276 times higher than the conventional one on a read-write combined workload.

Keywords: Software Tools for Benchmarking and Reproducibility, Motion and Path Planning, Collision Avoidance
Abstract: The ability to autonomously navigate safely, especially within dynamic environments is paramount for mobile robotics. In recent years, DRL approaches have shown superior performance in dynamic obstacle avoidance. However, these learning-based approaches are often developed in specially designed simulation environments and are hard to test against conventional planning approaches. Furthermore, the integration and deployment of these approaches into real robotic platforms are not yet completely solved. In this paper, we present Arena-bench, a benchmark suite to train, test, and evaluate navigation planners on different robotic platforms within 3D environments. It provides tools to design and generate highly dynamic evaluation worlds, scenarios, and tasks for autonomous navigation and is fully integrated into the robot operating system. To demonstrate the functionalities of our suite, we trained a DRL agent on our platform and compared it against a variety of existing different model-based and learning-based navigation approaches on a variety of relevant metrics. Finally, we deployed the approaches towards real robots and demonstrated the reproducibility of results. The code is publicly available on github.com/arena-rosnav-3D.

Keywords: Wearable Robotics, Compliance and Impedance Control, Human Performance Augmentation
Abstract: The fit of a wearable device, such as a prosthesis, can be quantitatively characterized by the mechanical coupling at the user-device interface. It is thought that the mechanical impedance, specifically the stiffness and damping, of wearable device interfaces can significantly impact human performance while using them. To test this theory, we develop a forearm-mounted testbed with a motorized, two degree of freedom (2-DOF) gimbal to simulate variations in the mechanical fit of an upper-extremity wearable device during pointing and target tracking tasks. The two gimbal motors are impedance-controlled to vary the mechanical stiffness and damping between the user and the device’s laser pointer end-effector. In this paper, experiments are conducted to determine the torque constants of the motors before implementation in the testbed, and to validate the accuracy of the joint impedance controller. The completed impedance-controlled wearable interface testbed is validated further by comparing the gimbal joint displacements and torques, recorded during 2-DOF base excitation experiments, to MATLAB Simulink simulation data.

Keywords: Wearable Robotics, Physical Human-Robot Interaction, Rehabilitation Robotics
Abstract: Lower Limb Exoskeleton (LLE) has received considerable interest in strength augmentation, rehabilitation, and walking assistance scenarios. For strength augmentation, the LLE is expected to have the capability of reducing metabolic energy. However, the energy for adjusting the Center of Gravity (CoG) is the main part of the energy consumption during walking, especially the walking with loads. This paper proposes a novel Human-exoskeleton Cooperative Balance (HCB) strategy for giving balance ability to the assistive torque and combined with the direction selected by the pilot to realize the balance walking of the human-exoskeleton system. In which, a Dynamic Torque Primitive Model (DTPM) is designed to plan a bionic assistive torque, and the balance parameter obtained by an Inverted Pendulum Model (IPM) is superimposed on it. Finally, the improved balance performance can break the limitation of traditional strategies and substantially increase the efficiency of assistance. We demonstrated the effectiveness of the proposed HCB strategy in the HUman-powered Augmentation Lower EXoskeleton (HUALEX) system. Experimental results indicate that the proposed HCB strategy is more efficient than traditional strategies.

Keywords: Wearable Robotics, Mechanism Design, Prosthetics and Exoskeletons
Abstract: Despite the large amount of available exoskeletons, their use in daily life is still limited due to the absence of testing in real-life environments. This paper aims at presenting the design of a wearable ankle exoskeleton for walking assistance. The system was tested on irregular terrains. Our RANK (Robotic ANKle) is equipped with a servomotor. A four-bar linkage mechanism is used for the torque transmission. An adjustable 3D-printed brace was realized in PC/ABS to connect the exoskeleton to the user shank. The control system processes the output signals of three piezoresistive sensors placed on the insole. Assistive torque was provided during swing phase in order to limit the plantarflexion to avoid the drop-foot. Two healthy male subjects were enrolled in the study. Experimental testing consists of walking tasks performed on three different terrain conditions (flat, soft, and irregular) with and without exoskeleton. Human kinematics was gathered via inertial measurements units (IMUs). The effects of ankle exoskeleton on lower limb joint angles were assessed in terms of range of motion (ROM). Statistical parametric map method was also applied to compare joint angle curves. As expected, a reduction of the ankle ROM in all terrain conditions was found between the trails performed with and without exoskeleton. No effects induced on the hip and knee joint were observed. Moreover, no significant differences have been observed over the almost totality of the gait cycle independently of the terrain conditions. Results demonstrate the capability of the exoskeleton to work properly regardless the type of walking surface.

Keywords: Wearable Robotics, Gesture, Posture and Facial Expressions, Soft Sensors and Actuators
Abstract: Advances in soft sensors coupled with machine learning are enabling increasingly capable wearable systems. Since hand motion in particular can convey useful information for developing intuitive interfaces, glove-based systems can have a significant impact on many application areas. A key remaining challenge for wearables is to capture, process, and analyze data from the high-degree-of-freedom hand in real time.

We propose using a commercially available conductive knit to create an unobtrusive network of resistive sensors that spans all hand joints, coupling this with an accelerometer, and deploying machine learning on a low-profile microcontroller to process and classify data. This yields a self-contained wearable device with rich sensing capabilities for hand pose and orientation, low fabrication time, and embedded activity prediction.

To demonstrate its capabilities, we use it to detect static poses and dynamic gestures from American Sign Language (ASL). By pre-training a long short-term memory (LSTM) neural network and using tools to deploy it in an embedded context, the glove and an ST microcontroller can classify 12 ASL letters and 12 ASL words in real time. Using a leave-one-experiment-out cross validation methodology, networks successfully classify 96.3% of segmented examples and generate correct rolling predictions during 92.8% of real-time streaming trials.

Keywords: Wearable Robotics, Prosthetics and Exoskeletons, Hydraulic/Pneumatic Actuators
Abstract: Upper limb impairments and weakness are common post-stroke and with advanced aging. Rigid exoskeletons have been developed as a potential solution, but have had limited impact. In addition to user concerns about safety, their weight and appearance, the rigid attachment and typical anchoring methods can result in skin damage. In this paper, we present a soft, fabric-based pneumatic sleeve, which can shrink from a loose fit to a tight fit in order to anchor to the limbs temporarily, thereby enabling the application of mechanical assistance only when needed. The sleeve is comfortable, ergonomic and can be embedded unobtrusively with clothing. A mathematical model is built to simulate and design sleeves with different geometric parameters. The best sleeve was capable of generating a friction force of 98 N on the limb when inflated to 25 kPa. This sleeve was used to create a wearable assistive device, integrated with a cable-driven actuator. This device was able to lift a 1.44 kg forearm rig up to 95 degree at low pressure of 20 kPa. The device was tested with six healthy participants, in terms of fit, comfort and assistive functionality. The average acceptable sleeve pressure was found to be 33±4.7 kPa. All participants liked the appearance of the sleeve, with a high average perceived assistance score of 7.33±1.6 (out of 10). The shrink-to-fit sleeve is expected to significantly increase the development and adoption of soft robotic assistive devices and emerging powered clothing.

Keywords: Wearable Robotics, Human Performance Augmentation, Rehabilitation Robotics
Abstract: This paper proposes and verifies a new method, the ring-pull mechanism, to overcome the disadvantages of existing wearable robotic gloves. By attaching a ring to the metacarpopha- langeal (MCP) joint of the finger, the ring-pull mechanism supplements the grasping force of the user, while reducing the weight of the entire wearable robotic glove system. Ring-pull mechanism experiments were conducted to determine which finger combinations had the most positive effect on muscle strength assistance, and through this, the Ring-Pull type Soft Glove (RPSG) was developed. The main body of the developed RPSG is composed of single polymer silicon, a soft material, and is driven by tendon-driven actuation. The tendon path is secured through a tube attached to the palm that matches the direction of the flexor digitorum superficialis (FDS). The new type of wearable robotic glove was manufactured with the proposed mechanism, and excellent fit and strength support effects were confirmed. The RPSG increased the subject's grasping force by 25.69% on average, and the %MVIC data analysis demonstrated that the activation of FDS decreased by about 23.51%. As a result, it was confirmed that the user's muscle efficiency was increased due to the muscle support and muscle function improvement provided by the RPSG.

Keywords: Wearable Robotics, Physically Assistive Devices, Human Performance Augmentation
Abstract: Typical industrial work activities may include a variety of different gestures, entailing the execution of dynamic and static movements. Occupational upper-limb exoskeletons can assist the shoulder complex in both static and dynamic gestures, but the required assistance level may be different according to the tasks. This article presents the design, development, and experimental evaluation of a novel kinematics-based adaptive assistance algorithm for a semi-passive upper-limb exoskeleton. The algorithm uses kinematic signals gathered by onboard sensors to set the assistance amplitude according to the type of movement being executed. Experimental activities were performed to assess the algorithm’s performance. Results show that the algorithm can effectively provide different assistance levels according to the type of task being executed, such as the minimum level for more dynamic tasks and the maximum level for the most static activities. Additionally, compared to working without the exoskeleton, the exoskeleton controlled by the proposed adaptive algorithm can reduce the users’ flexor muscular activity in both dynamic and static tasks, respectively by 24 ± 6% and 42 ± 2%. Similar results were reported for extensor muscles, which reduced their activations by 7 ± 3%, and 40 ± 4% in dynamic and static tasks.

Keywords: Wearable Robotics, Soft Sensors and Actuators, Hydraulic/Pneumatic Actuators
Abstract: Traditional pneumatic artificial muscles (PAMs) have cannot fully satisfy the requirements in wearable applications for safe and sufficient interaction with human body. The requirements include high contraction ratio, self-contained sensing, reconfiguration, locking abilities and no squeezing force worked on human tissue. In this paper, a reconfigurable self-sensing pneumatic artificial muscle (RSPAM) based on modular multi-chamber soft actuator with locking ability is developed, which provides an alternative that can satisfy all the requirements for wearable applications. Contraction principle of the RSPAM, which is transforming the expansion of the soft actuator into contraction by fabric winding, has inherent high contraction ratio (details seen in Appendix) no squeezing on human tissue at all. Self-sensing of contraction stroke based on liquid metal and locking ability based positive pressure jamming are integrated. A driving force of 70.14 N and contraction ratio of 71% were validated at 3 bar air pressure. Modular design of the actuator makes it possible to change configuration by adjusting the actuator amount and the fabric length according to application. In addition, the actuator unit can move on the fabric by itself, allowing an interesting self-adjustment ability to the muscle. The displacement self-sensing ability is verified through square-wave tracking experiments with closed-loop control. Finally, a preliminary test of assisting elbow movement verifies that RSPAM can reduce 25.52% muscle fatigue.

Keywords: Wearable Robotics, Rehabilitation Robotics, Soft Robot Applications
Abstract: This paper proposes a new mechanism, a four-bar linkage-based support hinge mechanism, for assisting with shoulder abduction with artificial muscle based on a shape memory alloy (SMA). An artificial muscle using the SMA coils is designed to lighten the entire system and support the wearer’s movement both actively and passively. It can generate up to 273 N and 180 N with and without energy input, respectively, while weighing only 0.04 kg. Furthermore, to consider the rotation axis shifts of the arm during shoulder abduction, the trajectory of the arm along the shoulder abduction is modeled using the scapulohumeral rhythm, a combined movement of the scapula and humerus. The mechanism is designed to follow the modeled arm trajectory and achieve the required torque to perform shoulder abduction based on a four-bar linkage mechanism. It can generate up to 10.1 Nm and 6.3 Nm of torque with and without energy input, respectively. To verify the assistive effect of the proposed mechanism, electromyography is measured while performing the same exercise requiring shoulder abduction with and without the support of the mechanism. The results show that the proposed mechanism reduces not only muscle load but also fatigue while performing the shoulder abduction.

Keywords: Intention Recognition, Telerobotics and Teleoperation, Human-Robot Collaboration
Abstract: Shared control can help in teleoperated object manipulation by assisting with the execution of the user’s intention. To this end, robust and prompt intention estimation is needed, which relies on behavioral observations. Here, an intention estimation framework is presented, which uses natural gaze and motion features to predict the current action and the target object. The system is trained and tested in a simulated environment with pick and place sequences produced in a relatively cluttered scene and with both hands, with possible hand-over to the other hand. Validation is conducted across different users and hands, achieving good accuracy and earliness of prediction. An analysis of the predictive power of single features shows the predominance of the grasping trigger and the gaze features in the early identification of the current action. In the current framework, the same probabilistic model can be used for the two hands working in parallel and independently, while a rule-based model is proposed to identify the resulting bimanual action. Finally, limitations and perspectives of this approach to more complex, full-bimanual manipulations are discussed.

Keywords: Intention Recognition, Representation Learning, Deep Learning Methods
Abstract: Equipping robots with the ability to infer human intent is a vital precondition for effective collaboration. Most computational approaches towards this objective derive a probability distribution of "intent" conditioned on the robot's perceived state. However, these approaches typically assume task-specific labels of human intent are known a priori. To overcome this constraint, we propose the Disentangled Sequence Clustering Variational Autoencoder (DiSCVAE), a clustering framework capable of learning such a distribution of intent in an unsupervised manner. The proposed framework leverages recent advances in unsupervised learning to disentangle latent representations of sequence data, separating time-varying local features from time-invariant global attributes. As a novel extension, the DiSCVAE also infers a discrete variable to form a latent mixture model and thus enable clustering over these global sequence concepts, e.g. high-level intentions. We evaluate the DiSCVAE on a real-world human-robot interaction dataset collected using a robotic wheelchair. Our findings reveal that the inferred discrete variable coincides with human intent, holding promise for collaborative settings, such as shared control.

Keywords: Intention Recognition, Prosthetics and Exoskeletons, Rehabilitation Robotics
Abstract: This letter presents a method for data-driven user-specific gait speed estimation for people with Spinal Cord Injuries (SCIs) walking in lower-limb exoskeletons. The scarcity of training data for this population is addressed by leveraging common patterns across users that relate gait changes to speed changes. To bootstrap the process, widely available walking data from uninjured individuals was used as a base dataset. The distribution of this data was first transformed to match smaller user-specific training sets from walking trials of subjects with SCIs. User-specific trials were then selected based on the mutual information between gait speed and features for the combined dataset. The resulting selected data was finally used to build a model for estimating the user's intended gait speed. The performance of this approach was evaluated using data from two users with SCIs walking in an EksoGT exoskeleton with a walker or crutches. Estimation trials were compared when using the base data alone versus when providing personalization via the addition of novel data. The average successful estimation of speed-up and slow-down changes increased from 52% to 67% with personalization using only 8 to 12 steps' worth of user-specific data, with a best-case improvement of 32%, from 48% to 80%. Overall, the proposed method uses the mutual information between gait features and speed to provide a reliable alternative to manual data selection while pooling data from healthy and injured individuals.

Keywords: Intention Recognition, Probabilistic Inference, Motion and Path Planning
Abstract: Motion prediction of road users in traffic scenes is critical for autonomous driving systems that must take safe and robust decisions in complex dynamic environments. We present a novel motion prediction system for autonomous driving. Our system is based on the Bayesian inverse planning framework, which efficiently orchestrates map-based goal extraction, a classical control-based trajectory generator and a mixture of experts collection of light-weight neural networks specialised in motion profile prediction. In contrast to many alternative methods, this modularity helps isolate performance factors and better interpret results, without compromising performance. This system addresses multiple aspects of interest, namely multi-modality, motion profile uncertainty and trajectory physical feasibility. We report on several experiments with the popular highway dataset NGSIM, demonstrating state-of-the-art performance in terms of trajectory error. We also perform a detailed analysis of our system's components, along with experiments that stratify the data based on behaviours, such as change-lane versus follow-lane, to provide insights into the challenges in this domain. Finally, we present a qualitative analysis to show other benefits of our approach, such as the ability to interpret the outputs.

Keywords: Intention Recognition, Intelligent Transportation Systems, Agent-Based Systems
Abstract: Predicting object motion behaviour is a challenging but crucial task for safe decision making and path planning for an autonomous vehicle. It is challenging in large part due to the uncertain, multi-modal, and practically intractable set of possible human-human and human-space interactions, especially in urban driving settings. Models solely based on constant velocity or social force have an inherent bias and may lead to inaccurate predictions across the prediction horizon whereas purely data driven approaches suffer from a lack of a holistic set of rules governing predictions. We tackle this problem by introducing MPC-PF: a novel potential field-based trajectory predictor that incorporates social interaction and is able to tradeoff between inherent model biases across the prediction horizon. Through evaluation on a variety of common urban driving scenarios, we show that our model is capable of producing accurate predictions for both short and long term timesteps. We also demonstrate the significance of our model architecture through an ablation study.

Keywords: Intention Recognition, Datasets for Human Motion, Human-Robot Collaboration
Abstract: How to design an optimal wearable device for human movement recognition is vital to reliable and accurate human-machine collaboration. Previous works mainly fabricate wearable devices heuristically. Instead, this paper raises an academic question: can we design an optimization algorithm to optimize the fabrication of wearable devices such as figuring out the best sensor arrangement automatically? Specifically, this work focuses on optimizing the placement of Forcemyography (FMG) sensors for FMG armbands in the application of arm movement recognition. Firstly, based on graph theory, the armband is modeled considering sensors’ signals and connectivity. Then, a Graph-based Armband Modeling Network (GAM-Net) is introduced for arm movement recognition. Afterward, the sensor placement optimization for FMG armbands is formulated and an optimization algorithm with greedy local search is proposed. To study the effectiveness of our optimization algorithm, a dataset for mechanical maintenance tasks using FMG armbands with 16 sensors is collected. Our experiments show that using only 4 sensors optimized with our algorithm can help maintain a comparable recognition accuracy to using all sensors. Finally, the optimized sensor placement result is verified from a physiological view. This work would like to shed light on the automatic fabrication of wearable devices considering downstream tasks, such as human biological signal collection and movement recognition.

Keywords: Intention Recognition, Computer Vision for Transportation, Intelligent Transportation Systems
Abstract: Anticipating the future behavior of pedestrians is a crucial part of deploying Automated Driving Systems (ADS) in urban traffic scenarios. Most recent works utilize a convolutional neural network (CNN) to extract visual information, which is then input to a recurrent neural network (RNN) along with pedestrian-specific features like location and speed to obtain temporal features. However, the majority of these approaches lack the ability to parse the relationships of the related objects in the specific traffic scene, which leads to omitting the interactions between the pedestrians and the interactions between the pedestrians and the traffic. For this purpose, we propose a graph-structured model which can dig out pedestrians' dynamic constraints by constructing a traffic-aware scene graph within each frame. In addition, to capture pedestrian movement more effectively, we also introduce a temporal feature representation model, which first uses inter-frame and intra-frame GRU (II-GRU) to mine inter-frame information and intra-frame information together, and then employs a novel attention mechanism to adaptively generate attention weights. Extensive experiments on the JAAD and PIE datasets prove that our proposed model is effective in reaching and enhancing the state-of-the-art performance.

Keywords: Intention Recognition, AI-Enabled Robotics, Human-Centered Robotics
Abstract: As robots across domains start collaborating with humans in shared environments, algorithms that enable them to reason over human intent are important to achieve safe interplay. In our work, we study human intent through the problem of predicting trajectories in dynamic environments. We explore domains where navigation guidelines are relatively strictly defined but not clearly marked in their physical environments. We hypothesize that within these domains, agents tend to exhibit short-term motion patterns that reveal context information related to the agent's general direction, intermediate goals and rules of motion, e.g., social behavior. From this intuition, we propose Social-PatteRNN, an algorithm for recurrent, multi-modal trajectory prediction that exploits motion patterns to encode the aforesaid contexts. Our approach guides long-term trajectory prediction by learning to predict short-term motion patterns. It then extracts sub-goal information from the patterns and aggregates it as social context. We assess our approach across three domains: humans crowds, humans in sports and manned aircraft in terminal airspace, achieving state-of-the-art performance.

Keywords: Control Architectures and Programming, Planning, Scheduling and Coordination, Engineering for Robotic Systems
Abstract: Performing a complex autonomous mission with a multi-robot system requires to integrate several deliberative approaches to perform task allocation, optimization, and execution control. Implementing such a deliberative architecture is a complex task: it requires the developer to master the decision algorithms themselves (e.g., automated planning models), to have a good knowledge of the involved robotic platforms, and to think about how these elements will be assembled as a system architecture.

We propose a framework to help designing such deliberative architectures. The framework relies on the concept of a hierarchical structure of actors, each actor managing goals with specific planning or optimization approaches, and delegating sub-goals to other actors.

Keywords: Omnidirectional Vision, Data Sets for Robotic Vision, Semantic Scene Understanding
Abstract: Surround-view cameras are a primary sensor for automated driving, used for near-field perception. It is one of the most commonly used sensors in commercial vehicles primarily used for parking visualization and automated parking. Four fisheye cameras with a 190∘ field of view cover the 360∘ around the vehicle. Due to its high radial distortion, the standard algorithms do not extend easily. Previously, we released the first public fisheye surround-view dataset named WoodScape. In this work, we release a synthetic version of the surround-view dataset, covering many of its weaknesses and extending it. Firstly, it is not possible to obtain ground truth for pixel-wise optical flow and depth. Secondly, WoodScape did not have all four cameras annotated simultaneously in order to sample diverse frames. However, this means that multi-camera algorithms cannot be designed to obtain a unified output in birds-eye space, which is enabled in the new dataset. We implemented surround-view fisheye geometric projections in CARLA Simulator matching WoodScape’s configuration and created SynWoodScape. We release 80 k images from the synthetic dataset with annotations for 10+ tasks. We also release the baseline code and supporting scripts.

Keywords: Semantic Scene Understanding, Deep Learning for Visual Perception, RGB-D Perception
Abstract: We present a new solution to the fine-grained retrieval of clean CAD models from a large-scale database in order to recover detailed object shape geometries for RGBD scans. Unlike previous work simply indexing into a moderately small database using an object shape descriptor and accepting the top retrieval result, we argue that in the case of a large-scale database a more accurate model may be found within a neighborhood of the descriptor. More importantly, we propose that the distinctiveness deficiency of shape descriptors at the instance level can be compensated by a geometry-based re-ranking of its neighborhood. Our approach first leverages the discriminative power of learned representations to distinguish between different categories of models and then uses a novel robust point set distance metric to re-rank the CAD neighborhood, enabling fine-grained retrieval in a large shape database. Evaluation on a real-world dataset shows that our geometry-based re-ranking is a conceptually simple but highly effective method that can lead to a significant improvement in retrieval accuracy compared to the state-of-the-art.

Keywords: Semantic Scene Understanding, Object Detection, Segmentation and Categorization, Recognition
Abstract: Several methods of semantic segmentation using light detection and ranging (LiDAR) sensors have been proposed for the recognition of surrounding objects by autonomous driving cars. LiDAR is a sensor that compensates for the weaknesses of other sensors, such as cameras or radar systems, and semantic segmentation assigns a class label to each point in the LiDAR point cloud. Recently, real-time semantic segmentation methods that are capable of processing LiDAR point clouds at frame rates have been proposed. Real-time semantic segmentation is essential for the autonomous driving system because it can output class labels for LiDAR point clouds at high speeds. However, this segmentation method suffers from a delay equal to processing time. To address this challenge, we propose a novel method that combines SalsaNext cite{salsanext}, a method of real-time LiDAR semantic segmentation, and semantic forecasting, which predicts the results of future semantic segmentation. We quantitatively evaluate our method using the Semantic-KITTI dataset, which comprises point cloud data acquired from the LiDAR sensor in the real world, and compare the latency and accuracy of our method with other semantic segmentation methods. Consequently, our method is found to be capable of operating in real-time and with low-latency, and it can achieve a performance similar to that of previously reported real-time semantic segmentation methods.

Keywords: Semantic Scene Understanding, Object Detection, Segmentation and Categorization
Abstract: Context information is important for instance segmentation on point clouds. Existing methods either only use local surroundings by stacking multiple convolution layers or use non-local methods to model long-range interactions. However, they usually directly operate on points which is an unstructured and low-level representation and is highly dependent on context. To address this issue, we propose an effective framework named Implicit-Part Context Aggregation (IPCA), which adopts implicit parts as an intermediate representation and achieves context aggregation through message passing along the implicit part graph. Specifically, we first organize unstructured points into geometrically consistent implicit parts and construct the implicit part graph according to the geometric adjacency. Then, an initial part embedding is extracted using the proposed Implicit Part Network (IPN) which can aggregate point features and capture the intrinsic geometric shape of the part. We further refine the part embedding by a graph reasoning module named Context Aggregation Network (CAN), which helps to make a more precise prediction by well exploiting the context information. Instance proposals are then generated by grouping implicit parts. Finally, we propose an additional step to attribute the entire instance proposal to a Semantic Criterion Net (SCN) to infer the semantics of the instance. The purpose is to correct the semantic prediction errors caused by not knowing the boundary and overall shape of the object in the previous steps. Extensive experiments on two large datasets, ScanNet and 3RScan demonstrate the effectiveness of our method. It outperforms all existing methods on the ScanNet test benchmark and its AP@50 is 9.5 points higher than the baseline.

Keywords: Semantic Scene Understanding, Transfer Learning, Deep Learning for Visual Perception
Abstract: Unsupervised domain adaptation for point cloud semantic segmentation has attracted great attention due to its effectiveness in learning with unlabeled data. Most of existing methods use global-level feature alignment to transfer the knowledge from the source domain to the target domain, which may cause the semantic ambiguity of the feature space. In this paper, we propose a graph-based framework to explore the local-level feature alignment between the two domains, which can reserve semantic discrimination during adaptation. Specifically, in order to extract local-level features, we dynamically construct local feature graphs on both of the two domains and then build a memory bank with the graphs from the source domain. In particular, we use optimal transport to generate the graph matching pairs. Then, based on the assignment matrix, we can meticulously align the feature distributions between the two domains by the graph-based local feature loss. Furthermore, we consider the correlation between the features of different categories and design a category-guided contrastive loss to guide the segmentation model to learn discriminative features on the target domain. Extensive experiments on different synthetic-to-real and real-to-real domain adaptation scenarios demonstrate that our method can achieve state-of-the-art performance.

Keywords: Semantic Scene Understanding, Localization, Intelligent Transportation Systems
Abstract: Place recognition is seen as a crucial factor to correct cumulative errors in Simultaneous Localization and Mapping (SLAM) applications. Most existing studies focus on visual place recognition, which is inherently sensitive to environmental changes such as illumination, weather and seasons. Considering these facts, more recent attention has been attracted to use 3-D Light Detection and Ranging (LiDAR) scans for place recognition, which demonstrates more credibility by exerting accurate geometric information. Different from pure geometric-based studies, this paper proposes a novel global descriptor, named SectionKey, which leverages both semantic and geometric information to tackle the problem of place recognition in large-scale urban environments. The proposed descriptor is robust and invariant to viewpoint changes. Specifically, the encoded three-layers key serves as a pre-selection step and a `candidate center' selection strategy is deployed before calculating the similarity score, thus improving the accuracy and efficiency significantly. Then, a two-step semantic iterative closest point (ICP) algorithm is applied to acquire the 3-D pose (x, y, theta) that is used to align the candidate point clouds with the query frame and calculate the similarity score. Extensive experiments have been conducted on public Semantic KITTI dataset to demonstrate the superior performance of our proposed system over state-of-the-art baselines.

Keywords: Semantic Scene Understanding
Abstract: Fisheye object detection is a difficult task in robotics and autonomous driving. One of the reasons is that the fisheye datasets are inferior to standard image datasets in scale and quantity, which inspires the idea of using standard image datasets for fisheye object detection. However, the models trained on standard image datasets do not perform well with fisheye data. In this work, we explore the effect of fisheye images on different stages of the YOLOX with published weights generated by standard image datasets. We also propose a new regression strategy for 24-points object representation method, which is insensitive to image distortion. The experiments show that the feature extraction part is robust to fisheye image features, while the regression part of location and category performs poorly. The strategy can achieve the position of discrete points without calculating the IOU of irregular-shaped boxes. Theoretically, the strategy can be widely adopted to regress the irregular bounding boxes composed of discrete points. Source code is at https://github.com/IN2-ViAUn/Exploration-of-Potential

Keywords: Semantic Scene Understanding, RGB-D Perception, Motion and Path Planning
Abstract: Disinfection robots have applications in promoting public health and reducing hospital-acquired infections and have drawn considerable interest due to the COVID-19 pandemic. To disinfect a room quickly, motion planning can be used to plan robot disinfection trajectories on a reconstructed 3D map of the room's surfaces. However, existing approaches discard semantic information of the room and, thus, take a long time to perform thorough disinfection. Human cleaners, on the other hand, disinfect rooms more efficiently by prioritizing the cleaning of high-touch surfaces. To address this gap, we present a novel GPU-based volumetric semantic TSDF (Truncated Signed Distance Function) integration system for semantic 3D reconstruction. Our system produces 3D reconstructions that distinguish high-touch surfaces from non-high-touch surfaces at approximately 50 frames per second on a consumer-grade GPU, which is approximately 5 times faster than existing CPU-based TSDF semantic reconstruction methods. In addition, we extend a UV disinfection motion planning algorithm to incorporate semantic awareness for optimizing coverage of disinfection trajectories. Experiments show that our semantic-aware planning outperforms geometry-only planning by disinfecting up to 20% more high-touch surfaces under the same time budget. Further, the real-time nature of our semantic reconstruction pipeline enables future work on simultaneous disinfection and mapping.

Keywords: Semantic Scene Understanding, Perception for Grasping and Manipulation, Task Planning
Abstract: Assistive robot systems have been developed to help people accomplish daily manipulation tasks especially for those with disabilities, where scene understanding plays a crucial role in enabling robots to interpret the surroundings and behave accordingly. Most of the current systems approach scene understanding without considering the functional dependencies between objects. However, it is only valuable to interact with some objects when their function-relevant counterparts are considered. In this paper, we augment an assistive robotic arm system with an end-to-end semantic relationship reasoning model. It incorporates functional relationships between pairs of objects for semantic scene understanding. To ensure good generalization to unseen objects and relationships, the model works in a category-agnostic manner. We evaluate our design and three baseline methods on a self-collected benchmark with two levels of difficulty. To further demonstrate the effectiveness, the model is integrated with a symbolic planner for goal-oriented, multi-step manipulation task on a real-world assistive robotic arm platform.

Keywords: Multi-Robot Systems, Task Planning, Mapping
Abstract: This paper presents a novel approach for multi-robot unknown area exploration. Recently, the frontier tree data structure was used in single robot exploration to memorize frontiers, their positions, exploration state, and the map. This tree could be queried to decide on further exploration steps. In this paper, we take the concept further for multi-robot exploration by proposing a new abstraction called the ‘group,’ meant to share information through a common frontier tree, requisite operations at the group level, and a method to assign goals to multiple robots. A group is a set of robots, the union of whose explored regions forms a contiguous region (a single connected region in a topological sense). As a group has precisely one tree, the robots share a common state of the exploration task. We propose techniques to merge groups and their frontier trees once their maps overlap. Finally, we suggest a method to designate and assign exploration goals to the individual robots by choosing nodes from the frontier tree. The proposed approach outperforms seven state-of-the-art research works in simulation.

Keywords: Multi-Robot Systems, Path Planning for Multiple Mobile Robots or Agents, Surveillance Robotic Systems
Abstract: We consider a multi-robot patrolling scenario with intermittent connectivity constraints, which ensure that data from the robots finally arrive at a base station. In particular, each robot traverses a closed tour periodically and meets with the robots on neighboring tours to exchange data. We model the problem as a variant of the min-max vertex cycle cover problem (MMCCP), which is the problem of covering all vertices with a given number of disjoint tours such that the largest tour length is minimal. In this work we introduce the minimum idleness connectivity-constrained multi-robot patrolling problem, show that it is NP-hard, and model it as a mixed integer linear program (MILP). The computational complexity of exactly solving this problem restrains practical applications, and therefore we develop approximate algorithms taking a solution for MMCCP as input. Our simulation experiments on 10 vertices and up to 3 robots compare the results of different solution approaches (including solving the MILP formulation) and show that our greedy algorithm can obtain an objective value close to the one of the MILP formulations with much shorter computation times. Experiments on instances with up to 100 vertices and up to 10 robots indicate that the greedy approximation algorithm tries to keep the length of the longest tour small by extending smaller tours for data exchange.

Keywords: Multi-Robot Systems, Path Planning for Multiple Mobile Robots or Agents, Swarm Robotics
Abstract: We present a manifold optimization approach to solve inference and planning problems with range constraints. The core of our approach is the definition of a manifold that represents points or poses with range constraints. We discover that the manifold of range-constrained points is homogeneous under the rigid transformation group action, and utilize the group action to derive the tangent space, retraction and topology of the manifold. We evaluate the performance of manifold optimization approach on solving range-constrained inference problems over state-of-the-art constrained optimization methods, and the results show that manifold optimization with the range-constraint manifold achieves both faster speed and better constraint satisfaction. We further study the conditions of inference problems that we can treat range measurements as constraints in practice.

Keywords: Multi-Robot Systems, Sensor Networks, Distributed Robot Systems
Abstract: This paper presents a coverage based control algorithm to coordinate a group of autonomous robots. Most of the solutions presented in the literature rely on an exact Voronoi partitioning, whose computation requires complete knowledge of the environment to be covered. This can be achieved only by robots with unlimited sensing capabilities, or through communication among robots in a limited sensing scenario. To overcome these limitations, we present a distributed control strategy to cover an unknown environment with a group of robots with limited sensing capabilities and in the absence of reliable communication. The control law is based on a limited Voronoi partitioning of the sensing area, and we demonstrate that the group of robots can optimally cover the environment using only information that is locally detected (without communication). The proposed method is validated by means of simulations and experiments carried out on a group of mobile robots.

Keywords: Multi-Robot Systems, Path Planning for Multiple Mobile Robots or Agents, Task and Motion Planning
Abstract: In this work, we consider the Multi-Agent Pickup and Delivery (MAPD) problem, where agents constantly engage with new tasks and need to plan collision-free paths to execute them. To execute a task, an agent needs to visit a pair of goal locations, consisting of a pickup location and a delivery location. We propose two variants of an algorithm that assigns a sequence of tasks to each agent using the anytime algorithm Large Neighborhood Search (LNS) and plans paths using the Multi-Agent Path Finding (MAPF) algorithm Priority-Based Search (PBS). LNS-PBS is complete for well-formed MAPD instances, a realistic subclass of MAPD instances, and empirically more effective than the existing complete MAPD algorithm CENTRAL. LNS-wPBS provides no completeness guarantee but is empirically more efficient and stable than LNS-PBS. It scales to thousands of agents and thousands of tasks in a large warehouse and is empirically more effective than the existing scalable MAPD algorithm HBH+MLA*. LNS-PBS and LNS-wPBS also apply to a more general variant of MAPD, namely the Multi-Goal MAPD (MG-MAPD) problem, where tasks can have different numbers of goal locations.

Keywords: Multi-Robot Systems, Path Planning for Multiple Mobile Robots or Agents, Networked Robots
Abstract: We present a novel overconstraining and constraint-discarding method for asynchronous, real-time, decentralized, multi-robot trajectory planning that ensures collision avoidance. Our approach utilizes communication between robots. The communication medium is best-effort: messages may be dropped, re-ordered or delayed. Robots conservatively constrain themselves against others assuming they may be working with outdated information, and discard constraints when they receive update messages from others. Our method can augment existing synchronized decentralized receding horizon planning algorithms that utilize separating hyperplanes for collision avoidance thereby making them applicable to asynchronous setups. As an example, we extend an existing model predictive control based, synchronized, decentralized multi-robot planner using our method. We show our method’s effectiveness under asynchronous planning and imperfect communication by comparing our extension to the base version. Our extension does not result in any collisions or synchronization-induced deadlocks to which the base version is prone.

Keywords: Multi-Robot Systems, Path Planning for Multiple Mobile Robots or Agents, Optimization and Optimal Control
Abstract: This paper presents an algorithm for a team of mobile robots to simultaneously learn a spatial field over a domain and spatially distribute themselves to optimally cover it. Drawing from previous approaches that estimate the spatial field through a centralized Gaussian process, this work leverages the spatial structure of the coverage problem and presents a decentralized strategy where samples are aggregated locally by establishing communications through the boundaries of a Voronoi partition. We present an algorithm whereby each robot runs a local Gaussian process calculated from its own measurements and those provided by its Voronoi neighbors, which are incorporated into the individual robot’s Gaussian process only if they provide sufficiently novel information. The performance of the algorithm is evaluated in simulation and compared with centralized approaches.

Keywords: Multi-Robot Systems, Path Planning for Multiple Mobile Robots or Agents, Motion and Path Planning
Abstract: We present a centralized algorithm for labeled, disk-shaped Multi-Agent Path Planning (MPP) in a continuous workspace with polygonal boundaries. Our method automatically constructs a discrete pebble-graph out of the workspace and then routes the agents on the graph. To construct the pebble graph,

we identify inscribed circles in the workspace via medial axis transform and organize agents into layers within each inscribed circle. We show that our layered pebble-graph allows the agents to perform both pebble and rotation motions, such that all the MPP instances restricted to the pebble-graph are feasible. MPP instances with continuous start and goal positions can then be solved via local navigations that route agents from and to graph vertices. We experiment with our method on a row of 5 environments with a high agent-packing density (up to 61:6% of the free space). Such density violates the well-separated assumptions made by state-of-the-art MPP planners, while our method achieves a high success rate.

Keywords: Multi-Robot Systems, Task Planning, Human Factors and Human-in-the-Loop
Abstract: Leveraging both the autonomy of robots and the expert knowledge of humans can enable a multi-robot system to complete missions in challenging environments with a high degree of adaptivity and robustness. This paper proposes a multimodal task-based graphical user interface for controlling a heterogeneous multi-robot team. The core of the interface is an integrated multi-robot task allocation system to allow the user to encode his/her intents to guide the heterogeneous multi-robot team. The design of the interface aims to provide the human operator continuous situational awareness and effective control for rapid decision-making in time-critical missions. Team CSIRO Data61 came in second place utilizing this interface for the DARPA Subterranean (SubT) Challenge. The ideas used for this user interface can apply to other multi-robot applications.

Keywords: Soft Sensors and Actuators, AI-Based Methods, Soft Robot Materials and Design
Abstract: Robots using classical control have revolutionised assembly lines where the environment and manipulated objects are restricted and predictable. However, they have proven less effective when the manipulated objects are deformable due to their complex and unpredictable behaviour. The use of tactile sensors and continuous monitoring of tactile feedback is therefore particularly important for pick-and-place tasks using these materials. This is in part due to the need to use multiple points of contact for the manipulation of deformable objects which can result in slippage with inadequate coordination between manipulators. In this paper, continuous monitoring of tactile feedback, using a liquid metal soft force sensor, for grasping deformable objects is presented. The trained data-driven model distinguishes between successful grasps, slippage and failure during a manipulation task for multiple deformable objects. Slippage could be anticipated before failure occurred using data acquired over a 30 ms period with a greater than 95% accuracy using a random forest classifier. The results were achieved using a single sensor that can be mounted on the fingertips of existing grippers and contributes to the development of an automated pick-and-place process for deformable objects.

Keywords: Soft Sensors and Actuators, Soft Robot Applications
Abstract: This research focuses on soft robotic bladders that are used to monitor and control the interaction between a user’s head and the shell of a Smart Helmet. Compression of these bladders determines impact dissipation; hence the focus of this paper is sensing and estimation of bladder compression. An IR rangefinder-based solution is evaluated using regression techniques as well as a Neural Network to estimate bladder compression. A Hall-Effect (HE) magnetic sensing system is also examined where HE sensors embedded in the base of the bladder sense the position of a magnet in the top of the bladder. The paper presents the HE sensor array, signal processing of HE voltage data, and then a Neural Network (NN) for predicting bladder compression. Efficacy of different training data sets on NN performance is studied. Different NN configurations are examined to determine a configuration that provides accurate estimates with as few nodes as possible. Different bladder compression profiles are evaluated to characterize IR range finding and HE based techniques in application scenarios.

Keywords: Soft Sensors and Actuators, Soft Robot Materials and Design
Abstract: Whiskers are widely used by animals for sensing physical interactions with their environments. By combining the Kirigami skin pop-up feature and flexible conducting layer, we designed a deployable Kirigami whisker sensor. The sensor can deploy from a flat state to a sensing state while whisker stiffness and initial pop-up angle can be tuned by adjusting the pre-stretch strain. Preliminary results show that the sensor works well both in air and underwater. The sensor is capable of measuring both externally applied forces and water flow.

Keywords: Soft Sensors and Actuators, Modeling, Control, and Learning for Soft Robots, In-Hand Manipulation
Abstract: Soft sensorised skins are essential for improving robotic manipulation capabilities towards that of humans. Integration of sensors into existing robotic hands is challenging due to rigidity of components, low packing density or poor sensor response. We propose a sensorised skin, based-on barometric sensing, which can be molded over a skeletal robot hand. The sensors connect air chambers embedded in the soft skin to wrist-mounted pressure sensors, allowing sensor spacing 2-4 mm, force ranges from 23 mN to 5700 mN and bandwidth of 20 Hz. Integrating this with a skeletal hand allows us to showcase the potential of these sensors to aid robotic manipulation. We demonstrate 3-axis contact modelling, useful for in-hand manipulation and exploration. In addition, by grasping a chopstick and sensing forces transmitted from the environment, the system can remotely detect small environmental features, e.g., hole finding using tools.

Keywords: Soft Sensors and Actuators, Force and Tactile Sensing
Abstract: We create a virtual 2D tactile array for soft pneumatic actuators using embedded audio components. We detect contact-specific changes in sound modulation to infer tactile information. We evaluate different sound representations and learning methods to detect even small contact variations. We demonstrate the acoustic tactile sensor array by the example of a PneuFlex actuator and use a Braille display to individually control the contact of 29 x 4 pins with the actuator's 90 x 10 mm palmar surface. Evaluating the spatial resolution, the acoustic sensor localizes edges in x- and y-direction with a root-mean-square regression error of 1.67 mm and 0.0 mm, respectively. Even light contacts of a single Braille pin with a lifting force of 0.17 N are measured with high accuracy. Finally, we demonstrate the sensor's sensitivity to complex contact shapes by successfully reading the 26 letters of the Braille alphabet from a single display cell with a classification rate of 88%.

Keywords: Soft Sensors and Actuators, Soft Robot Applications
Abstract: A soft pneumatic actuator (SPA) is one of the most prominent components in a soft robotic system. The sensing of SPAs is challenging owing to their elasticity and deformability. Sensors for specific factors are required to successfully sense an SPA. A flexible sensor is an important component for sensing the conditions of SPAs, such as the shape and deformation due to contact events during applications. Developing versatile sensors with high flexibility and tolerability for SPAs is challenging. Data-driven sensing approaches involve individual machine learning models for different actuators. In contrast, it is an enormous advantage to have versatile sensors as hardware and machine learning models as software employed on various actuators. Therefore, we propose a transferable shape estimation method based on active vibro-acoustic sensing to achieve tolerability and increase versatility. We created easily transferable sensing devices for SPAs. In addition, we employed a data-driven approach that utilizes a simple transfer learning technique on a two-dimensional convolutional neural network model. We confirm the feasibility and versatility of the proposed method through evaluation experiments. A transferable estimation method was used on SPAs to estimate the bending angle and length under various sensing and environmental conditions, and the average errors were less than 3.5 degrees and 2.1 mm, respectively.

Keywords: Soft Sensors and Actuators, Soft Robot Applications, Modeling, Control, and Learning for Soft Robots
Abstract: In recent times, soft manipulators have gardened immense interest given their dexterous abilities. A critical aspect for their feedback control involves the reconstruction of the manipulator shape. The research, for the first time, presents shape reconstruction of a soft manipulator through sensor fusion of information available from Inertial Measurement Units (IMUs) and visual tracking. The manipulator is modeled using multi-segment continuous curvature Pythagorean Hodograph (PH) curves. PH curves are a class of continuous curvature curves with an analytical expression for the hodograph (slope). The shape reconstruction is formulated as an optimization problem that minimizes bending energy of the curve with a length constraint and the information from IMUs and/or visual markers. The paper experimentally investigates the robustness of shape reconstruction for scenarios when position of all visual markers, or slope at all the knots (placement of sensors) are known. Occlusion of manipulator segments is frequent, hence, this scenario is simulated by fusing information of available slopes (IMUs) at all knots and position (vision) at some knots. The experiments are performed on a planar tensegrity manipulator with IMUs feedback and visual tracking. The robustness study indicates reliability of these models for real world applications. Additionally, the proposed sensor fusion algorithm provides promising results where, for most cases, the shape estimates benefit from additional position information. Finally, the low dimensionality of the optimization problem argues for extension of the approach for real-time applications.

Keywords: Soft Sensors and Actuators, Soft Robot Applications, Sensor Fusion
Abstract: In this paper, we propose a novel variable-length estimation approach for shape sensing of extensible soft robots utilizing fiber Bragg gratings (FBGs). Shape reconstruction from FBG sensors has been increasingly developed for soft robots, while the narrow stretching range of FBG fiber makes it difficult to acquire accurate sensing results for extensible robots. Towards this limitation, we newly introduce an FBG-based length sensor by leveraging a rigid curved channel, through which FBGs are allowed to slide within the robot following its body extension/compression, hence we can search and match the FBGs with specific constant curvature in the fiber to determine the effective length. From the fusion with the above measurements, a model-free filtering technique is accordingly presented for simultaneous calibration of a variable-length model and temporally continuous length estimation of the robot, enabling its accurate shape sensing using solely FBGs. The performances of the proposed method have been experimentally evaluated on an extensible soft robot equipped with an FBG fiber in both free and unstructured environments. The results concerning dynamic accuracy and robustness of length estimation and shape sensing demonstrate the effectiveness of our approach.

Keywords: Soft Sensors and Actuators, Soft Robot Materials and Design, Soft Robot Applications
Abstract: Underwater robots have a variety of potential uses, including marine resource research, ecological research, and disaster relief. Most of the underwater robots currently in practical use have screw propulsion systems, which have several noises, collision, and entrainment problems. There is a lot of research on underwater robots using soft actuators to solve these problems. However, current soft actuators have disadvantages, such as the need for special fluids, pressure sources, and high voltage circuits. Therefore, we have developed a soft-skin actuator based on magnetohydrodynamics (MHD). The soft-skin MHD actuator is made of soft material and the structure is prepared as thin, which allows it to attach to the surface of an object, including curved surfaces, to provide the object with a propulsive function in the sea. Since it has no moving parts, it does not generate mechanical noise, and there is no danger of entrapment. Because it can pump seawater directly, it does not require a special working fluid, and its structure is simple and easy to miniaturize. This paper investigates the thrust and power consumption of the developed soft-skin MHD actuator when attached to a flat surface. As a result, we obtained a thrust of 1.37 mN from a single soft-skin MHD actuator with a maximum power of about 140 W. We also measured the thrust force by attaching it to a curved surface. We obtained a higher thrust on a curved surface by adjusting the crossing of the magnetic field and the current than when using a flat surface. We developed an untethered robot that can remove oil from the sea using soft-skin MHD actuators. We demonstrated the adaptability of the soft-skin MHD actuator by attaching it to a commercial underwater camera weighing about 253.5 g and providing propulsion.

Keywords: Path Planning for Multiple Mobile Robots or Agents, Networked Robots
Abstract: Multi agent coverage and robot navigation are two very important research fields within robotics. However, their intersection has received limited attention. In multi agent coverage, perfect navigation is often assumed, and in robot navigation, the focus is often to minimize the localization error with the aid of stationary features from the environment. The need for integration of the two becomes clear in environments with very sparse features or landmarks, for example when a group of Autonomous Underwater Vehicles (AUVs) are to search a uniform seafloor for mines or other dangerous objects. In such environments, localization systems are often deprived of detectable features to use that could increase their accuracy. In this paper we propose an algorithm for doing navigation aware multi agent coverage in areas with no landmarks. Instead of using identical lawn mower patterns, we propose to mirror every other pattern to enable the agents to meet up and make inter-agent measurements and share information regularly. This improves performance in two ways, global drift in relation to the area to be covered is reduced, and local coverage gaps between adjacent patterns are reduced. Further, we show that this can be accomplished within the constraints of very limited sensing, computing and communication resources that most AUVs have available. The effectiveness of our method is shown through statistically significant simulated experiments.

Keywords: Path Planning for Multiple Mobile Robots or Agents, Multi-Robot Systems, Planning, Scheduling and Coordination
Abstract: For enabling efficient, large-scale coordination of unmanned aerial vehicles (UAVs) under the labeled setting, in this work, we develop the first polynomial time algorithm for the reconfiguration of many moving bodies in three- dimensional spaces, with provable 1.x asymptotic makespan optimality guarantee under high robot density. More precisely, on an m1×m2×m3 gird, m1>= m2>= m3, our method computes solutions for routing up to m1*m2*m3/3 uniquely labeled robots with uniformly randomly distributed start and goal configurations within a makespan of m1+2m2+2m3+o(m1), with high probability. Because the makespan lower bound for such instances is m1+m2+m3-o(m1), also with high probability, as m1 -> infty, (m1+2m2 +2m3)/(m1+m2+m3) optimality guarantee is achieved. (m1 +2m2+2m3)/(m1+m2+m3)in (1,5/3 ],yielding 1.x optimality.In contrast, it is well-known that multi-robot path planning is NP-hard to optimally solve. In numerical evaluations, our method readily scales to support the motion planning of over 100 000 robots in 3D while simultaneously achieving 1.x optimality. We demonstrate the application of our method in coordinating many quadcopters in both simulation and hardware experiments.

Keywords: Task and Motion Planning, Planning, Scheduling and Coordination, Environment Monitoring and Management
Abstract: In many environmental monitoring applications robots are often tasked to visit various distinct locations to make observations and/or collect specific measurements. The problem of scheduling and assigning robots to the various tasks and planning feasible paths for the robots can be posed as an Orienteering Problem (OP). In the standard OP, routing and scheduling is achieved by maximizing an objective function by visiting the most rewarding locations while respecting a limited travel budget. However, traditional formulations for such problems usually neglect some environmental features that can greatly impact the tour, e.g., flows, such as wind or ocean currents. This is of particular importance for applications in marine and atmospheric environments where vehicle motions can be significantly impacted by the environmental dynamics and the environment exerts a non-negligible force on the vehicles. In this paper, we tackle the OP in fluid environments where robots must operate in the presence of ocean and/or atmospheric currents. We introduce a novel multi-objective formulation that combines both task and path planning problems, and whose goals are to (i) maximize the collected reward, while (ii) minimizing the energy expenditure by leveraging the environmental dynamics wherever possible. We validate our strategy using simulated ocean model data to show that our approach can generate a diverse set of solutions that have an adequate compromise between both objectives.

Keywords: Path Planning for Multiple Mobile Robots or Agents, Multi-Robot Systems, Deep Learning Methods
Abstract: Portfolio-based algorithm selection can help in choosing the best suited algorithm for a given task while leveraging the complementary strengths of the candidates. Solving the Multi-Agent Path Finding (MAPF) problem optimally has been proven to be NP-Hard. Furthermore, no single optimal algorithm has been shown to have the fastest runtime for all MAPF problem instances, and there are no proven approaches for when to use each algorithm. To address these challenges, we develop MAPFASTER, a smaller and more accurate deep learning based architecture aiming to be deployed in fleet management systems to select the fastest MAPF solver in a multi-robot setting. MAPF problem instances are encoded as images and passed to the model for classification into one of the portfolio's candidates. We evaluate our model against state-of-the-art Optimal-MAPF-Algorithm selectors, showing +5.42% improvement in accuracy while being 7.1times faster to train. The dataset, code and analysis used in this research can be found at href{https://github.com/jeanmarcalkazzi/mapfaster}{https://github.com/jeanmarcalkazzi/mapfaster}.

Keywords: Task and Motion Planning, Task Planning, Manipulation Planning
Abstract: Robotic planning in real-world scenarios typically requires joint optimization of logic and continuous variables. A core challenge to combine the strengths of logic planners and continuous solvers is the design of an efficient interface that informs the logical search about continuous infeasibilities. In this paper we present a novel iterative algorithm that connects logic planning with nonlinear optimization through a bidirectional interface, achieved by the detection of minimal subsets of nonlinear constraints that are infeasible. The algorithm continuously builds a database of graphs that represent (in)feasible subsets of continuous variables and constraints, and encodes this knowledge in the logical description. As a foundation for this algorithm, we introduce Planning with Nonlinear Transition Constraints (PNTC), a novel planning formulation that clarifies the exact assumptions our algorithm requires and can be applied to model Task and Motion Planning (TAMP) efficiently. Our experimental results show that our framework significantly outperforms alternative optimization-based approaches for TAMP. Webpage: https://quimortiz.github.io/graphnlp/

Keywords: Path Planning for Multiple Mobile Robots or Agents, Multi-Robot Systems, Swarm Robotics
Abstract: Online coverage path planning to explore an unknown workspace with multiple homogeneous robots could be either centralized or distributed. While distributed planners are computationally faster, centralized planners can produce more efficient paths, reducing the duration of completing a coverage mission significantly. To exploit the power of a centralized framework, we propose a receding horizon centralized online multi-robot planner. In each planning horizon, it generates collision-free paths that guide the robots to visit some obstacle-free locations (aka goals) not visited so far, which in turn help them explore some new regions with their laser rangefinders. We formally prove that, under reasonable conditions, it enables the robots to cover a workspace completely and subsequently analyze its time complexity. We evaluate our planner for ground and aerial robots by performing experiments with up to 128 robots on six 2D grid-based benchmark obstacle maps, establishing scalability. We also perform Gazebo simulations with 10 quadcopters and real experiments with 2 four-wheel ground robots, demonstrating its practical feasibility. Furthermore, a comparison with a state-of-the-art distributed planner establishes its superiority in coverage completion time.

Keywords: Path Planning for Multiple Mobile Robots or Agents, Integrated Planning and Control, Multi-Robot Systems
Abstract: This paper deals with Multi-robot Trajectory Planning, that is, the problem of computing trajectories for multiple robots navigating in a shared space. Approaches based on trajectory optimization can solve this problem optimally. However, such methods are hampered by complex robot dynamics and collision constraints that couple robot's decision variables. We propose a distributed multi-robot optimization algorithm (DiMOpt) which addresses these issues by exploiting (1) consensus optimization strategies to tackle coupling collision constraints, and (2) a single-robot sequential convex programming (SCP) method for efficiently handling non-convexities introduced by dynamics. We compare DiMOpt with a baseline sequential convex programming algorithm tailored to the multi-robot case (M-SCP). We empirically demonstrate that DiMOpt scales well for large fleets of robots, while computing solutions faster and with lower costs than M-SCP. Moreover, DiMOpt is an iterative algorithm that finds feasible trajectories before converging to an optimal solution, and results suggest the quality of such fast initial solutions is comparable to a converged solution computed via M-SCP. Finally, we also investigate how other factors, including path length, affect the performance of DiMOpt.

Keywords: Path Planning for Multiple Mobile Robots or Agents, Multi-Robot Systems, Planning under Uncertainty
Abstract: We provide theoretical bounds on the worst case performance of the greedy algorithm in seeking to maximize a normalized, monotone, but not necessarily submodular objective function under a simple partition matroid constraint. We also provide worst case bounds on the performance of the greedy algorithm in the case that limited information is available at each planning step. We specifically consider limited information as a result of unreliable communications during distributed execution of the greedy algorithm. We utilize notions of curvature for normalized, monotone set functions to develop the bounds provided in this work. To demonstrate the value of the bounds provided in this work, we analyze a variant of the benefit of search objective function and show, using real-world data collected by an autonomous underwater vehicle, that theoretical approximation guarantees are achieved despite non-submodularity of the objective function.

Keywords: Path Planning for Multiple Mobile Robots or Agents, Reinforcement Learning, Automation at Micro-Nano Scales
Abstract: For biomedical applications in targeted therapy delivery and interventions, a large swarm of micro-scale particles ("agents") has to be moved through a maze-like environment ("vascular system") to a target region ("tumor"). Due to limited on-board capabilities, these agents cannot move autonomously; instead, they are controlled by an external global force that acts uniformly on all particles. In this work, we demonstrate how to use a time-varying magnetic field to gather particles to a desired location. We use reinforcement learning to train networks to efficiently gather particles. Methods to overcome the simulation-to-reality gap are explained, and the trained networks are deployed on a set of mazes and goal locations. The hardware experiments demonstrate fast convergence, and robustness to both sensor and actuation noise. To encourage extensions and to serve as a benchmark for the reinforcement learning community, the code is available at Github.

Keywords: Incremental Learning, Transfer Learning, Recognition
Abstract: The ability to evolve is fundamental for any valuable autonomous agent whose knowledge cannot remain limited to that injected by the manufacturer. Consider for example a home assistant robot: it should be able to incrementally learn new object categories when requested, but also to recognize the same objects in different environments (rooms) and poses (hand-held/on the floor/above furniture), while rejecting unknown ones. Despite its importance, this scenario has started to raise interest in the robotic community only recently and the related research is still in its infancy, with existing experimental testbeds but no tailored methods. With this work, we propose the first learning approach that deals with all the previously mentioned challenges at once by exploiting a single contrastive objective. We show how it learns a feature space perfectly suitable to incrementally include new classes and is able to capture knowledge which generalizes across a variety of visual domains. Our method is endowed with a tailored effective stopping criterion for each learning episode and exploits a self-paced thresholding strategy that provides the classifier with a reliable rejection option. Both these novel contributions are based on the observation of the data statistics and do not need manual tuning. An extensive experimental analysis confirms the effectiveness of the proposed approach in establishing the new state-of-the-art. The code is available at https://github.com/FrancescoCappio/Contrastive_Open_World.

Keywords: Machine Learning for Robot Control, Reinforcement Learning, Transfer Learning
Abstract: In order to be effective general purpose machines in real world environments, robots not only will need to adapt their existing manipulation skills to new circumstances, they will need to acquire entirely new skills on-the-fly. One approach to achieving this capability is via Multi-task Reinforcement Learning (MTRL). Most recent work in MTRL trains a single policy to solve all tasks at once. In this work, we investigate the feasibility of instead training separate policies for each task, and only transferring from a task once the policy for it has finished training. We describe a method of finding near optimal sequences of transfers to perform in this setting, and use it to show that performing the optimal sequence of transfer is competitive with other MTRL methods on the MetaWorld MT10 benchmark. Lastly, we describe a method for finding nearly optimal transfer sequences during training that is able to improve on training each task from scratch.

Keywords: Reinforcement Learning, Transfer Learning, Machine Learning for Robot Control
Abstract: In manufacturing, assembly tasks have been a challenge for learning algorithms due to variant dynamics of different environments. Reinforcement learning (RL) is a promising framework to automatically learn these tasks, yet it is still not easy to apply a learned policy or skill, that is the ability of solving a task, to a similar environment even if the deployment condition is only slightly different. For safety and feasibility reasons, state-of-the-art methods require policy training in simulation to prevent undesired behavior followed by domain transfer, or guided policy search for similar environments. In this paper, we address the challenge of transferring knowledge within a family of similar tasks by leveraging multiple skill priors. We propose to learn prior distribution over the specific skill required to accomplish each task and compose the family of skill priors to guide learning the policy for a new task by comparing the similarity between the target task and the prior ones. Our method learns a latent action space representing the skill embedding from demonstrated trajectories for each prior task. We have evaluated our method on a task in simulation and a set of peg-in-hole insertion tasks and demonstrate better generalization to new tasks that have never been encountered during training. Our Multi-Prior Regularized RL (MPR-RL) method is deployed directly on a real-world Franka Panda arm, requiring only a set of demonstration trajectories from similar, but cruicially not identical, problem instances.

Keywords: Reinforcement Learning, Machine Learning for Robot Control, Transfer Learning
Abstract: Current reinforcement learning (RL) in robotics often experiences difficulty in generalizing to new downstream tasks due to the innate task-specific training paradigm. To alleviate it, unsupervised RL, a framework that pre-trains the agent in a task-agnostic manner without access to the task-specific reward, leverages active exploration for distilling diverse experience into essential skills or reusable knowledge. For exploiting such benefits also in robotic manipulation, we propose an unsupervised method for transferable manipulation skill discovery that ties structured exploration toward interacting behavior and transferable skill learning. It not only enables the agent to learn interaction behavior, the key aspect of the robotic manipulation learning, without access to the environment reward, but also to generalize to arbitrary downstream manipulation tasks with the learned task-agnostic skills. Through comparative experiments, we show that our approach achieves the most diverse interacting behavior and significantly improves sample efficiency in downstream tasks including the extension to multi-object, multitask problems.

Keywords: Transfer Learning, Object Detection, Segmentation and Categorization, Deep Learning for Visual Perception
Abstract: Autonomous agents need to perceive the world in a robust way, such that the shift in data distribution does not lead to faulty perception results. When agents cannot be trained with abundant data, agents may need to operate on real world environments while trained on simulated data, and suffer from domain shift. This paper proposes an effective and robust unsupervised domain adaptation (UDA) method that can resolve these situations. In the UDA setup, we are given a labeled source domain and an unlabeled target domain that share the same set of classes but are sampled from different distributions. This domain shift prevents agents which employ deep neural networks from generalizing well on the target domain. Recent methods adopt the strategy of self-training the networks with pseudo labeled target samples. However, falsely labeled samples cause negative transfer and deteriorate generalization of a network. To reduce negative transfer we propose an algorithm that can filter the pseudo labels, and use the filtered labels to align the domains in the feature space. The samples whose labels have not passed the filtering process can be used as an index to tune the hyperparameters of our method. Across various benchmarks, we validate the performance of our method. Especially, our method achieves strong performance on the synthetic-to-real adaptation scenario.

Keywords: Transfer Learning, Machine Learning for Robot Control, Reinforcement Learning
Abstract: In this work, we propose a data-driven approach to optimize the parameters of a simulation such that control policies can be directly transferred from simulation to a real-world quadrotor. Our neural network-based policies take only onboard sensor data as input and run entirely on the embedded hardware. In real-world experiments, we compare low-level Pulse-Width Modulated control with higher-level control structures such as Attitude Rate and Attitude, which utilize Proportional-Integral-Derivative controllers to output motor commands. Our experiments show that low-level controllers trained with Reinforcement Learning require a more accurate simulation than higher-level control policies at the expense of being less robust towards parameter uncertainties.

Keywords: Transfer Learning, Semantic Scene Understanding, Computer Vision for Transportation
Abstract: Unsupervised domain adaptation (UDA) aims to learn domain-invariant representations between the labeled source domain and the unlabeled target domain. Existing self-training-based UDA methods use ground truth and pseudo-labels to supervise source data and target data respectively. However, strong supervision in the source domain and pseudo-label noise in the target domain lead to some problems, such as biased predictions and over-fitting. To tackle these issues, we propose a novel Bilateral Knowledge Distillation (BKD) framework for UDA in semantic segmentation, which adopts different knowledge distillation strategies depending on the domain. Specifically, we first introduce a Source-Flow Distillation (SD) to smooth the labels of source images, which weakens the supervision in the source domain. Meanwhile, a Target-Flow Distillation (TD) is designed to extract the inter-class knowledge in the probability map output from the teacher model, which alleviates the influence of pseudo-label noise in the target domain. Considering the class imbalance in semantic segmentation, we further propose an Image-Wise Hard Pixel Mining (HPM) to address this issue without estimating class frequency in the unlabeled target domain. The effectiveness of our framework against existing state-of-the-art methods is demonstrated by extensive experiments on two benchmarks: GTA5-to-Cityscapes and SYNTHIA-to-Cityscapes.

Keywords: Transfer Learning, Recognition, Deep Learning Methods
Abstract: Domain adaptation is an important property in robot vision, which enables the neural networks pre-trained on source domains to adapt target domains automatically without any annotation efforts. During this process, source data is not always accessible due to the constraints of expensive storage overhead and data privacy protection. Therefore, the source domain pre-trained model is expected to optimize with only unlabeled target data, termed as source-free unsupervised domain adaptation. In this paper, we view this problem as a special case of noisy label learning, since the given pre-trained model can generate noisy labels for unlabeled target data via network inference. The potential semantic cues for unsupervised domain adaptation exactly lie on these noisy labels. Inspired by this problem modeling, we propose a simple yet effective Self-Supervised Noisy Label Learning method, which injects self-supervised learning to impose the intrinsic data structure and facilitate label-denoising. Extensive experiments have been conducted on diverse benchmarks to validate the effectiveness. Our method achieves state-of-the-art performance.

Keywords: Transfer Learning, Incremental Learning, Deep Learning in Grasping and Manipulation
Abstract: Randomization is currently a widely used approach in Sim2Real transfer for data-driven learning algorithms in robotics. Still, most Sim2Real studies report results for a specific randomization technique and often on a highly customized robotic system, making it difficult to evaluate different randomization approaches systematically. To address this problem, we define an easy-to-reproduce experimental setup for a robotic reach-and-balance manipulator task, which can serve as a benchmark for comparison. We compare four randomization strategies with three randomized parameters both in simulation and on a real robot. Our results show that more randomization helps in Sim2Real transfer, yet it can also harm the ability of the algorithm to find a good policy in simulation. Fully randomized simulations and fine-tuning show differentiated results and translate better to the real robot than the other approaches tested.

Keywords: Additive Manufacturing, Product Design, Development and Prototyping
Abstract: In recent years, with the combination of tissue engineering and additive manufacturing technologies, the possibility of fabricating scaffolds with porosity and complex structure has been improved. Since the properties of most biomaterial inks are influenced by temperature and thereby affect the quality of the scaffolds, a controlled printing environment is very important. This study focuses on temperature monitoring from the nozzle to the working platform. A compact heating jacket is developed to heat the needle and sense its temperature inside the nozzle. It makes it very different from common cartridge heating mechanisms. Moreover, a semi-closed printing environment composed of an air curtain and temperature circulation device is developed to create a stable cooling environment. It improves the uniformity of the work platform and increases by 50% the cooling time efficiency. To demonstrate the robustness for a wide range of temperatures, this study presents two experiments of printing two biomaterial inks at body and low temperatures, respectively.

Keywords: Assembly, Dual Arm Manipulation, Computer Vision for Automation
Abstract: We introduce a robotic assembly system that streamlines the design-to-make workflow for going from a CAD model of a product assembly to a fully programmed and adaptive assembly process. Our system captures (in the CAD tool) the intent of the assembly process for a specific robotic workcell and generates a recipe of task-level instructions. By integrating visual sensing with deep-learned perception models, the robots infer the necessary actions to assemble the design from the generated recipe. The perception models are trained directly from simulation, allowing the system to identify various parts based on CAD information. We demonstrate the system with a workcell of two robots to assemble interlocking 3D part designs. We first build and tune the assembly process in simulation, verifying the generated recipe. Finally, the real robotic workcell assembles the design using the same behavior.

Keywords: Assembly, Reinforcement Learning, Task and Motion Planning
Abstract: Robot assembly discovery (RAD) is a challenging problem that lives at the intersection of resource allocation and motion planning. The goal is to combine a predefined set of objects to form something new while considering task execution with the robot-in-the-loop. In this work, we tackle the problem of building arbitrary, predefined target structures entirely from scratch using a set of Tetris-like building blocks and a robotic manipulator. Our novel hierarchical approach aims at efficiently decomposing the overall task into three feasible levels that benefit mutually from each other. On the high level, we run a classical mixed-integer program for global optimization of block-type selection and the blocks’ final poses to recreate the desired shape. Its output is then exploited to efficiently guide the exploration of an underlying reinforcement learning (RL) policy. This RL policy draws its generalization properties from a flexible graph-based representation that is learned through Q-learning and can be refined with search. Moreover, it accounts for the necessary conditions of structural stability and robotic feasibility that cannot be effectively reflected in the previous layer. Lastly, a grasp and motion planner transforms the desired assembly commands into robot joint movements. We demonstrate our proposed method’s performance on a set of competitive simulated RAD environments, showcase real-world transfer, and report performance and robustness gains compared to an unstructured end-to-end approach.

Keywords: Assembly, Manipulation Planning
Abstract: This paper addresses the problem of copying an unknown assembly of primitives with known shape and appearance using information extracted from a single photograph by an off-the-shelf procedure for object detection and pose estimation. The proposed algorithm uses a simple combination of physical stability constraints, convex optimization and Monte Carlo tree search to plan assemblies as sequences of pick-and-place operations represented by STRIPS operators. It is efficient and, most importantly, robust to the errors in object detection and pose estimation unavoidable in any real robotic system. The proposed approach is demonstrated with thorough experiments on a UR5 manipulator.

Keywords: Additive Manufacturing, Task and Motion Planning, Cooperating Robots
Abstract: We present a new algorithm for coordinating the motion of multiple extruders to increase throughput in fused filament fabrication (FFF)/fused deposition modeling (FDM) additive manufacturing. Platforms based on FFF are commonly available and advantageous to several industries, but are limited by slow fabrication time and could be could be significantly improved through efficient use of multiple extruders. We propose the coordinated toolpath planning problem for systems of extruders mounted as end-effectors on robot arms with the objective of maximizing utilization and avoiding collisions. Building on the idea of dependency graphs introduced in our earlier work, we develop a planning and control framework that precomputes a set of multi-layer toolpath segments from the input model and efficiently assigns them to individual extruders such that executed toolpaths are collision-free. Our method overcomes key limitations of existing methods, including utilization loss from workspace partitioning, precomputed toolpaths subject to collisions with the partially fabricated object, and wasted motion resulting from strict layer-by-layer fabrication. We report simulation results that show a major increase in utilization compared to single and multi-extruder methods, and favorable fabrication results using commodity hardware that demonstrate the feasibility of our method in practice.

Keywords: Industrial Robots, Assembly, Human-Robot Collaboration
Abstract: Work-related musculoskeletal disorders (MSD) are one of the major causes of injuries and absenteeism at work. These lead to important costs in the manufacturing industry. Human-robot collaboration can help decrease this issue by appropriately distributing the tasks and decreasing the workload of the factory worker. This paper proposes a novel generic task allocation approach based on hierarchical finite-state machines for human-robot assembly tasks. The developed framework decomposes first the main task into sub-tasks modeled as state machines. Based on capabilities considerations, workload, and performance estimations, the task allocator assigns the sub-task to the human or robot agent. The algorithm was validated on the assembly of a crusher unit of a smoothie machine using the collaborative Franka Emika Panda robot and showed promising results in terms of productivity thanks to task parallelization, with an improvement of more than 30% of the total assembly time with respect to a collaborative scenario, where the agents perform the tasks sequentially.

Keywords: Assembly, Swarm Robotics
Abstract: The emerging field of passive macro-scale tile-based self-assembly (TBSA) shows promise in enabling effective manufacturing processes by harnessing TBSA’s intrinsic parallelism. However, current TBSA methodologies still do not fulfill their potentials, largely because such assemblies are often prone to errors, and the size of an individual assembly is limited due to insufficient mechanical stability. Moreover, the instability issue worsens as assemblies grow in size. Using a novel type of magnetically-bonded tiles carried by bristle-bot drives, we propose here a framework that reverses this tendency; i.e., as an assembly grows, it becomes more stable. Stability is achieved by introducing two sets of tiles that move in opposite directions, thus zeroing the assembly net force. Using physics-based computational experiments, we compare the performance of the proposed approach with the common orbital shaking method, proving that the proposed system of tiles indeed possesses self-stabilizing characteristics. Our approach enables assemblies containing hundreds of tiles to be built, while the shaking approach is inherently limited to a few tens of tiles. Our results indicate that one of the primary limitations of mechanical, agitation-based TBSA approaches, instability, might be overcome by employing a swarm of free-running, sensorless mobile robots, herein represented by passive tiles at the macroscopic scale.

Keywords: Compliant Assembly, Force and Tactile Sensing, Sensor Fusion
Abstract: Snap-fit peg-in-hole assembly widely exists in both industry and daily life, especially for consumer electronics. The buckle mechanism leads to a damping zone inside the port where insertion force needs to be increased. It is much difficult to automate this process by robots, for size and clearance of the components are always small, and the damping buckle should be perceived and distinguished from solid inner walls of the port. End-effector position control might be invalid, since grasping error will make it difficult to locate the plug accurately. In this article, we undertake this assembly challenge by taking advantage of fingertip tactile perception combined with visual images and force feedback. Raw sensor data is collected, processed, and fused together to be state input of a reinforcement learning network, generating continuous action vectors. We also propose a novel damping zone predictor through feature extraction and multimodal fusion, which is able to identify whether the plug has touched the buckle mechanism, so as to adjust the insertion force. The whole framework is implemented through a common USB Type-C insertion experiment on Franka Panda robot platform, reaching a success rate of 88%. Furthermore, system robustness is verified, and comparisons of different modalities are also conducted.

Keywords: Compliant Assembly, Perception-Action Coupling, Contact Modeling
Abstract: In this paper, we propose a novel and general method for autonomous robotic assembly of arbitrary and complex-shaped parts in the presence of 6-dimensional uncertainty. When a nominal assembly motion of the robot holding a part is stopped by contact due to uncertainty, our method finds the best estimate for the uncertainty and the contact configuration of the part based on sensed force/torque and uses that information to find a more accurate estimate of the goal configuration to guide a recovery motion of the part. It is based on a general, surface-based sphere tree representation of parts, a constrained optimization strategy to find the best estimate of the contact configuration under an uncertainty estimate, and a learned force/torque calibration model to relate computed force/torque and the sensed real force/torque. The method is applied and evaluated on different complex-shaped multi-peg-in-hole tasks. The results show that our method can achieve successful assembly with the presence of realistic 6-D uncertainties more than 10 times of the tight task clearances in terms of orientation clearance (<0.015 rad) and position clearance (<1.5 mm), in all the test cases.

Keywords: Motion and Path Planning, Collision Avoidance, Autonomous Vehicle Navigation
Abstract: Autonomously driving vehicles must be able to navigate in dynamic and unpredictable environments in a collision-free manner. So far, this has only been partially achieved in driverless cars and warehouse installations where marked structures such as roads, lanes, and traffic signs simplify the motion planning and collision avoidance problem. We are presenting a new control approach for car-like vehicles that is based on an unprecedentedly fast-paced A* implementation that allows the control cycle to run at a frequency of 30 Hz. This frequency enables us to place our A* algorithm as a low- level replanning controller that is well suited for navigation and collision avoidance in virtually any dynamic environment. Due to an efficient heuristic consisting of rotate-translate-rotate motions laid out along the shortest path to the target, our Short- Term Aborting A* (STAA*) converges fast and can be aborted early in order to guarantee a high and steady control rate. While our STAA* expands states along the shortest path, it takes care of collision checking with the environment including predicted states of moving obstacles, and returns the best solution found when the computation time runs out. Despite the bounded computation time, our STAA* does not get trapped in corners due to the following of the shortest path. In simulated and real-robot experiments, we demonstrate that our control approach eliminates collisions almost entirely and is superior to an improved version of the Dynamic Window Approach with predictive collision avoidance capabilities.

Keywords: Motion and Path Planning, Collision Avoidance, Robotics and Automation in Agriculture and Forestry
Abstract: This paper presents a path planning refinement technique that allows the efficient collision and rollover-free motion planning for mobile manipulator robots working on rough terrain. First, the necessary theoretical background on a mobile manipulator's kinematics and dynamic stability measure is introduced. Then, after the brief introduction of the sampling-based path planning problem, the additional refinement stage and its problem formulation will be introduced. Within the refinement stage, the novel B'{e}zier control point addition method is introduced to allow for fast, collision and rollover-free path smoothing using curvature-continuous parametrized curves. Analytical proofs and simulated comparisons are provided in the paper to show effectiveness. The beneficial effect of the refined path on trajectory planning will also be demonstrated through simulation.

Keywords: Motion and Path Planning, Autonomous Vehicle Navigation, Mapping
Abstract: In recent years, mobile robots are becoming ambitious and deployed in large-scale scenarios. Serving as a high-level understanding of environments, a sparse skeleton graph is beneficial for more efficient global planning. Currently, existing solutions for skeleton graph generation suffer from several major limitations, including poor adaptiveness to different map representations, dependency on robot inspection trajectories and high computational overhead. In this paper, we propose an efficient and flexible algorithm generating a trajectory-independent 3D sparse topological skeleton graph capturing the spatial structure of the free space. In our method, an efficient ray sampling and validating mechanism are adopted to find distinctive free space regions, which contributes to skeleton graph vertices, with traversability between adjacent vertices as edges. A cycle formation scheme is also utilized to maintain skeleton graph compactness. Benchmark comparison with state-of-the-art works demonstrates that our approach generates sparse graphs in a substantially shorter time, giving high-quality global planning paths. Experiments conducted in real-world maps further validate the capability of our method in real-world scenarios. Our method will be made open source to benefit the community.

Keywords: Motion and Path Planning, Autonomous Vehicle Navigation, Collision Avoidance
Abstract: This paper addresses a safe planning and control problem for mobile robots operating in communication- and sensor-limited dynamic environments. In this case the robots cannot sense the objects around them and must instead rely on intermittent, external information about the environment, as e.g., in underwater applications. The challenge in this case is that the robots must plan using only this stale data, while accounting for any noise in the data or uncertainty in the environment. To address this challenge we propose a compositional technique which leverages neural networks to quickly plan and control a robot through crowded and dynamic environments using only intermittent information. Specifically, our tool uses reachability analysis and potential fields to train a neural network that is capable of generating safe control actions. We demonstrate our technique both in simulation with an underwater vehicle crossing a crowded shipping channel and with real experiments with ground vehicles in communication- and sensor-limited environments.

Keywords: Motion and Path Planning, Multi-Robot Systems, Machine Learning for Robot Control
Abstract: In this paper, we present a decentralized control approach based on a Nonlinear Model Predictive Control (NMPC) method that employs barrier certificates for safe navigation of multiple nonholonomic wheeled mobile robots in unknown environments with static and/or dynamic obstacles. This method incorporates a Learned Barrier Function (LBF) into the NMPC design in order to guarantee safe robot navigation, i.e., prevent robot collisions with other robots and the obstacles. We refer to our proposed control approach as NMPC-LBF. Since each robot does not have a priori knowledge about the obstacles and other robots, we use a Deep Neural Network (DeepNN) running in real-time on each robot to learn the Barrier Function (BF) only from the robot's LiDAR and odometry measurements. The DeepNN is trained to learn the BF that separates safe and unsafe regions. We implemented our proposed method on simulated and actual Turtlebot3 Burger robot(s) in different scenarios. The implementation results show the effectiveness of the NMPC-LBF method at ensuring safe navigation of the robots.

Keywords: Motion and Path Planning, Autonomous Vehicle Navigation, Intelligent Transportation Systems
Abstract: Trajectory planning is a critical component in autonomous vehicles directly responsible for driving safety and efficiency during deployment. The ability to find the optimal trajectory in real-time is critical for autonomous driving. This paper presents a novel general framework using the Fast Iterative Search and Sampling (FISS) strategy for sampling-based trajectory planning, which can find the optimal trajectory from an enormous number of candidates with high efficiency in real-time. Specifically, before generating any trajectories, the proposed method utilizes historical planning results as prior information in heuristics to estimate the cost distribution over the sampling space. On this basis, the Fast Iterative Search and Sampling strategy is employed to explore the sampling space for possible candidates and generate trajectories for verification during the search process. Experimental results show that our method can significantly outperform existing frameworks by order of magnitude in planning efficiency while ensuring safety and maintaining high accuracy.

Keywords: Motion and Path Planning, Collision Avoidance, Constrained Motion Planning
Abstract: This paper proposes a path planner that reshapes a global path locally in response to sensor-based observations of obstacles in the environment. Two fundamental concepts enable the resultant algorithm (a) a path-following synthetic vehicle whose steering actions are non-myopically optimized to result in a smooth traversible path that meets path curvature constraints, and (b) a path-aware turning moment-field that enables obstacle avoidance while eluding the typical local-minimum-induced stagnation associated with potential field methods. The use of the combination of the two concepts results in a reduced action space over which optimization needs to be performed towards minimizing the path deviation subject to obstacle avoidance, and thus results in an efficient algorithm that can be implemented online. We demonstrate the algorithm in simulations as well as in field experiments, performing real time local path planning and obstacle avoidance on two different vehicle platforms (an ackerman steering vehicle two-axled vehicle, and a differential-steering 4-axled vehicle) in an unstructured off-road terrain.

Keywords: Motion and Path Planning, Collision Avoidance
Abstract: Sampling-based motion planners are widely used in robotics due to their simplicity, flexibility and computational efficiency. However, in their most basic form, these algorithms operate under the assumption of static scenes and lack the ability to avoid collisions with dynamic (i.e. moving) obstacles. This raises safety concerns, limiting the range of possible applications of mobile robots in the real world. Motivated by these challenges, in this work we present Temporal-PRM, a novel sampling-based path-planning algorithm that performs obstacle avoidance in dynamic environments. The proposed approach extends the original Probabilistic Roadmap (PRM) with the notion of time, generating an augmented graph-like structure that can be efficiently queried using a time-aware variant of the A* search algorithm, also introduced in this paper. Our design maintains all the properties of PRM, such as the ability to perform multiple queries and to find smooth paths, while circumventing its downside by enabling collision avoidance in highly dynamic scenes with a minor increase in the computational cost. Through a series of challenging experiments in highly cluttered and dynamic environments, we demonstrate that the proposed path planner outperforms other state-of-the-art sampling-based solvers. Moreover, we show that our algorithm can run onboard a flying robot, performing obstacle avoidance in real time.

Keywords: Motion and Path Planning, Collision Avoidance, Computational Geometry
Abstract: We present a hierarchical skeleton-guided motion planning algorithm to guide mobile robots. A good skeleton maps the connectivity of the subspace of c-space containing significant degrees of freedom and is able to guide the planner to find the desired solutions fast. However, sometimes the skeleton does not closely represent the free c-space, which often misleads current skeleton-guided planners. The hierarchical skeleton-guided planning strategy gradually relaxes its reliance on the workspace skeleton as Cspace is sampled, thereby incrementally returning a sub-optimal path, a feature that is not guaranteed in the standard skeleton-guided algorithm. Experimental comparisons to the standard skeleton guided planners and other lazy planning strategies show significant improvement in roadmap construction run time while maintaining path quality for multi-query problems in cluttered environments.

Keywords: Legged Robots, Deep Learning Methods, Reinforcement Learning
Abstract: We propose a simple imitation learning procedure for learning locomotion controllers that can walk over very challenging terrains. We use trajectory optimization (TO) to produce a large dataset of trajectories over procedurally generated terrains and use Reinforcement Learning (RL) to imitate these trajectories. We demonstrate with a realistic model of the ANYmal robot that the learned controllers transfer to unseen terrains and provide an effective initialization for fine-tuning on challenging terrains that require exteroception and precise foot placements. Our setup combines TO and RL in a simple fashion that overcomes the computational limitations and need for a robust tracking controller of the former and the exploration and reward-tuning difficulties of the latter.

Keywords: Legged Robots, Mechanism Design, Optimization and Optimal Control
Abstract: We present a versatile framework for the computational co-design of legged robots and dynamic maneuvers. Current state-of-the-art approaches are typically based on random sampling or concurrent optimization. We propose a novel bilevel optimization approach that exploits the derivatives of the motion planning sub-problem (i.e., the lower level). These motion-planning derivatives allow us to incorporate arbitrary design constraints and costs in an general-purpose nonlinear program (i.e., the upper level). Our approach allows for the use of any differentiable motion planner in the lower level and also allows for an upper level that captures arbitrary design constraints and costs. It efficiently optimizes the robot’s morphology, payload distribution and actuator parameters while considering its full dynamics, joint limits and physical constraints such as friction cones. We demonstrate these capabilities by designing quadruped robots that jump and trot. We show that our method is able to design a more energy-efficient Solo robot for these tasks.

Keywords: Legged Robots, Human and Humanoid Motion Analysis and Synthesis, Motion and Path Planning
Abstract: The complex dynamics of agile robotic legged locomotion requires motion planning to intelligently adjust footstep locations. Often, bipedal footstep and motion planning use mathematically simple models such as the linear inverted pendulum, instead of dynamically-rich models that do not have closed-form solutions. We propose a real-time optimization method to plan for dynamical models that do not have closed form solutions and experience irrecoverable failure. Our method uses a data-driven approximation of the step-to-step dynamics and of a failure margin function. This failure margin function is an oriented distance function in state-action space where it describes the signed distance to success or failure. The motion planning problem is formed as a nonlinear program with constraints that enforce the approximated forward dynamics and the validity of state-action pairs. For illustration, this method is applied to create a planner for an actuated spring-loaded inverted pendulum model. In an ablation study, the failure margin constraints decreased the number of invalid solutions by between 24 and 47 percentage points across different objectives and horizon lengths. While we demonstrate the method on a canonical model of locomotion, we also discuss how this can be applied to data-driven models and full-order robot models.

Keywords: Legged Robots, Energy and Environment-Aware Automation, Learning from Demonstration
Abstract: This work reports on developing a deep inverse reinforcement learning method for legged robots terrain traversability modeling that incorporates both exteroceptive and proprioceptive sensory data. Existing works use robot-agnostic exteroceptive environmental features or handcrafted kinematic features; instead, we propose to also learn robot-specific inertial features from proprioceptive sensory data for reward approximation in a single deep neural network. Incorporating the inertial features can improve the model fidelity and provide a reward that depends on the robot's state during deployment. We train the reward network using the Maximum Entropy Deep Inverse Reinforcement Learning (MEDIRL) algorithm and propose simultaneously minimizing a trajectory ranking loss to deal with the suboptimality of legged robot demonstrations. The demonstrated trajectories are ranked by locomotion energy consumption, in order to learn an energy-aware reward function and a more energy-efficient policy than demonstration. We evaluate our method using a dataset collected by an MIT Mini-Cheetah robot and a Mini-Cheetah simulator. The code is publicly available at https://github.com/ganlumomo/minicheetah-traversability-irl.

Keywords: Legged Robots, Machine Learning for Robot Control, Motion Control
Abstract: Deep reinforcement learning has emerged as a popular and powerful way to develop locomotion controllers for quadruped robots. Common approaches have largely focused on learning actions directly in joint space, or learning to modify and offset foot positions produced by trajectory generators. Both approaches typically require careful reward shaping and training for millions of time steps, and with trajectory generators introduce human bias into the resulting control policies. In this paper, we present a learning framework that leads to the natural emergence of fast and robust bounding policies for quadruped robots. The agent both selects and controls actions directly in task space to track desired velocity commands subject to environmental noise including model uncertainty and rough terrain. We observe that this framework improves sample efficiency, necessitates little reward shaping, leads to the emergence of natural gaits such as galloping and bounding, and eases the sim-to-real transfer at running speeds. Policies can be learned in only a few million time steps, even for challenging tasks of running over rough terrain with loads of over 100% of the nominal quadruped mass. Training occurs in PyBullet, and we perform a sim-to-sim transfer to Gazebo and sim-to-real transfer to the Unitree A1 hardware. For sim-to-sim, our results show the quadruped is able to run at over 4 m/s without a load, and 3.5 m/s with a 10 kg load, which is over 83% of the nominal quadruped mass. For sim-to-real, the Unitree A1 is able to bound at 2 m/s with a 5 kg load, representing 42% of the nominal quadruped mass.

Keywords: Legged Robots, Motion Control, Multi-Contact Whole-Body Motion Planning and Control
Abstract: This work investigates a data-driven template model for trajectory planning of dynamic quadrupedal robots. Many state-of-the-art approaches involve using a reduced-order model, primarily due to computational tractability. The spirit of the trajectory planning approach in this work draws on recent advancements in the area of behavioral systems theory. Here, we aim to capitalize on the knowledge of well-known template models to construct a data-driven model, enabling us to obtain an information rich reduced-order model. In particular, this work considers input-output states similar to that of the single rigid body model and proceeds to develop a data-driven representation of the system, which is then used in a predictive control framework to plan a trajectory for quadrupeds. The optimal trajectory is passed to a low-level and nonlinear model-based controller to be tracked. Preliminary experimental results are provided to establish the efficacy of this hierarchical control approach for trotting and walking gaits of a high-dimensional quadrupedal robot on unknown terrains and in the presence of disturbances.

Keywords: Legged Robots, Dynamics
Abstract: Existing work in legged robot navigation in cluttered environments often seeks collision-free paths that avoid obstacle interactions. Here we present a new approach for multi-legged robots to utilize leg-obstacle collisions to generate desired dynamics across obstacle fields. To predict the change of robot state (i.e., position and orientation) under repeated leg-obstacle collisions, we construct a discretized directed graph model: each node of the graph represents a different robot state, whereas the directed edges pointing from one node to another represent the transitions from one robot state to the next within one stride. These obstacle-modulated state transitions can depend on the robot gaits used. To capture this dependence, an empirical interaction model is used to compute the change of robot state based on initial contact positions between robot legs and obstacles. To validate the prediction of robot state transitions, we experimentally measure the state of a quadrupedal robot as it traverses a periodic obstacle field with three different gaits: bound, trot, and pace. We observed that the robot could passively converge to different steady state orientations, and these steady states corresponded well with the Strongly-Connected-Path-Components (SCPCs) within the directed graph. Searching over the graph for connected paths of SCPCs allows development of a gait planner that can generate gait switching strategies for a robot to achieve desired states by simply engaging with a sequence of leg-obstacle collisions. We demonstrate in experiments that, by using the gait sequences generated by our planner, an open-loop quadrupedal robot was able to successfully achieve desired orientations within a periodic obstacle field without any sensory input or active steering.

Keywords: Legged Robots, Hardware-Software Integration in Robotics, Sensorimotor Learning
Abstract: Inspired by the digital twinning systems, a novel real-time digital double framework is developed to enhance robot perception of the terrain conditions. Based on the very same physical model and motion control, this work exploits the use of such simulated digital double synchronized with a real robot to capture and extract discrepancy information between the two systems, which provides high dimensional cues in multiple physical quantities to represent differences between the modelled and the real world. Soft, non-rigid terrains cause common failures in legged locomotion, whereby visual perception solely is insufficient in estimating such physical properties of terrains. We used digital double to develop the estimation of the collapsibility, which addressed this issue through physical interactions during dynamic walking. The discrepancy in sensory measurements between the real robot and its digital double are used as input of a learning-based algorithm for terrain collapsibility analysis. Although trained only in simulation, the learned model can perform collapsibility estimation successfully in both simulation and real world. Our evaluation of results showed the generalization to different scenarios and the advantages of the digital double to reliably detect nuances in ground conditions.

Keywords: Compliant Joints and Mechanisms, Mechanism Design, Biologically-Inspired Robots
Abstract: The vast majority of state-of-the-art walking robots employ flat or ball feet for locomotion, presenting limitations while stepping on obstacles, slopes, or unstructured terrain. Moreover, traditional feet for quadrupeds lack sensing systems that are able to provide information about the environment and about the foot interaction with the surroundings. This further diminishes their value. Inspired by our previous work on soft feet for bipedal robots, we present the SoftFoot-Q, an articulated adaptive foot for quadrupeds. This device is conceived to be robust and able to overcome the limitations of currently employed feet. The core idea behind our adaptive foot design is first introduced and validated through a simplified mathematical formulation of the problem. Subsequently, we present the chosen mechanical implementation to attempt overcoming current limitations. The realized prototype of adaptive foot is integrated and tested on the compliantly actuated quadrupedal robot ANYmal together with a ROS based real-time foot pose reconstruction software. Both extensive field tests and indoor experiments show noticeable performance improvements, in terms of reduced slippage of the robot, with respect to both flat and ball feet.

Keywords: Art and Entertainment Robotics, Force and Tactile Sensing, Sensorimotor Learning
Abstract: As Liszt once said "(a virtuoso) must call up scent and blossom, and breathe the breath of life", a virtuoso plays the piano with passion, poetry, and extraordinary technical ability. Hence, piano playing, being a task that is quintessentially human, becomes a hallmark for roboticians and artificial intelligence researchers to pursue. In this paper, we advocate an end-to-end reinforcement learning (RL) paradigm to demonstrate how an agent can learn directly from machine-readable music score to play the piano with touch-augmented dexterous hands on a simulated piano. To achieve the desired tasks, we design useful touch- and audio-based reward functions and a series of tasks. Empirical results show that the RL agent can not only find the correct key position but also deal with the various rhythmic, volume, and fingering requirements. As a result, the agent demonstrates its effectiveness in playing simple pieces that have different musical requirements which show the potential of leveraging reinforcement learning approach for the piano playing tasks.

Keywords: Dual Arm Manipulation, Multi-Robot Systems, Reinforcement Learning
Abstract: We develop two consensus-based learning algorithms for multi-robot systems applied on complex tasks involving collision constraints and force interactions, such as the cooperative peg-in-hole placement. The proposed algorithms integrate multi-robot distributed consensus and normalizing-flow-based reinforcement learning. The algorithms guarantee the stability and the consensus of the multi-robot system's generalized variables in a transformed space. This transformed space is obtained via a diffeomorphic transformation parameterized by normalizing-flow models that the algorithms use to train the underlying task, learning hence skillful, dexterous trajectories required for the task accomplishment. We validate the proposed algorithms by parameterizing reinforcement learning policies, demonstrating efficient cooperative learning, and strong generalization of dual-arm assembly skills in a dynamics-engine simulator.

Keywords: Dual Arm Manipulation, Task Planning, Cooperating Robots
Abstract: We investigate the problem of coordinating two robot arms to solve non-monotone tabletop multi-object rearrangement tasks. In a non-monotone rearrangement task, complex object-object dependencies exist that require moving some objects multiple times to solve an instance. In working with two arms in a large workspace, some objects must be handed off between the robots, which further complicates the planning process. For the challenging dual-arm tabletop rearrangement problem, we develop effective task planning algorithms for scheduling the pick-n-place sequence that can be properly distributed between the two arms. We show that, even without using a sophisticated motion planner, our method achieves significant time savings in comparison to greedy approaches and naive parallelization of single-robot plans.

Keywords: Biological Cell Manipulation, Computer Vision for Automation
Abstract: Visual localization, which is a key technology to realize the automation of cell manipulation, has been widely studied. Since the depth of field of the microscope is narrow, the planar localization and depth estimation are usually coupled together. At present, most methods adopt the serial working mode of focusing first and then planar localization, but they usually do not have good real-time performance and stability. In this paper, a simultaneous depth estimation and localization network was developed for cell manipulation. The network takes a focused image and a defocus-offset image as inputs, and outputs the defocus in the depth direction and the offset in the plane at the same time after going through defocus-offset information extraction, defocus classification mapping and offset regression mapping. To train and test our network, we also create two datasets: An Adherent Cell dataset and an Injection Micropipette dataset. The experimental results demonstrated that the proposed method achieves the detection of all test samples with a frame rate of more than 40Hz, and the maximum errors of depth estimation and localization are 2.44μm and 0.49μm, respectively. The proposed method has good stability, which is mainly reflected in its strong generalization ability and anti-noise ability.

Keywords: Dual Arm Manipulation, Multi-Robot Systems, Cooperating Robots
Abstract: We address the problem of cooperative manipulation of an object whose tasks are specified by a Signal Temporal Logic (STL) formula. We employ the Prescribed Performance Control (PPC) methodology to guarantee predefined transient and steady-state performance on the object trajectory in order to satisfy the STL formula. More specifically, we first provide a way that translates the problem of satisfaction of an STL task to the problem of state evolution within a user-defined time-varying funnel. We then design a control strategy for the robotic agents that guarantees compliance with this funnel. The control strategy is decentralized, in the sense that each agent calculates its own control signal, and does not use any information on the agents' and object's dynamic terms, which are assumed to be unknown. We experimentally verify the results on two manipulator arms, cooperatively working to manipulate an object based on a STL formula.

Keywords: Art and Entertainment Robotics, Motion Control
Abstract: We present drozBot: le robot portraitiste, a robotic system that draws artistic portraits of people. The input images for the portrait are taken interactively by the robot itself. We formulate the problem of drawing portraits as a problem of coverage which is then solved by an ergodic control algorithm to compute the strokes. The ergodic computation of the strokes for the portrait gives an artistic look to them. The specific ergodic control algorithm that we chose is inspired by the heat equation. We employed a 7-axis Franka Emika robot for the physical drawings and used an optimal control strategy to generate joint angle commands. We explain the influence of the different hyperparameters and show the importance of the image processing steps. The attractiveness of the results was evaluated by conducting a survey where we asked the participants to rank the portraits produced by different algorithms.

Keywords: Physically Assistive Devices, Deep Learning for Visual Perception, Perception for Grasping and Manipulation
Abstract: Robotic manipulation of highly deformable cloth presents a promising opportunity to assist people with several daily tasks, such as washing dishes; folding laundry; or dressing, bathing, and hygiene assistance for individuals with severe motor impairments. In this work, we introduce a formulation that enables a collaborative robot to perform visual haptic reasoning with cloth--the act of inferring the location and magnitude of applied forces during physical interaction. We present two distinct model representations, trained in physics simulation, that enable haptic reasoning using only visual and robot kinematic observations. We conducted quantitative evaluations of these models in simulation for robot-assisted dressing, bathing, and dish washing tasks, and demonstrate that the trained models can generalize across different tasks with varying interactions, human body sizes, and object shapes. We also present results with a real-world mobile manipulator, which used our simulation-trained models to estimate applied contact forces while performing physically assistive tasks with cloth. Videos can be found at our project webpage: https://sites.google.com/view/visualhapticreasoning/home

Keywords: Dual Arm Manipulation, In-Hand Manipulation, Cooperating Robots
Abstract: The main purpose of this paper is to demonstrate that smart exploitation of force/tactile feedback can enable successful physical cooperation of two robot manipulators to handle a common object with a high degree of dexterity. The novelty of the paper is that dexterity is provided not only by the degrees of freedom of the robot arms but also by the grasp controller of the sensorized parallel grippers, which allow the robots to manipulate the object either with a tight grasp or with a one-degree-of-freedom rolling contact. The coordinated motion of the robots depends on both the desired motion of the carried object and the control of the internal forces during transportation and in-hand manipulation. The solution exploits only kinematic models of the robots and a dynamic model of the distributed soft contact, which includes both linear force and torsional moment.

Keywords: Deep Learning in Grasping and Manipulation, Object Detection, Segmentation and Categorization, Grasping
Abstract: Object manipulation in cluttered scenes is a difficult and important problem in robotics. To efficiently manipulate objects, it is crucial to understand their surroundings, especially in cases where multiple objects are stacked one on top of the other, preventing effective grasping. We here present DUQIM-Net, a decision-making approach for object manipulation in a setting of stacked objects. In DUQIM-Net, the hierarchical stacking relationship is assessed using Adj-Net, a model that leverages existing Transformer Encoder-Decoder object detectors by adding an adjacency head. The output of this head probabilistically infers the underlying hierarchical structure of the objects in the scene. We utilize the properties of the adjacency matrix in DUQIM-Net to perform decision making and assist with object-grasping tasks. Our experimental results show that Adj-Net surpasses the state-of-the-art in object-relationship inference on the Visual Manipulation Relationship Dataset (VMRD), and that DUQIM-Net outperforms comparable approaches in bin clearing tasks.

Keywords: Aerial Systems: Mechanics and Control, Motion Control
Abstract: Tracking position and orientation independently affords more agile maneuver for over-actuated multirotor Unmanned Aerial Vehicles (UAVs) while introducing undesired downwash effects; downwash flows generated by thrust generators may counteract others due to close proximity, which significantly threatens the stability of the platform. The complexity of modeling aerodynamic airflow challenges control algorithms from properly compensating for such a side effect. Leveraging the input redundancies in over-actuated UAVs, we tackle this issue with a novel control allocation framework that considers downwash effects and explores the entire allocation space for an optimal solution. This optimal solution avoids downwash effects while providing high thrust efficiency within the hardware constraints. To the best of our knowledge, ours is the first formal derivation to investigate the downwash effects on over-actuated UAVs. We verify our framework on different hardware configurations in both simulation and experiment.

Keywords: Aerial Systems: Applications, Deep Learning for Visual Perception, Data Sets for Robotic Vision
Abstract: Although the manipulating of the unmanned aerial manipulator (UAM) has been widely studied, vision-based UAM approaching, which is crucial to the subsequent manipulating, generally lacks effective design. The key to the visual UAM approaching lies in object tracking, while current UAM tracking typically relies on costly model-based methods. Besides, UAM approaching often confronts more severe object scale variation issues, which makes it inappropriate to directly employ state-of-the-art model-free Siamese-based methods from the object tracking field. To address the above problems, this work proposes a novel Siamese network with pairwise scale-channel attention (SiamSA) for vision-based UAM approaching. Specifically, SiamSA consists of a pairwise scale-channel attention network (PSAN) and a scale-aware anchor proposal network (SA-APN). PSAN acquires valuable scale information for feature processing, while SAAPN mainly attaches scale awareness to anchor proposing. Moreover, a new tracking benchmark for UAM approaching, namely UAMT100, is recorded with 35K frames on a flying UAM platform for evaluation. Exhaustive experiments on the benchmarks and real-world tests validate the efficiency and practicality of SiamSA with a promising speed. Both the code and UAMT100 benchmark are now available at https:// github.com/vision4robotics/SiamSA.

Keywords: Aerial Systems: Mechanics and Control, Simulation and Animation, Modeling, Control, and Learning for Soft Robots
Abstract: Flying animals possess highly complex physical characteristics and are capable of performing agile maneuvers using their wings. The flapping wings generate complex wake structures that influence the aerodynamic forces, which can be difficult to model. While it is possible to model these forces using fluid-structure interaction, it is very computationally expensive and difficult to formulate. In this paper, we follow a simpler approach by deriving the aerodynamic forces using a relatively small number of states and presenting them in a simple state-space form. The formulation utilizes Prandtl's lifting line theory and Wagner's function to determine the unsteady aerodynamic forces acting on the wing in a simulation, which then are compared to experimental data of the bat-inspired robot called the Aerobat. The simulated trailing-edge vortex shedding can be evaluated from this model, which then can be analyzed for a wake-based gait design approach to improve the aerodynamic performance of the robot.