Keywords: Grasping and Manipulation, Control, Humanoid Robot Systems
Abstract: Coordinated dual-arm manipulation tasks can be broadly characterized as possessing absolute and relative motion components. Relative motion tasks, in particular, are inherently redundant in the way they can be distributed between end-effectors. In this work, we analyse cooperative manipulation in terms of the asymmetric resolution of relative motion tasks. We discuss how existing approaches enable the asymmetric execution of a relative motion task, and show how an asymmetric relative motion space can be defined. We leverage this result to propose an extended relative Jacobian to model the cooperative system, which allows a user to set a concrete degree of asymmetry in the task execution. This is achieved without the need for prescribing an absolute motion target. Instead, the absolute motion remains available as a functional redundancy to the system. We illustrate the properties of our proposed Jacobian through numerical simulations of a novel differential Inverse Kinematics algorithm.

Keywords: Control
Abstract: In this work, we leverage the alternating direction method of multipliers (ADMM) framework to solve task assignment for a multi-robot team. While ADMM is a well-established method, it has yet to be utilized in multi-robot problems as the standard formulation requires a centralized update step, a paradigm that conflicts with decentralization as a means of robustness. Therefore, we describe the formulation of separable optimizations in order to produce decentralized ADMM algorithms. Here, our aim is to provide an additional tool for solving cooperative team-based problems in robotics. For the decentralized algorithms, we discuss the conditions for convergence to the optimal centralized solution. We present simulation results for task assignment to demonstrate the benefits of ADMM compared to state-of-the-art methods.

Keywords: Control, AI-enabled Robotics
Abstract: Motion planning problems can be simplified by admissible projections of the configuration space to sequences of lower-dimensional quotient-spaces, called sequential simplifications. To exploit sequential simplifications, we present the Quotient-space Rapidly-exploring Random Trees (QRRT) algorithm. QRRT takes as input a start and a goal configuration, and a sequence of quotient-spaces. The algorithm grows trees on the quotient-spaces both sequentially and simultaneously to guarantee a dense coverage. QRRT is shown to be (1) probabilistically complete, and (2) can reduce the runtime by at least one order of magnitude. However, we show in experiments that the runtime varies substantially between different quotient-space sequences. To find out why, we perform an additional experiment, showing that the more narrow an environment, the more a quotient-space sequence can reduce runtime.

Keywords: Control
Abstract: This paper investigates optimization-based planning methods for generating trajectories which are robust to state uncertainty in undersensed and underactuated systems. Specifically, these methods are applied to an undersensed robotic hill climbing system. In previous work, divergence metrics based on contraction analysis were used to quantify robustness of a trajectory to state uncertainty in conjunction with a kinodynamic RRT planner to guide the planner towards more convergent directions. Resulting trajectories were sub-optimal or needed to be smoothed prior to implementation. This work proposes an optimization framework to plan optimally robust and smooth trajectories which can also be readily implemented on the robotic hill climbing problem. A new hill climbing controller is also presented which can guarantee for the first time the strongest result of contraction analysis, global asymptotic convergence, where possible. Trajectories created using the new trajectory optimization framework and hill controller are shown to be smoother and more robust than previous methods as well as an asymptotically optimal versions of previous methods.

Keywords: Control
Abstract: As autonomous systems rely increasingly on onboard sensors for localization and perception, the parallel tasks of motion planning and uncertainty minimization become increasingly coupled. This coupling is well-captured by augmenting the planning objective with a posterior-covariance penalty – however, online optimization can be computationally intractable, particularly for observation models with latent environmental dependencies (e.g., unknown landmarks). This paper addresses a number of fundamental challenges in efficient minimization of the posterior covariance. First, we provide a measurement bundling approximation that enables high-rate sensors to be approximated with fewer, low-rate updates. This allows for landmark marginalization (crucial in the case of unknown landmarks), for which we provide a novel recipe for computing the gradients necessary for optimization. Finally, we identify a large class of measurement models for which the contributions from each landmark can be combined, so evaluation of the total information gained at each timestep can be carried out (nearly) independently of the number of landmarks. We evaluate our trajectory-generation framework for both a Dubin’s car and a quadrotor, demonstrating significant estimation improvement and moderate computation time.

Keywords: Control, Grasping and Manipulation, Human-Robot Interaction
Abstract: Cartesian impedance control has been widely used in controlling robotic manipulators for manipulation and assembly tasks to compute the required joint torques and/or end-effector forces. Current congruence mapping of the Cartesian stiffness and damping matrices to joint spaces are not valid for general cases. In this paper, we derive from first principles, the general form of the mapping of the stiffness and damping matrices between Cartesian and joint space, which applies to all general cases. The new results show the coupling of Cartesian damping in the stiffness after mapping to the joint space, which is not found in the literature. By applying principle of vibration analysis and including the null space tasks, we can choose the stiffness and damping matrices in order to achieve prescribed dynamic response. Such analysis also enables us to gain a deeper understanding of the responses versus the parameters of robot manipulators; for example, certain elements only affect dynamic responses in specific directions.

Keywords: Soft Robotics, Bioinspired Robotics, Design
Abstract: A Cayley map for the special Euclidean group SE(3) is introduced to relate, for a soft continuum robot, the Lie algebra of internal deformations with the Lie group of rigid-body motions. This Cayley map is used for approximated and exact kinematic shape reconstruction of soft continuum robots, under the hypothesis of constant deformations. This map could be used for deriving computationally efficient interpolation schemes for soft robots, since it does not involve transcendental functions as those introduced by the exponential parametrization of soft robot kinematics.

Keywords: Control, Humanoid Robot Systems
Abstract: This paper discusses an optimization method of the wholebody compliance for stable and robust control of a humanoid robot. In a previous study, one of the authors proposed resolving the virtual viscoelasticity at the center of gravity into the joint viscoelasticity, considering the redundant degrees of freedom, and named this method as resolved viscoelasticity control (RVC). However, the author considered only the relationship based on statics. In this study, the authors extend the previous work on the RVC by considering dynamics. This extension helps to realize stable and robust balancing. We also provide a comparison between the RVC and the control method based on the operational space formulation. The proposed method is validated using forward dynamics simulations.

Keywords: Bioinspired Robotics, Grasping and Manipulation, Control
Abstract: We document the reliably repeatable dynamical mounting and dismounting of wheeled stools and carts, and of fixed ledges, by the Minitaur robot. Because these tasks span a range of length scales that preclude quasi-static execution, we use a hybrid dynamical systems framework to variously compose and thereby systematically reuse a small lexicon of templates (low degree of freedom behavioral primitives). The resulting behaviors comprise the key competences beyond mere locomotion required for robust implementation on a legged mobile manipulator of a simple version of the warehouseman's problem.

Keywords: AI-enabled Robotics
Abstract: A novel modality of ultrasound imaging known as elastography has been shown to improve cancer detection for women with dense breast tissue. However, the scanning procedure for this technique is often difficult for a human to perform in a consistent manner and could conceivably benefit from robot assistance. In this work, we present a novel robot-assisted probabilistic elasticity mapping algorithm which uses Gaussian filter techniques to produce elastograms and uncertainty maps. We demonstrate the proposed approach using a 7-DOF robot manipulator on a gelatin phantom designed to imitate the elasticity of human tissue. The results indicate the algorithm is capable of imaging a 6.5 mm lesion and reducing map uncertainty in the observable region.

Keywords: Grasping and Manipulation
Abstract: This paper studies robustness in planar grasping from a geometric perspective. By treating grasping as a process that shapes the free-space of an object over time, we can define three types of certificates to guarantee success of a grasp: (a) invariance under an initial set, (b) convergence towards a goal grasp, and (c) observability over the final object pose. We develop convex-combinatorial models for each of these certificates, which can be expressed as simple semi-algebraic relations under mild-modeling assumptions. By leveraging these models to synthesize certificates, we optimize certifiable grasps of arbitrary planar objects composed as a union of convex polygons, using manipulators described as point-fingers. We validate this approach with simulations and real robot experiments, by grasping random polygons, comparing against other standard grasp planning algorithms, and performing sensorless grasps over different objects.

Keywords: Grasping and Manipulation, Robot Vision, Robot Learning
Abstract: We would like robots to achieve purposeful manipulation by placing any instance from a category of objects into a desired set of goal states. Existing manipulation pipelines typically specify the desired configuration as a target 6-DOF pose and rely on explicitly estimating the pose of the manipulated objects. However, representing an object with a parameterized transformation defined on a fixed template cannot capture large intra-category shape variation, and specifying a target pose at a category level can be physically infeasible or fail to accomplish the task, e.g. knowing the pose and size of a mug relative to a canonical mug is not sufficient to hang it on a rack by its handle. We propose a novel formulation of category-level manipulation that uses semantic 3D keypoints as the object representation. This keypoint representation enables a simple and interpretable specification of the manipulation target as geometric costs and constraints on the keypoints, which flexibly generalizes existing pose-based manipulation methods. Using this formulation, we factor the manipulation policy into instance segmentation, 3D keypoint detection, optimization-based robot action planning and local dense-geometry-based action execution. This factorization allows us to leverage advances in these sub-problems and combine them into a general and effective perception-to-action manipulation pipeline. Our pipeline is robust to large intra-category shape variation and topology changes as the keypoint representation ignores task-irrelevant geometric details. Extensive hardware experiments demonstrate our method can reliably accomplish tasks with never-before seen objects in a category, such as placing shoes and mugs with significant shape variation into category level target configurations.

Keywords: Grasping and Manipulation, Control, Design
Abstract: In this paper, we present the achievement of aerial manipulation and grasping by a novel multilinked aerial robot called DRAGON, in which a pair of rotors is embedded in each link. First, the unique mechanical design of this robot is briefly introduced. The key to performing stable manipulation and grasping is the two degrees-of-freedom rotor vectoring apparatus. Second, the autonomous flight control framework is presented, which is followed by the derivation of the external wrench compensation during interaction with an environment or object. Third, on the basis of external wrench compensation, the motion planning methods involving the optimization problem are presented for different cases: single end-effector manipulation and two-point (i.e., the ends) grasping. Finally, we show the experimental motion results for the movement of a rectangle plate by contact with different parts of the plate, as well as the grasping of different objects by the two ends of the robot.

Keywords: Grasping and Manipulation, Human-Robot Interaction
Abstract: Assistive robot manipulators must be able to autonomously pick and place a wide range of novel objects to be truly useful. However, current assistive robots lack this capability. Additionally, assistive systems need to have an interface that is easy to learn, to use, and to understand. This paper takes a step forward in this direction. We present a robot system comprised of a robotic arm and a mobility scooter that provides both pick-and-drop and pick-and-place functionality for open world environments without modeling the objects or environment. The system uses a laser pointer to directly select an object in the world, with feedback to the user via projecting an interface into the world. Our evaluation over several experimental scenarios shows a significant improvement in both runtime and grasp success rate relative to a baseline from the literature [5], and furthermore demonstrates accurate pick and place capabilities for tabletop scenarios.

Keywords: Grasping and Manipulation, Robot Vision, AI-enabled Robotics
Abstract: Object search -- the problem of finding a target object in a cluttered scene -- is essential to solve for many robotics applications in warehouse and household environments. However, cluttered environments entail that objects often occlude one another, making it difficult to segment objects and infer their shapes and properties. Instead of relying on the availability of CAD or other explicit models of scene objects, we augment a manipulation planner for cluttered environments with a state-of-the-art deep neural network for shape completion as well as a volumetric memory system, allowing the robot to reason about what may be contained in occluded areas. We test the system in a variety of tabletop manipulation scenes composed of household items, highlighting its applicability to realistic domains. Our results suggest that incorporating both components into a manipulation planning framework significantly reduces the number of actions needed to find a hidden object in dense clutter.

Keywords: Grasping and Manipulation, Robot Vision
Abstract: A robot-assisted feeding system must successfully acquire many different food items. A key challenge is the wide variation in the physical properties of food, demanding diverse acquisition strategies that are also capable of adapting to previously unseen items. Our key insight is that items with similar physical properties will exhibit similar success rates across an action space, allowing the robot to generalize its actions to previously unseen items. To better understand which skewering strategy works best for each food item, we collected a dataset of 2450 robot bite acquisition trials for 16 food items with varying properties. Analyzing the dataset provided insights into how the food items' surrounding environment, fork pitch, and fork roll angles affect bite acquisition success. We then developed a bite acquisition framework that takes the image of a full plate as an input, segments it into food items, and then applies our Skewering-Position-Action network (SPANet) to choose a target food item and a corresponding action so that the bite acquisition success rate is maximized. SPANet also uses the surrounding environment features of food items to predict action success rates. We used this framework to perform multiple experiments on uncluttered and cluttered plates. Results indicate that our integrated system can successfully generalize skewering strategies to many previously unseen food items.

Keywords: Grasping and Manipulation, Soft Robotics
Abstract: Suction cups are the most common manipulation effectors in industry, but they are mostly only used for the purpose of pick-and-place. This paper proposes to use suction cups for a wider variety of tasks with the help of external contacts. A major hurdle in improving the dexterity of suction cups is the challenge of soft material modeling. Model error causes inaccurate contact location estimation as well as undesirable control performance. In this work, we propose a general framework for manipulation with suction cups under external contacts. The solution consists of a locally linear force-deformation model for suction cups with large deformation, and an estimation-control framework which utilizes contact constraints and feedback control to counter modeling errors. We verify the efficacy of our method experimentally by tilting a block on a table with a suction cup. Our method works reliably under modeling error even under large suction cup deformation (over 40 degree of bending). We also show the superiority of suction cups by performing tasks that are not possible with any normal fingertip.

Keywords: Grasping and Manipulation
Abstract: We present a method for fast and accurate physics-based predictions during non-prehensile manipulation planning and control. Given an initial state and a sequence of controls, the problem of predicting the resulting sequence of states is a key component of a variety of model-based planning and control algorithms. We propose combining a coarse (i.e. computationally cheap but not very accurate) predictive physics model, with a fine (i.e. computationally expensive but accurate) predictive physics model, to generate a hybrid model that is at the required speed and accuracy for a given manipulation task. Our approach is based on the Parareal algorithm, a parallel-in-time integration method used for computing numerical solutions for general systems of ordinary differential equations. We adapt Parareal to combine a coarse pushing model with an off-the-shelf physics engine to deliver physics-based predictions that are as accurate as the physics engine but run in substantially less wall-clock time, thanks to parallelization across time. We use these physics-based predictions in a model-predictive-control framework based on trajectory optimization, to plan pushing actions that avoid an obstacle and reach a goal location. We show that with hybrid physics models, we can achieve the same success rates as the planner that uses the off-the-shelf physics engine directly, but significantly faster. We present experiments in simulation and on a real robotic setup. Videos are available here: https://youtu.be/5e9oTeu4JOU

Keywords: Grasping and Manipulation
Abstract: Although point contact models are ubiquitous for robot grasp planning, they do not model the range of wrenches that finite-area soft contacts provide. This approximation leads to many false negatives. To reduce these, we propose REACH, a Robust Efficient Area Contact Hypothesis model. We consider its potential benefits and investigate two potential drawbacks: increased computational complexity and increased false positives. The REACH model computes the contact profile using constructive solid geometry intersection and barycentric integration and estimates the contact's ability to resist external wrenches (e.g., gravity) under perturbations in object pose and material properties. We evaluate the performance of REACH with 2,625 physical grasps of 21 diverse objects with an ABB YuMi robot. We compare performance of a soft point contact model, an elliptical area contact model, and a rigid-body dynamic simulation model using NVIDIA Flex. The REACH model reduces false negatives by 17% compared to the point contact model, achieving 72% average recall. The REACH model also compares favorably to full dynamic simulation in Flex and is two orders of magnitude faster, with 50~ms average computation time. Experimental data and supplementary material are available at https://sites.google.com/berkeley.edu/reach.

Keywords: Grasping and Manipulation