Keywords: Principles and technologies for anthropomorphic/bionic design and control, Humanoids for human science and engineering, Software and hardware architecture, system integration
Abstract: We have developed and studied musculoskeletal humanoids. Our goal is to realize a more human-like humanoid as a real human simulator, which has the same muscle and joint arrangements as human's and can do natural and dynamic motions as well as humans. Especially, it is very challenging to design musculoskeletal structure which can contain a large number of high powered muscles. Now, we design new fullbody musculoskeletal humanoid Kenshiro. This paper presents the concepts of this new robot and also shows the outline of our latest results Kenshiro, which is the succeeding version of our previous robot Kojiro.

Keywords: Novel materials, devices, mechanisms, energy system for humanoids, Principles and technologies for anthropomorphic/bionic design and control
Abstract: One of the motivations behind the development of humanoid robots is the will to comply with, and fruitfully integrate in the human environment, a world forged by humans for humans, where the importance of the hand shape dominates prominently. This paper presents the novel hand under-actuation frame- work which goes under the name of synergies. In particular two incarnations of this concept are considered, soft synergies and adaptive synergies. They are presented and their substantial equivalence is demonstrated. After this, it presents the first implementation of THE UNIPI-hand, a prototype which conciliates the idea of adaptive synergies for actuation with an high degree of integration, in a humanoid shape. The hand is validated experimentally through some grasps and measurements. Results are reported also in the attached video.

Keywords: Principles and technologies for anthropomorphic/bionic design and control, Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics
Abstract: In this paper, we propose a design of a robotic ankle-foot complex based on the human functional-anatomic ankle-foot structure. The proposed foot consists of three links, two joints, and four plantar muscles, whose mechanical stiffness can be controlled by utilizing McKibben pneumatic actuators. With this structure, a deformable medial longitudinal arch in a human foot can be emulated. We developed a musculoskeletal biped robot to which the proposed feet are implemented and measured its walking motion, especially the deformation of the robot foot. It is found that the foot generates a truss mechanism and a windlass mechanism, which are important functions of a human foot for shock absorption and energy storage and reuse. We also conducted a walking experiment with various parameters of a plantar muscle’s tonus to see how the tonus affects to ground reaction forces (GRFs) and its walking behavior. It is found that the GRF had two peaks as well as human walking and the shape of the GRF curve changes according to the tonus of the plantar muscle. We analyzed the impulse of GRF, finding out that a truss mechanism and a windlass mechanism works effectively with appropriate tonus of the plantar aponeurosis.

Keywords: Neuro-robotics and brain-robot interfaces for humanoids and humans, Cyborgs, prostheses, assistive devices and sensor/motor suits
Abstract: In this paper, we introduce our attempt to develop an assistive robot system which can contribute to Brain-Machine Interface (BMI) rehabilitation. For the BMI rehabilitation, we construct a Electroencephalogram(EEG)-Exoskeleton robot system, where the exoskeleton robot is connected to the EEG system so that the users can control the exoskeleton robot by using their brain activities. We use a classification method which considers covariance matrices of measured EEG signals as inputs to decode brain activities. The decoded brain activities are used to control exoskeleton movements. In this study, we consider assisting the stand-up movement which is one of the most frequently appeared movements in daily life and also a standard movement as a rehabilitation training. To assist the stand-up movement, we develop a force control model which takes dynamics of tendon string into account for the pneumatic-electric hybrid actuation system used in our exoskeleton robot. The results show that the exoskeleton robot successfully assisted user stand-up movements, where the assist system was activated by the decoded brain activities.

Keywords: Whole-body dynamics, control, sensing, informatics, Humanoid locomotion, manipulation, perception, planning
Abstract: Simplified models such as the inverted pendulum model are often used in humanoid robot control because the full dynamics model of humanoid robots is too complex to design a controller. These models are usually derived from simple mechanical systems that represent the essential properties of the robot dynamics. This method for deriving simplified models is a manual process that heavily relies on the controller developer's intuition. Moreover, mapping the state and input between the original and simplified models requires model-specific code. This paper describes a general method for systematically obtaining simplified models of humanoid robots. We demonstrate an application of derived models to humanoid robot balance control using linear quadratic regulators.

Keywords: Whole-body dynamics, control, sensing, informatics, Humanoid locomotion, manipulation, perception, planning
Abstract: We propose in this paper a general analytic scheme based on Gauss’ principle of least constraint for the derivation of the Lagrangian dynamics equation of motion of arbitrarily parameterized free-floating-base articulated mechanisms. The free-floating base of the mechanism is a non-actuated rigid object evolving in the 6D Lie group SE(3), the SO(3) component of which can be parameterized using arbitrary coordinate charts with equality constraints, for instance unit quaternions (also known as Euler parameters). This class of systems includes humanoid robots, and the presented formalism is particularly suitable for the whole-body dynamics modeling and control problem of such humanoid systems. Example motions of humanoid in arbitrary contact states with the environment demonstrate the originality of the approach.

Keywords: Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics, Human and humanoid skills/cognition/interaction
Abstract: This paper presents a generic stabilization framework which is applicable for both compliant and stiff humanoids. The proposed control framework is applied to the passive compliant humanoid robot COMAN which is equipped with series elastic actuators. The stabilization control framework combines the compliance control and the intrinsic angular momentum modulation to achieve an agile and compliant interaction against external perturbations. The admittance based compliance control uses the force/torque sensing in both feet to regulate the active compliance for the position controlled system. The physical elasticity in the new full body COMAN is exploited for the reduction and absorption of the instantaneous impacts while the admittance control further dissipates the excessive elastic energy. The angular momentum controller reduces the overall inertia effect for providing more rapid reactions. Both the theoretical work and experimental validation were presented. The effectiveness of the control scheme is demonstrated by COMAN's capabilities of withstanding various types of perturbations applied over the body, balancing on a moving platform and stabilizing while walking. Experimental data of the ground reaction force/torque, center of mass references and estimations, and the stored elastic energy are presented and analyzed.

Keywords: Whole-body dynamics, control, sensing, informatics, Humanoid locomotion, manipulation, perception, planning, Measuring, modeling and simulating humans
Abstract: This paper presents a modification for a broad class of controllers based on the LIPM dynamics. We use a change of variables such that instead of controlling the center of mass, we control an “augmented center of mass”, which is unaffected by upper body angular accelerations. Instead, we use upper body orientation as an additional source of control authority, allowing us to use both upper body rotation and center of pressure modulation for control. We demonstrate an improved robustness to external pushes with this additional control authority through simulated standing and walking experiments. We also demonstrate the modified controller on our force-controlled humanoid robot.

Keywords: Adaptation, learning and cognitive development of humanoids, Human and humanoid skills/cognition/interaction
Abstract: Present formulations of periodic dynamic movement primitives (DMPs) do not encode the transient behavior required to start the rhythmic motion, although these transient movements are an important part of the rhythmic movements (i.e. when walking, there is always a first step that is very different from the subsequent ones). An ad-hoc procedure is then necessary to get the robot into the periodic motion. In this contribution we present a novel representation for rhythmic Dynamic Movement Primitives (DMPs) that encodes both the rhythmic motion and its transient behaviors. As with previously proposed DMPs, we use a dynamical system approach where an asymptotically stable limit cycle represents the periodic pattern. Transients are then represented as trajectories converging towards the limit cycle, different trajectories representing varying transients from different initial conditions. Our approach thus constitutes a generalization of previously proposed rhythmic DMPs. Experiments conducted on the humanoid robot ARMAR-III demonstrate the applicability of the approach for movement generation.

Keywords: Adaptation, learning and cognitive development of humanoids, Human and humanoid skills/cognition/interaction, Applications: home, field, space, social, industrial, medical,health/mental care, art/entertainment, education
Abstract: In this paper we present a new methodology for robot learning that combines ideas from statistical generalization and reinforcement learning. First we apply statistical generalization to compute an approximation for the optimal control policy as defined by training movements that solve the given task in a number of specific situations. This way we obtain a manifold of movements, which dimensionality is usually much smaller than the dimensionality of a full space of movement primitives. Next we refine the policy by means of reinforcement learning on the approximating manifold, which results in a learning problem constrained to the low dimensional manifold. We show that in some situations, learning on the low dimensional manifold can be implemented as an error learning algorithm. We apply golden section search to refine the control policy. Furthermore, we propose a reinforcement learning algorithm with an extended parameter set, which combines learning in constrained domain with learning in full space of parametric movement primitives, which makes it possible to explore actions outside of the initial approximating manifold. The proposed approach was tested for learning of pouring action both in simulation and on a real robot

Keywords: Adaptation, learning and cognitive development of humanoids, Human and humanoid skills/cognition/interaction, Humanoid locomotion, manipulation, perception, planning
Abstract: Learning the visual appearance and physical properties of unknown objects is an important capability for humanoid robots that are supposed to be working in an open environment. We present an approach that enables a robot to discover new, unknown objects, segment them from the background and grasp them. This gives the robot full control over the object and allows its further multimodal exploration.

In order to discover an unknown object in a cluttered scene and segment it from the (likewise unknown) background, we generate hypotheses based on visual input and try to verify one of them by pushing it. The induced motion solves visual ambiguities and allows a clear object-background segmentation.

The acquired estimation of the object position and extent allows the robot to try grasping it. As we do not have exact shape information, we apply a reactive grasping approach. Based on tactile sensor feedback of the hand, we execute correction movements until the object can be grasped in a stable manner.

Keywords: Adaptation, learning and cognitive development of humanoids, Social interaction and acceptability, Measuring, modeling and simulating humans
Abstract: Culture is not the first aspect that comes to mind when discussing human robot interaction. But our cultural upbringing does to a large degree influence our patterns of behavior and interpretation. Thus, culture is present in the development of robotic systems right from the start, unconsciously influencing how robots look, what we envision with them to do, and how they are programmed to interact with the user. In this paper we argue that is is beneficial to make this unconscious influence explicit and take it into account during the development (and evaluation) of humanoid robots. To this end we present a principled approach of capturing various cultural influences during the development process of humanoid robots and exemplify this approach with a case study of affective body movements.

Keywords: Adaptive and Learning Humanoids, Skill Acquisition and Learning, Healthcare Humanoids
Abstract: Automation of neurologic rehabilitation becomes more important as the number of age correlated dysfunctions increases. Starting from existing Assist-As-Needed control schemes we designed new algorithms specifically for bilateral tasks. We propose three approaches for controlling an arm rehabilitation device that assists hemiparetic subjects. Design goals are 1) Users should use there healthy arm to generate the desired movement pattern. 2) Force contribution and freedom of movement should be as high as possible for the affected arm. 3) Large errors of the affected arm must be compensated by the device. Simulations and first clinical evaluations support the following results. Goal 1) is achieved by making support independent of time and a set trajectory. Goals 2) and 3) can be achieved with an iterative learning controller that learns a supportive force. Adaptation speed and error sensitivity can be improved with model based approaches. Frequency dependent shifting of the force profile shifts the acceleration to the movement phase in which it is needed resulting in earlier acceleration and deceleration. An additional stiffness should only be used, if the patient's behavior varies to much.

Keywords: Biped Systems, Motion Planning, Walking Patterns, and CPG, Passive Walking
Abstract: This paper investigates the stability principle underlying limit cycle walking of an underactuated biped and proposes the method for generating fast-convergent bipedal gaits. First, we revisit our previous results and simplify the transition function for the state error of the stance phase. The simplified function without including step period is written as the ratio of the steady angular velocities. Second, we apply a discrete-time output deadbeat control to biped gait generation. We mathematically show that the output deadbeat control has the potentiality to achieve fast-convergent gait generation. The validity of the proposed method is investigated through numerical simulations.

Keywords: Grasping and Manipulations, Skill Acquisition and Learning, Applications
Abstract: Autonomous snap assemblies is a highly desirable robotic functionality. While much work has been done in active sensing for peg-in-hole assemblies and general compliant motions, snap assembly state estimation remains an open research problem. This work presents a probabilistic framework designed to evaluate, based on belief states, if each state in the assembly as well as the overall result succeeded or not. In this work, an anthropomorphic robot HIRO performed a cantilever-snap assembly using the Pivot Approach control strategy and our snap verification system. The latter used a Bayesian Filter within its hierarchical taxonomy yielding belief states at two levels of the taxonomy. The last layer of the system, derived thresholds from training data to assess the test assembly outcomes. The framework was effective in correctly assessing the outcome of all test assemblies.

Keywords: Human and humanoid skills/cognition/interaction
Abstract: In this paper we address the problem of recognizing everyday sound events in indoor environments with a consumer robot. Sounds are represented in the spectro-temporal domain using the stabilized auditory image (SAI) representation. The SAI is well suited for representing pulse-resonance sounds and has the interesting property of mapping a time-varying signal into a fixed-dimension feature vector space. This allows us to map the sound recognition problem into a supervised classification problem and to adopt a variety of classifications schemes. We present a complete system that takes as input a continuous signal, splits it into significant isolated sounds and noise, and classifies the isolated sounds using a catalogue of learned sound-event classes. The method is validated with a large set of audio data recorded with a humanoid robot in a house. Extended experiments show that the proposed method achieves state-of-the-art recognition scores with a twelve-class problem, while requiring extremely limited memory space and moderate computing power. A first real-time embedded implementation in a consumer robot show its ability to work in real conditions.

Keywords: Human and humanoid skills/cognition/interaction, Adaptation, learning and cognitive development of humanoids, Applications: home, field, space, social, industrial, medical,health/mental care, art/entertainment, education
Abstract: Manipulation tasks are characterized by continuous motion trajectories containing a set of key phases. In this paper, we propose a probabilistic method to autonomously segment the motion trajectories for estimating the key phases embedded in such a task. The autonomous segmentation process relies on principal component analysis to adaptively project into one of the low-dimensional subspaces, in which a Gaussian mixture model is learned based on Bayesian information criterion and expectation-maximization algorithms. The basis skills are estimated by a set of Gaussians approximating quasi-linear key phases, and those times spent calculated from the segmentation points between two consecutive Gaussians representing the local changes of dynamics and directions of the trajectories. The basis skills are then used to build novel motion trajectories with possible motion alternatives and optional parts. After sequentially reorganizing the basis skills, a Gaussian mixture regression process is used to retrieve smooth motion trajectories. Two experiments are presented to demonstrate the capability of the autonomous segmentation approach.

Keywords: Human and humanoid skills/cognition/interaction, Adaptation, learning and cognitive development of humanoids, Social interaction and acceptability
Abstract: This paper presents an experiment in which the iCub humanoid robot learns to recognize faces through proprioceptive information. We take inspiration in the way blind people recognize people's faces, i.e. through tactile exploration of the person's face. The iCub robot's tactile sensors are used to provide compliance in the hand motion so as to smoothly scan the facial features. The displacement of the fingers, as the robot explores the face, is used to build a model of the face using Hidden Markov Models. We show that the robot can successfully distinguish across the faces of a standard doll and the faces of three humanoid robots, the HOAP-3 robot, a Robota doll robot and MyDreamBaby, a commercial robotic doll.

Keywords: Human and humanoid skills/cognition/interaction, Humanoids for human science and engineering, Social interaction and acceptability
Abstract: In this paper we present a new system for on-line audio-visual detection of speaking persons using the humanoid robot NAO. The proposed method is based on multimodal statistical signal processing and implemented on top of a platform-independent middleware. The system is able to process the binocular visual data and the binaural auditory data at 17Hz. The adaptations of an audio-visual fusion model done to port the algorithm to Nao is given together with the implementation details. We obtained good performance results in different scenarios recorded in a room with no special acoustic or lighting properties or arrangements.

Keywords: Humanoid locomotion, manipulation, perception, planning
Abstract: This work presents a Central Pattern Generator approach where the designer is able to build a basic motor repertoire that enables a biped robot to walk. The presented locomotor system includes a phase regulation feedback, which elicits or delays the transitions between swing and stance step phases according to load sensory information, adjusting the nominal walk to the environment.

The approach is tested on two simulated humanoid robots and in a DARwIn-OP robot, achieving a multitude of locomotor tasks, showing how general the proposed locomotor system can be. Simulations and experiments also demonstrate the role of phase regulation on addressing small external perturbations.

Keywords: Humanoid locomotion, manipulation, perception, planning
Abstract: A bipedal robot should be robust and able to move in various directions on slope or stairs. However, up to date many research studies have been focussing on walking in the up or down direction only. Therefore, a strategy to realize walking along slope is investigated. In conventional methods, CoM is moved up and down during walking in this situation. In this paper, a method named as dual length linear inverted pendulum method with Newton-Raphson is proposed in order to move CoM always in horizontal. Different lengths of pendulums are applied at left and right legs in order to represent the CoM height. By using the proposed method, maximum impact forces are reduced as verified via the simulation results.

Keywords: Humanoid locomotion, manipulation, perception, planning, Adaptation, learning and cognitive development of humanoids
Abstract: In a variety of problem domains, such as math and motion planning, humans use a dual strategy of generation and recall to find solutions. `Generation' uses production rules and models to search for novel solutions to novel problems, whereas `recall' reuses previously found solutions for similar previously encountered problems. As we expect the advantages of this dual strategy to carry over to the robotics domain, we compare and evaluate generation and recall strategies for motion planning on a set of reaching tasks. The specific implementations we use are the lazy variant of the Rapidly-exploring Random Trees and Dynamic Movement Primitives, and we compare these two methods on the commercially available REEM robot. Quantifying the differences and advantages of these methods constitutes is required to make informed decisions about which approach is most suitable for which application domain and task contexts.

This paper is accompanied by a video: http://youtu.be/QJy9QiIplE0

Keywords: Humanoid locomotion, manipulation, perception, planning, Others (not listed above)
Abstract: We propose an odometric system for localizing a walking humanoid robot using standard sensory equipment, i.e., a camera, an Inertial Measurement Unit, joint encoders and foot pressure sensors. Our method has the prediction-correction structure of an Extended Kalman Filter. At each sampling instant, position and orientation of the torso are predicted on the basis of the differential kinematic map from the support foot to the torso, using encoder data from the support joints. The actual measurements coming from the camera (head position and orientation reconstructed by a V-SLAM algorithm) and the Inertial Measurement Unit (torso orientation) are then compared with their predicted values to correct the estimate. The filter is made aware of the current placement of the support foot by an asynchronous update mechanism triggered by the pressure sensors. An experimental validation on the humanoid NAO shows the satisfactory performance of the proposed method.

Keywords: Humanoid locomotion, manipulation, perception, planning, Others (not listed above), Measuring, modeling and simulating humans
Abstract: While humans are highly efficient in dealing with balance perturbations, current biped humanoid robots are far from showing similar skills. This is mainly due to the limited capacity of current robot controllers to deal with the inherently complex dynamics of biped robots. Though Model Predictive Control schemes have shown improved robustness to perturbations, they still suffer from a few shortcomings such as not considering the upper body inertial effects or non-optimal step durations. We propose here a Model Predictive Control scheme that specifically addresses these shortcomings and generates human-like responses to perturbations, involving appropriate combinations of ankle, hip and stepping strategies, with automatically adjusted step durations.

Keywords: Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics
Abstract: Planning collision-free, dynamically-stable movements for humanoid robots is a challenging problem. An effective approach consists of first planning a motion satisfying geometric and kinematic constraints (such as collision avoidance, joint angle limits, velocity limits, etc.) and, in a second stage, modifying this motion so that it respects dynamic stability criteria, such as those relative to the Zero Moment Point (ZMP). However, this approach currently suffers from the issue that the modified motion may give rise to new collisions with respect to the original motion, which can be very costly to deal with, especially for systems with many degrees of freedom and cluttered environments. Here we present an algorithm to modify the motions of humanoid robots under ZMP constraints without changing the original motion path, making thereby new collision checks unnecessary. We do so by adapting the minimum-time path parameterization under torque constraints algorithm of Bobrow et al. to the case of ZMP constraints. In contrast with a previous approach based on finite differences and iterative optimization to find the optimal path parameterization under ZMP constraints, our Bobrow-based algorithm finds this optimal parameterization in a single pass. We demonstrate the efficiency of this algorithm by simulations.

Keywords: Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics
Abstract: This paper proposes a balance control method that consists of an imaginary full-actuated controller and an input transformation method. The controller is designed for maintaining the balance of bipeds under the assumption that bipeds have a direct control input to the torso. This controller has a symmetric structure that is independent of the stance phase and includes internal impedance of the chest and hip to increase the controller’s robustness with respect to external disturbances. The transformation method converts the full-actuated controller to a realizable under-actuated controller and guarantees that the power provided by the under-actuated controller equals that by the full-actuated controller in the case that bipeds contact the environment at multiple points. Simulations were executed to validate the proposed input transformation and the controller to clarify the effects of the mass distribution models and internal impedance on balance.

Keywords: Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics
Abstract: Typically, gait pattern generation for bipedal robots utilizes a simplified model, neglecting the rate of change of the angular momentum. In this paper a linear parameter optimization problem for this simplified model, extended by an estimation of the angular momentum, is proposed to generate stable walking patterns that can be solved online and still consider the nonlinear effects of the multi-body model. The scheme was used successfully to generate trajectories for a full size humanoid robot.

Keywords: Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics, Humanoids for human science and engineering
Abstract: In this paper, we demonstrate methods for bipedal walking control based on the Capture Point (CP) methodology. In particular, we introduce a method to intuitively derive a CP reference trajectory from the next three steps and extend the linear inverted pendulum (LIP) based CP tracking controller (CPT) introduced in Englsberger et al. (IROS 2011), generalizing it to a model that contains vertical CoM motions and changes in angular momentum. Respecting the dynamics of general multibody systems, we propose a measurement-based compensation of multi-body effects, which leads to a stable closed-loop dynamics of bipedal walking robots. In addition we propose a ZMP projection method, which prevents the robots feet from tilting and ensures the best feasible CP tracking. The extended CP controller's performance is validated in OpenHRP3 simulations and compared to a previous version of the controller proposed in Englsberger et al. (IROS 2011).

Keywords: Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics, Principles and technologies for anthropomorphic/bionic design and control
Abstract: We propose a method for resolving a hybrid task description consisting of generalized forces and task-space trajectories into joint accelerations. The method is motivated by the problem of biped walking control for which a natural task description consists of contact forces and task-space trajectories. We present both a general method based on the full equations of motion of the multibody system and a more efficient approach tailored to the problem of biped walking control. We analyze the method and compare it to the previously proposed dynamics error compensation method, which can reduce the planning error resulting from simplified robot models. The controller is evaluated in walking experiments and compared to a different walking controller using an inverse kinematics algorithm. The method introduced in this paper allows us to exactly track the desired contact forces in real-time.

Keywords: Humanoids for human science and engineering, Human and humanoid skills/cognition/interaction, Applications: home, field, space, social, industrial, medical,health/mental care, art/entertainment, education
Abstract: We present a real-time algorithm that segments unstructured and highly cluttered scenes. The algorithm robustly separates objects of unknown shape in congested scenes of stacked and partially occluded objects. The model-free approach finds smooth surface patches, using a depth image from a Kinect camera, which are subsequently combined to form highly probable object hypotheses. The real-time capabilities and the quality of the algorithm are evaluated on a benchmark database. Advantages compared to existing approaches as well as weaknesses are discussed. We also report on an autonomous grasping experiment with the Shadow Robot Hand which employs the estimated shape and pose of objects given by our algorithm in a task in which it cleans a table.

Keywords: Novel materials, devices, mechanisms, energy system for humanoids, Cyborgs, prostheses, assistive devices and sensor/motor suits, Human and humanoid skills/cognition/interaction
Abstract: This paper introduces a novel, fabric-based, flexible, and stretchable tactile sensor, capable of seamlessly covering natural shapes. Our design allows for several tactile cells to be embedded in a single sensor patch, and can have an arbitrary perimeter and can cover freeform surfaces. The sensor remains operational on top of soft padding, facilitating the possibility to build human-like artificial skin. It provides force measurements from subtle to high forces (0.1--30N), which easily covers the common range for everyday human manual interactions. Due to a layered construction, the sensor is very robust and can withstand huge normal forces without sustaining damage. We discuss the construction of the sensor and evaluate its performance. As an exciting application for the sensor, we describe the construction of a wearable tactile dataglove with 54 tactile cells.

Keywords: Principles and technologies for anthropomorphic/bionic design and control, Human and humanoid skills/cognition/interaction, Humanoid locomotion, manipulation, perception, planning
Abstract: The ability to manipulate deformable objects, such as textiles or paper, is a major prerequisite to bringing the capabilities of articulated robot hands closer to the level of manual intelligence exhibited by humans. We concentrate on the manipulation of paper, which affords us a rich interaction domain that has not yet been solved for anthropomorphic robot hands. Robust tracking and physically plausible modeling of the paper as well as feedback based robot control are crucial components for this task. This paper makes two novel contributions to this area. The first concerns real-time modeling and visual tracking. Our technique not only models the bending of a sheet of paper, but also paper crease lines which allows us to monitor deformations. The second contribution concerns enabling an anthropomorphic robot to fold paper, and is accomplished by introducing a set of tactile- and vision-based closed loop controllers.

Keywords: Principles and technologies for anthropomorphic/bionic design and control, Superhuman humanoids / future humanoids: possibility, technology and meaning, Others (not listed above)
Abstract: In this paper, we propose a method to determine the optimal multiarticular muscle arrangement of a robot for a specific motion. Multiarticular muscles contribute to the realization of a musculoskeletal robot with dynamic motion ability similar to humans and animals. However, the ability is affected by the multiarticular muscle arrangement. In previous researches, the muscle arrangements of robots were based on those of humans and animals. However, we suppose that a more suitable muscle arrangement than any other one of humans and animals can be found by limiting the variety of motions. Therefore, we focused on a specific robot motion, and furthermore, searched for the optimal multiarticular muscle arrangement for the specific motion. Specifically, we estimated the impulse properties of robots with a great variety of possible muscle arrangements. The impulse properties were calculated with dynamics simulation corresponding to the specific motion. Then, on the basis of the evaluation of the impulse properties of the robots, we searched for the optimal muscle arrangement from among 736,281 of possible muscle arrangements. Furthermore, we demonstrated the effectiveness of our method using simulation experiments. The simulation results indicate that there is a more suitable muscle arrangement than any other one of humans and animals on condition that we focused on the specific motion.

Keywords: Sensing, Sensors and Actuators, Kinematics, Dynamics, and Control, Biomimetics
Abstract: In this paper we suggest basic principles to design a novel type of passive variable impedance actuator aimed at replicating a specific property of human co-contraction, related to the ability to cope with uncertainties affecting any physical/biological system. In particular the dynamical model of the proposed actuator is such that the variance of the state vector in response to noisy disturbances can be reduced by tuning the passive stiffness of the system. By means of a linearization analysis we characterize the mathematical properties that a non-linear dynamical system should have in order to possess this noise rejection property. We provide a practical example of such a system based on non-linear springs whose critical feature is to attach some elastic elements to a fixed reference (e.g. "ground"). We then show that this antagonist actuator structure is actually analog to Hill’s model of the human muscle/tendon system, emphasizing its biological relevance. We finally illustrate how time-varying stiffness can be efficiently planned feedforwardly to reject disturbances that may affect task achievement. To this aim, we use the formalism of stochastic optimal control to derive open-loop controls anticipating the consequences of unpredictability and instability linked to the task. We conclude that the suggested actuators are well-suited to mimic the main features of human co-contraction and plan to implement this type of actuator on the robot platform iCub in a near future.

Keywords: Skill Acquisition and Learning, Motion Planning, Walking Patterns, and CPG, Learning from/by Demonstration
Abstract: A service robot has to be flexible and fast in order to solve a manipulation task in the human environment with differing start configurations, objects, obstacles and a restricted work space. Based on a sophisticated task model, which captures the relevant constraints and goals of a task, constrained motion planning can be used to generate robot trajectories autonomously with high generalization capabilities. The major drawbacks are high planning times and non-repeatability of the results. In this work, search heuristics, which restrict the search space during motion planning, are learned incrementally whenever the robot uses the task model to plan a solution. The number of learned search heuristics is restricted by using a combination of constrained motion planning and a fast local control algorithm to increase the number of situations, in which the search heuristic can be applied. The approach combines two major approaches in Programming by Demonstration (PbD), i.e. learning and goal-directed planning with a general task description and learning efficient encodings of low-level trajectories, in a consistent way.

Keywords: Social interaction and acceptability, Humanoid ethics and philosophy
Abstract: Familiar to everyone, expressions such as "familiar, to familiarize, familiarity" are commonly used in reviews on HRI. Nevertheless, "the familiar" is a notion that lacks a precise definition and only rare studies tackle it directly. This paper is a preliminary study on the human experience of the familiar when interacting with a robot. Which signals and elements trigger it? What are the mental and emotional states it associates with? How dependent is it on anthropomorphism and social acceptance? Can the familiar really be considered as a key-concept to study Human-Robot interaction and better understand what makes it a motivating, satisfying and socially adapted one? This research question was investigated in an experiment involving three protagonists: one robot, NAO, and two participants. The experiment was thought in a way allowing a real socially-engaging situation to occur. It investigates participants' response to an unannounced interaction with a robot consisting of two main actions: performing social greetings and handing the envelope containing the questionnaire that participants agreed to answer. The robot's behaviour vary from a participant to another.

Keywords: Social interaction and acceptability, Humanoids for human science and engineering, Humanoid ethics and philosophy
Abstract: This research explores the concept of “The Frankenstein Syndrome” in order to develop a psychological tool for measuring acceptance of humanoid robots, as well as expectations and anxieties toward these technologies. It also aims to explore factors that influences this acceptance, such as the effects of age and experiences with robots. The tool developed is tentatively named “The Frankenstein Syndrome Questionnaire”, a questionnaire to measure acceptance of humanoids in humans. A preliminary survey conducted in Japan suggested that: 1) the elder population has positive expectations for the specific applications of humanoid robots in their daily lives and view the developers of these technologies more favourably, in comparison with the younger population; 2) these expectations can be made more positive through having experiences of real humanoid robots or experiences of the robots via media information; 3) such experiences of humanoid robots do not seem to affect general anxiety toward the robots in the younger population, but they do decrease the anxiety in the elder population; 4) experiences of humanoid robots increase apprehension toward risks the robots entail in the society in the younger population, although they do not affect such apprehension in the elder population. The paper discusses about implications from the results.

Keywords: Software and hardware architecture, system integration, Adaptation, learning and cognitive development of humanoids, Human and humanoid skills/cognition/interaction
Abstract: In this paper we describe a cognitive architecture for humanoids interacting with objects and caregivers in a developmental robotics scenario. The architecture is foundational to the MACSi project: it is designed to support experiments to make a humanoid robot gradually enlarge its repertoire of known objects and skills combining autonomous learning, social guidance and intrinsic motivation. This complex learning process requires the capability to learn affordances first. Here, we present the general framework for achieving these goals, focusing on the elementary action, perception and interaction modules. Preliminary experiments performed on the humanoid robot iCub are also discussed.

Keywords: Software and hardware architecture, system integration, Humanoid locomotion, manipulation, perception, planning
Abstract: Reducing the number of cables needed for the actuators and sensors of humanoid and other robots with high numbers of degrees of freedom (DoF) is a relevant problem, often solved by using a common bus for all communication, which may result in bandwidth limitation problems. This paper proposes an optimized method to re-order the commands sent to the joint-local controllers of a high DoF serial manipulator. The proposed method evaluates which local controller would benefit the most from an updated command given a cost function, and sends a command to this controller. As is demonstrated in both simulation and in experiments on a real robot, the resulting scheme can significantly improve system performance, equivalent to increasing the communication frequency by up to 3 times.

Keywords: Teleoperation, tele-experience, tele-presence using humanoids, Humanoid locomotion, manipulation, perception, planning
Abstract: We extend a recent low cost real-time method of hand tracking and pose estimation in order to control an anthropomorphic robot hand. The approach is data-driven and based on matching the current image of a color-gloved hand with the best fitting image in a database to retrieve the posture. Then, using depth information from a Kinect camera and a color-sensitive iterative closest point-to-triangle algorithm we can very accurately estimate the absolute position and orientation of the hand. The effectiveness of the approach is demonstrated in an application in which we actively control a 20 DOF anthropomorphic robot hand in a manual interaction grasping task.

Keywords: Teleoperation, tele-experience, tele-presence using humanoids, Humanoid locomotion, manipulation, perception, planning
Abstract: Robust manipulation with a dexterous robot hand is a grand challenge of robotics. Impressive levels of dexterity can be achieved through teleoperation. However, teleoperation devices such as a glove or force reflecting master-slave system can be expensive and can tie the robot down to a restricted workspace. We observe that inexpensive and widely available multi-touch interfaces can achieve excellent performance for a large range of telemanipulation tasks, making dexterous robot telemanipulation broadly accessible. Our key insight is that dexterous grasping and manipulation interactions frequently focus on precise control of the fingertips in a plane. Following this observation, our novel multi-touch interface focuses on reliable replication of planar fingertip trajectories, making previously difficult actions such as grasping, dragging, reorienting, rolling, and smoothing as intuitive as miming the action on a multi-touch surface. We demonstrate and evaluate these and other interactions using an iPad interface to a Shadow Hand mounted on a Motoman SDA10 robot.

Keywords: Whole-body dynamics, control, sensing, informatics, Humanoid locomotion, manipulation, perception, planning
Abstract: This paper presents an acceleration-based inverse dynamics method to control the floating base of a force-controlled bipedal robot. The desired accelerations of the floating base are derived by PD control in operational space and then used to calculate the accelerations of the joints. Given kinematic constraints to the feet, a relationship between the accelerations of the floating base and the desired external forces needed for those accelerations is obtained. The desired external forces are constrained by ZMP, friction and unilateral vertical forces, which introduces corresponding constraints on the accelerations. If the desired accelerations do not satisfy the constraints, quadratic programming is applied to determine optimal accelerations, which will satisfy the constraints. These optimal accelerations are used instead of the desired ones when calculating inverse dynamics. Our controller guarantees the desired external forces satisfy their constraints. The effectiveness of the proposed methods is demonstrated by tracking desired trajectories and recovering from disturbances on a force-controlled bipedal robot in simulation.

Keywords: Whole-body dynamics, control, sensing, informatics, Measuring, modeling and simulating humans
Abstract: This paper presents an efficient method for adaptive control of humanoid robot arms with biarticular muscles, which exhibits multiple beneficial properties. In our approach, sliding control was chosen to get joint torque first and then the joint torque was distributed to muscle forces. The muscle force was computed based on a Jacobian matrix between joint torque space and muscle force space. In addition, internal forces were used to optimize the computed muscle forces – and thus, two important benefits arise: through our proposed method, not only we are making sure that each muscle force stays within its predefined force boundary; but we are also enabling the muscles to work in the middle of their working range, which is considered to be an anti-fatigue state. Yet more benefit derived, comes from the fact that in our method all the dynamic parameters are updated in real-time, and can thus one can account for perturbations and disturbances during operation. Compared with previous work in parameter adaptation, a composite method was proposed which utilizes the prediction error to accelerate parameter convergence speed. Our method was tested for the case of reaching motions. The results clearly illustrate the benefits of the method.

Keywords: Whole-body dynamics, control, sensing, informatics, Software and hardware architecture, system integration, Humanoid locomotion, manipulation, perception, planning
Abstract: This paper addresses the problem of simulating the dynamics of a robot, which include sensors and actuators, in the context of robot grasping and manipulation. The simulated robot is equipped with tactile sensors using soft contacts and is actuated by angular motors. The simulation is tightly integrated with the real robot, which allows the transparent control of both robots with the same joints' velocity commands, and the perception of the same sensor readings.

In order to validate the simulation, three different types of manipulation experiments are performed and recorded, both on the real and on the simulated robots. Results show that the simulated robot replicates the real execution with sufficient accuracy, which enables the simulation tool to be used as a replacement for the real hardware for development and prototyping.

Keywords: Adaptation, learning and cognitive development of humanoids, Human and humanoid skills/cognition/interaction, Humanoid locomotion, manipulation, perception, planning
Abstract: Seeing is not just done by the eyes, it involves also the integration of other modalities such as auditory, proprioceptive and tactile information; e.g., to locate spatially objets, persons and our limbs. Accordingly, it is advanced that the neonate's perceptual system is broadly tuned to intersensory temporal synchrony and to intensity of multisensory signals which permit him to construct his coherent multimodal world. We hypothesize that the neural mechanism of gain-field modulation, found to process coordinate transform between modalities in the superior colliculus and in the parietal area, may play a key role to build his unified perceptual world. In experiments with a head-neck-eye's robot with a camera and microphones, we study how gain-field modulation in neural networkss can serve for transcribing one modality (e.g., audio) from its proper reference frame to another one centered on the body, the head or on the eye. This transformation permits to bind the modalities from each others in intermediate reference frames, which can influence in return the unisensory signals (multimodal enhancement). Sound signals, which are naturally mapped into the head coordinates, can be remapped in retina coordinates, in example, for audio-visual speech perception.

Keywords: Human and humanoid skills/cognition/interaction, Social interaction and acceptability, Adaptation, learning and cognitive development of humanoids
Abstract: Human-robot communication is a very important aspect in the field of Humanoid Robotics. Non-verbal communication is one of the components that make interaction natural. However, despite the most recent efforts, robots still can show only limited expression capabilities. The purpose of this work is to create a facial expression generator that can be applied to the new 24 DoF head of the humanoid robot KOBIAN-R. In this paper, we present a system based on relevant studies of human communication and facial anatomy, adapted to the specific robotic face. It makes use of polynomial classifiers and is able to produce over 600 thousands of facial cues combinations, together with appropriate neck movement. Results showed that the recognition rate of expressions produced by this system is comparable to the rate of recognition of the most common facial expressions. We conclude that our system can successfully improve the communication capabilities of the robot KOBIAN-R, and that there is potential for using it to implement more complex interaction.

Keywords: Humanoid locomotion, manipulation, perception, planning, Adaptation, learning and cognitive development of humanoids, Whole-body dynamics, control, sensing, informatics
Abstract: Movement primitives as basis of movement planning and control have become a popular topic in recent years. The key idea of movement primitives is that a rather small set of stereotypical movements should suffice to create a large set of complex manipulation skills. An interesting side effect of stereotypical movement is that it also creates stereotypical sensory events, e.g., in terms of kinesthetic variables, haptic variables, or, if processed appropriately, visual variables. Thus, a movement primitive executed towards a particular object in the environment will associate a large number of sensory variables that are typical for this manipulation skill. These association can be used to increase robustness towards perturbations, and they also allow failure detection and switching towards other behaviors. We call such movement primitives augmented with sensory associations Associative Skill Memories (ASM). This paper addresses how ASMs can be acquired by imitation learning and how they can create robust manipulation skill by determining subsequent ASMs online to achieve a particular manipulation goal. Evaluation for grasping and manipulation with a Barrett WAM/Hand illustrate our approach.

Keywords: Software and hardware architecture, system integration
Abstract: We present a new Interprocess Communication (IPC) mechanism and library. Ach is uniquely suited for coordinating drivers, controllers, and algorithms in complex robotic systems such as humanoid robots. Ach eliminates the Head-of-Line Blocking problem for applications that always require access to the newest message. Ach is efficient, robust, and formally verified. It has been tested and demonstrated on a variety of physical robotic systems, and we discuss the implementation on our humanoid robot Golem Krang. Finally, the source code for Ach is available under an Open Source permissive license.

Keywords: Adaptation, learning and cognitive development of humanoids
Abstract: Learning by imitation in humanoids is challenging due to the unpredictable environments these robots have to face during reproduction. Two sets of tools are relevant for this purpose: 1) probabilistic machine learning methods that can extract and exploit the regularities and important features of the task; and 2) dynamical systems that can cope with perturbation in real-time without having to replan the whole movement. We present a learning by imitation approach combining the two benefits. It is based on a superposition of virtual spring-damper systems to drive a humanoid robot's movement. The method relies on a statistical description of the springs attractor points acting in different candidate frames of reference. It extends dynamic movement primitives models by formulating the dynamical systems parameters estimation problem as a Gaussian mixture regression problem with projection in different coordinate systems. The robot exploits local variability information extracted from multiple demonstrations of movements to determine which frames are relevant for the task, and how the movement should be modulated with respect to these frames. The approach is tested on the new prototype of the COMAN compliant humanoid with time-based and time-invariant movements, including bimanual coordination skills.

Keywords: Applications: home, field, space, social, industrial, medical,health/mental care, art/entertainment, education, Human and humanoid skills/cognition/interaction, Software and hardware architecture, system integration
Abstract: We address the problem of real-time gesture recognition, and we prove that our system can be used in real scenarios presenting an original memory game; the object of the game is to perform the longest sequence of gestures that it is possible to remember. We explore the human-robot interaction field, letting the player confront a humanoid robot, iCub. Our main contribution is two-fold: on one hand, we present a robust and real-time gesture recognition system; on the other hand, we place the presented system in a real scenario, where its reliability and its effectiveness are remarkably stressed. This game has ranked 2nd at ChaLearn Kinect Demonstration Competition.

Keywords: Grand challenges, competitions, outreach, Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics
Abstract: We study the feasibility of having various humanoid robots undertake some tasks from those challenged by the DARPA's call on disaster operations. Hence, we focus on locomotion tasks that apparently require human-like motor skills to be achieved. We use virtual scenes under the fully-3D-modeled-environment assumption. The robot autonomously plans and executes the motion with a high-level goal specification, such as reaching a global position or a particular contact state. We assess the feasibility according not only to the robot kinematics, but also to whole-body dynamics, non-desired collision avoidance, friction limits, and actuation limits. The results --the controlled motions-- are demonstrated in the accompanying video.

Keywords: Human and humanoid skills/cognition/interaction, Applications: home, field, space, social, industrial, medical,health/mental care, art/entertainment, education, Social interaction and acceptability
Abstract: The recognition of emotions and the generation of appropriate responses is a key component for facilitating more natural human-robot interaction. Music, often called the ``language of emotions," is a particularly useful medium for investigating questions involving the expression of emotion. Likewise, movements and gestures, such as dance, can also communicate specific emotions to human observers. We apply an efficient, causal technique for estimating the emotions (mood) from music audio to enable a humanoid to perform gestures reflecting the musical mood. We implement this system using Hubo, an adult-sized humanoid that has been used in several applications of musical robotics. Our preliminary experiments indicate that the system is able to produce dance-like gestures that are judged by human observers to match the perceived emotion of the music.

Keywords: Human and humanoid skills/cognition/interaction, Humanoid locomotion, manipulation, perception, planning
Abstract: In this paper we introduce an online grasping system for the Nao robot [1] manufactured by Aldebaran Robotics. The proposed system consists of an object detector and a grasp motion planner. Thereby, known objects are detected by a stereo contour-based object detector and hand motion paths are planned by an A*-based algorithm while avoiding obstacles. Compared to skilled robots such as Justin [2] or ASIMO [3] online grasping with the Nao constitute as particular problem due to the limited processing power and the hand design. The methods proposed allow detecting and grasping objects in realtime on an affordable humanoid robot.

Keywords: Human and humanoid skills/cognition/interaction, Humanoid locomotion, manipulation, perception, planning, Adaptation, learning and cognitive development of humanoids
Abstract: Robotic researchers face fundamental challenges when designing autonomous humanoid robots, which are able to interact with real dynamic environments. In such unstructured environments, the robot has to autonomously segment objects, detect and categorize relevant situations, decide when to initiate and terminate actions. As humans are very good in these tasks, inspiration from models of human sensory-motor and cognitive processes may help design more flexible and autonomous robotic control architectures. Recently, we have extended a neurally inspired model for sequential organization with a representation of hierarchies of behaviors. Here, we implement this model on a robotic platform and demonstrate its functionality under constraints of a real-world implementation. The architecture generates hierarchically organized behavioral sequences on the Aldebaran’s humanoid robot NAO. The key dynamic components of serial organization – such as the intention, condition of satisfaction (CoS), and interactions within the hierarchy – are coupled to robotic sensors and motors and bring about flexible and autonomous behavior. We also demonstrate how continuous in time neural-dynamic parts of the controller may be seamlessly integrated with preprogramed algorithmic behaviors, introducing flexibility, autonomy, and ability to learn, while avoiding unnecessary complexity of the architecture.

Keywords: Human and humanoid skills/cognition/interaction, Measuring, modeling and simulating humans, Teleoperation, tele-experience, tele-presence using humanoids
Abstract: Humans explore and adapt neuro-motor strategies to cope with limitations of multi-joint impedance regulation mechanism. For instance, predictive control of degrees of redundancy further regulates the endpoint impedance in addition to co-contractions. Inspired by these observations, this paper proposes a novel Tele-Impedance algorithm that replicates the human’s arm endpoint stiffness in robot by controlling the common-mode and configuration-dependant stiffness. Design of the controller and its stability is addressed and experimentally evaluated in robotic peg-in-hole task. Results of the proposed method are compared to the ones derived from Tele-Impedance implemented using classical Cartesian stiffness control. The interaction performance achieved highlights the possibility of adopting common mode stiffness in robots with adequate degrees of redundancy, in order to realize the desired task space impedance.

Keywords: Human and humanoid skills/cognition/interaction, Software and hardware architecture, system integration, Humanoid locomotion, manipulation, perception, planning
Abstract: This paper addresses an active audio-visual integration framework which integrates audio and visual information with a robot’s active motion for noise-robust Voice Activity Detection (VAD). VAD is crucial for noise robust automatic speech recognition (ASR) because speech captured by a robot’s microphones is usually contaminated with other noise sources. To realize such noise-robust VAD, we propose Active Audio- Visual (AAV) integration framework which integrates auditory, visual and motion information using a Causal Bayesian Network (CBN). CBN is a subclass of Bayesian networks, which is able to estimate the effect on VAD performance caused by active motions. Since CBN is a general framework for information integration, we can naturally introduce various types of information such as the location of a speaker and a noise source which affect VAD performance to CBN, and CBN selects the optimal active motion for better perception of the robot using “intervention” mechanism in CBN. We implemented a prototype system based on the proposed framework on a humanoid robot called Hearbo. The proposed AAV-VAD is compared with three types of AV-VAD; simple AAV-VAD, multiregression- based AAV-VAD, and stationary (not active) AVVAD. A preliminary experiment using the prototype system showed that the VAD performance of the proposed AV-VAD was 14.4, 26.0, and 30.3 points higher than that of the simple active, multi-regression-based active, and stationary AV-VAD, respectively.

Keywords: Humanoid locomotion, manipulation, perception, planning
Abstract: This paper is about grasping known objects of arbitrary shapes with a humanoid robot. We extend our previous work, where we presented a grasp planning method using an object representation based on the medial axis transform (MAT). The MAT describes an object's topological skeleton and contains information about local symmetry properties and thickness valuable for grasp planning. So far, our previous work was only conducted in simulation.

The contribution of this paper is the transfer of our grasp planning method to the real world. We present grasping experiments with challenging arbitratrily shaped objects where we execute the grasps generated by our grasp planner on a real humanoid robot with a five-finger hand.

Keywords: Humanoid locomotion, manipulation, perception, planning
Abstract: This paper proposes a novel approach for the simultaneous localization and reconstruction of a 3D object. This reconstruction is done with a small humanoid platform by the only means of monocular vision. We combine (1) a stochastic control approach to decide how to move the robot such that the object can be both localized with respect to the robot and correctly reconstructed, and (2) a space carving approach for the 3D reconstruction of the object that merges the acquired monocular views into a coherent model. We present a set of very promising results obtained with a Nao platform.

Keywords: Humanoid locomotion, manipulation, perception, planning
Abstract: This paper proposes solutions to the relative lack of efficiency of walking trajectories generated online. Extensions of a previously introduced Linear Model Predictive Control law are proposed that allow to improve the walking performance through variation of the height of the CoM and toe flexion. Extensive simulations serve to demonstrate the capacity of the improved scheme to generate significantly more efficient walking motions.

Keywords: Humanoid locomotion, manipulation, perception, planning, Adaptation, learning and cognitive development of humanoids, Software and hardware architecture, system integration
Abstract: This paper explores to which extent contact information is enough to perform grasps on unknown objects. The specific scenario addressed consists of emptying a box which contains an undefined number of unknown objects using only force-torque and tactile data. The proposed approach to solve the task, follows the manipulation primitives paradigm, which is based on the use of atomic controllers specifically designed to perform robustly elemental manipulation actions. A set of sensor based primitives that implement a reactive controller that adapts to the uncertain real environment is implemented. Manipulation primitives are assembled within a Finite State Machine, which sequentially searches, grasps and transports all the objects in the box to a predefined location.

Two different blind exploration approaches are used to deal with the task. Performance results showing grasp attempts, corrections performed, error corrected and elapsed time are presented and discussed. Moreover a simple vision system is added to compare its results to the blind approaches.

Keywords: Humanoid locomotion, manipulation, perception, planning, Applications: home, field, space, social, industrial, medical,health/mental care, art/entertainment, education
Abstract: This paper shows how an accurate peg-in-hole assembly task can be easily achieved with nothing but cheap position sensors when resourcing to Variable Impedance Actuators (VIA). We present the use of a low-cost Variable Stiffness Torso platform, that consists of two 4-DOF non-planar VSA manipulators, for a peg-in-hole assembly task using both arms. One arm holds the peg and the other holds the hole. The task is accomplished without any force measurement and without calling for parallel-manipulator control techniques, exploiting the intrinsic mechanical elasticity of the actuator units. Indeed, a simple position control scheme is required. Simulations and experimental results are reported.

Keywords: Humanoid locomotion, manipulation, perception, planning, Human and humanoid skills/cognition/interaction
Abstract: The Surface-Approximation Polynomials (SAP) descriptor has been shown to be an appropriate global surface descriptor for object categorization tasks in robotic applications. Nevertheless, in the original formulation the SAP descriptor is not invariant against rotations around the camera axis. This paper explains and evaluates two methods which pre-process the input data to yield repeatably well-aligned point clouds for the computation of the SAP descriptor. We show that the SAP descriptor can be rendered robust against rotations while retaining almost the full performance of the original approach which is superior to GFPFH, GRSD and VFH.

Keywords: Humanoid locomotion, manipulation, perception, planning, Human and humanoid skills/cognition/interaction
Abstract: In order to realize in-hand manipulation of unknown objects, we introduce an extension to our previously developed manipulation framework, such that long manipulation sequences, involving finger regrasping, become feasible. To thisend, we propose a novel feedback controller, which searchesfor locally optimal contact points (suitable for regrasping), employing an online exploration process on the unknown object surface. The method autonomously estimates and follows the gradient of a smooth objective function. More concretely, we propose to dynamically switch between manipulability and grasp stability depending on the grasp stability level. Physics-based simulation experiments, involving artificial noise to model real-world sensor readings, prove the feasibility of our approach by rotating an object while readjusting the grasp configuration with all fingers in turn.

Keywords: Humanoid locomotion, manipulation, perception, planning, Human and humanoid skills/cognition/interaction, Whole-body dynamics, control, sensing, informatics
Abstract: Three different approaches of having a full-size humanoid throw the first pitch at a Major League Baseball game are tested and implemented. The approaches include kinematic mapping using a motion capture system to capture a human's throwing motion then mapping that to a full-size humanoid. The second method is a fully automated approach that uses the sparse reachable map to provide viable full body throwing trajectories to provide the end effector with the desired velocity. The third approach borrows from the animation industry. The key-frames of the desired trajectory are constructed by hand. The time between each key-frame is defined by the user. Interpolation methods are used to smoothly move between key frames while limiting the jerk. Each method is analyzed and tested in simulation and on physical hardware. The full-size humanoid used is the Hubo series robot. Based on the latter tests one method was chosen to successfully throw the ceremonial first pitch at a Major League Baseball game in April 2012.

Keywords: Humanoid locomotion, manipulation, perception, planning, Software and hardware architecture, system integration, Applications: home, field, space, social, industrial, medical,health/mental care, art/entertainment, education
Abstract: Solving arbitrary manipulation tasks is a key feature for humanoid service robots. However, especially when tasks involve handling complex mechanisms or using tools, a generic action description is hard to define. Different objects require different handling methods. Therefore we try to solve manipulation tasks from point of view of the object, rather than in the context of the robot. Action templates within the object context are introduced to resolve object specific task constraints. As part of a centralized world representation, the action templates are integrated into the planning process. This results in an intuitive way of solving manipulation tasks. The underlying architecture as well as the mechanisms are discussed within this paper. The proposed methods are evaluated in two experiments.

Keywords: Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics, Human and humanoid skills/cognition/interaction
Abstract: The visual perception of a humanoid robot bridges the physical world with the internal world representation through visual skills such as self-localization, object recognition, detection, classification and tracking. Unfortunately, these skills are affected by internal and external sources of uncertainty. These uncertainties are present at various levels ranging from noisy signals and calibration deviations of the embodiment up to mathematical approximations and limited granularity of the perception-planning-action cycle. This aggregated uncertainty deteriorates and limits the precision and efficiency of the humanoid robot visual perception.

In order to overcome these limitations, the depth perception uncertainty should be modeled in the skills of the humanoid robots. Due to the complexity of the aggregated uncertainty in humanoid systems, the visual depth uncertainty can be hardly modeled analytically. However, the uncertainty distribution can be conveniently attained by supervised learning. The role of the supervisor is to provide ground-truth spatial measurements corresponding to the humanoid uncertain visual depth perception. In this article, a supervised learning method for inferring a novel model of the visual depth uncertainty is presented. The acquisition of the model is autonomously attained by the humanoid robot ARMAR-IIIB, see Fig.1.

Keywords: Humanoid locomotion, manipulation, perception, planning, Whole-body dynamics, control, sensing, informatics, Software and hardware architecture, system integration
Abstract: Previously, we have realized an efficient 3-D biped walking based on the Passive Dynamic Autonomous Control (PDAC), which is one of the point-contact methods. Since the dynamics based method such as the point-contact method utilizes the inherent dynamics of the robot, an efficient walking can be realized compared with the model based method. However, the stable range of the point-contact method is small and the robustness of the walking is not large enough for practical use. In order to increase robustness of the point-contact method, this paper proposes a swing leg retraction and a double support phase for the point-contact method. The swing leg retraction enhances the convergence range of the walking system; also, by use of the double support phase, the robot dynamics converges to a stable orbit even if a disturbance happens at a foot-contact. Finally, the proposed biped walking algorithm is verified by experiments.

Keywords: Humanoids for human science and engineering, Humanoid locomotion, manipulation, perception, planning
Abstract: This paper describes the dynamics identification of humanoid robots. It is important to know correctly dynamics parameters of link and joint which constitute a robot for its control. Here, we identify the inertial parameters of legged systems using the base-link dynamics. This method generally set the base-link at the torso where gyroscope and accelerometer are installed in order to calculate base-link velocity and acceleration. We propose a technique to identify without using these sensors. This technique sets the base-link at the sole of the foot and uses the kinematic constraint of the leg connected to the ground. Thus the base-link velocity and acceleration are set to zero. Therefore we can identify without inertial sensor's noise. In this paper, we apply this technique to the humanoid robot HRP-2. First, we calculate the model and then we identify HRP-2 dynamics parameters with simulation environment.

Keywords: Humanoids for human science and engineering, Social interaction and acceptability, Measuring, modeling and simulating humans
Abstract: This paper presents how human interprets robot intention by its verbal and nonverbal communication in poker game. A poker playing robot to socially interact with humans should exhibit intentionality in the sense that it makes the human believe that it has beliefs, desires and intentions. The experiment was designed to measure the interpreted intention and perceived emotion according to the role of nonverbal behaviors, which means substitution, complementation and contradiction for verbal messages. The result showed significant effects of the robot nonverbal behavior in mixed communication with verbal messages. The robot nonverbal behaviors can substitute their cor- responding verbal messages. Its nonverbal behaviors complement the verbal messages meaning that participants perceived these communications are congruence. In the case of contradiction, the participants recognized the contradictory nonverbal behaviors as incongruence.

Keywords: Measuring, modeling and simulating humans, Others (not listed above)
Abstract: We present a general real-time continuous collision detection algorithm for arbitrary systems of moving bodies connected to each other in a kinematic tree of joints. Here “joint” as a general term refers to the measured relative motion between two bodies, which may be physically connected or not. We provide a basic set of joints covering revolute and prismatic joints, vehicle motion, and 3d positioning which is sufficient for many applications in particular those involved with mobile manipulators, e.g. industrial and humanoid robots, intelligent transportation systems, or equipment at construction sites. Each joint implementation either operates on a motion bound (an interval covering the braking distance) or an uncertainty bound (measurement error) and computes a volume that spatially bounds the effect of that motion or uncertainty. Aggregating all joints in a kinematic tree then yields a conservative continuous collision detection for the complex and uncertain motion of the whole system.

We further present a augmented reality visualization that overlays the collision volumes into the live image of a camera, which can be used to validate the collision model before bringing a system into service or support human while operating equipment, e.g., on construction sites.

Keywords: Measuring, modeling and simulating humans, Superhuman humanoids / future humanoids: possibility, technology and meaning, Applications: home, field, space, social, industrial, medical,health/mental care, art/entertainment, education
Abstract: In this paper, we propose a method to apply any new face to a retro-projected 3D face system, the Mask-bot, which we have developed as a human-robot interface. The robot face using facial animation projected onto a 3D face mask can be quickly replaced by a new face based on a single frontal image of any person. Our contribution is to apply an automatic face replacement technique with the modified texture morphable model fitting to the 3D face mask. Using our technique, a face model displayed on Mask-bot can be automatically replaced within approximately 3 seconds, which makes Mask-bot widely suitable to applications such as video conferencing and cognitive experiments.

Keywords: Novel materials, devices, mechanisms, energy system for humanoids
Abstract: High-powered humanoid robots are expected that it can move at a high speed in the same environment as human. In that case, when they receive shocks suddenly, instantaneous impact is applied to the structure of the feet and they can be broken. So, robots require shock-reduction mechanisms.In this paper, we consider a high-powered elastic humanoid robot. and we propose a mechanism to make a hybrid leg, which uses a tendon driven joint as an ankle and axial-driven joints as other joints. We designed and developed this shock-reduction mechanisms and verify the validity of this mechanisms.

Keywords: Novel materials, devices, mechanisms, energy system for humanoids, Humanoid locomotion, manipulation, perception, planning
Abstract: This article describes the design and testing process of a new pair of legs for the iCub. The iCub is an opensource humanoid robot that is being used by more than 20 institutes and research centers worldwide. Although the legs of the original iCub allow the robot to perform basic balancing and crawling locomotion the extension of its locomotion skills to bipedal walking could give to the iCub community additional possibilities for a wider range of experimentation. To extend the iCub locomotion capabilities to bipedal walking we thus undertook a design revision of the legs. A major novelty of the new design is the introduction of series elastic actuators (SEA) at the knee and ankle joints. Additional modifications were applied at the ankle joint to improve both the torque capability and range of motion. The present article will describe the conceptual design of the new system, and will present some preliminary data regarding the testing of the elastic module.

Keywords: Novel materials, devices, mechanisms, energy system for humanoids, Humanoid locomotion, manipulation, perception, planning, Principles and technologies for anthropomorphic/bionic design and control
Abstract: Exploiting variable impedance for dynamic tasks such as walking is both challenging and topical -- research progress in this area impacts not only autonomous, bipedal mobility but also prosthetics and exoskeletons. In this work, we present the design, construction and preliminary testing of a planar bipedal robot with joints capable of physically varying both their stiffness and damping independently -- the first of its kind. A wide variety of candidate variable stiffness and damping actuator designs are investigated. Informed by human biophysics and locomotion studies, we design an appropriate (heterogenous) impedance modulation mechanism that fits the necessary torque and stiffness range and rate requirements at each joint while ensuring the right form factor. In addition to hip, knee and ankle, the constructed robot is also equipped with a three part compliant foot modelled on human morphology. We describe in detail the hardware construction and the communication and control interfaces. We also present a full physics based dynamic simulation which matches the hardware closely. Finally, we test impedance modulation response characteristics and a basic walking gait realised through a simple movement controller, both in simulation and on the real hardware.

Keywords: Principles and technologies for anthropomorphic/bionic design and control, Humanoid locomotion, manipulation, perception, planning, Others (not listed above)
Abstract: This paper investigates the stability principle of a planar, underactuated, four-linked limit cycle walker which is strictly controlled to follow the desired-time trajectories. We first linearize the equation of motion of the robot and analytically derive the transition equations of the state error for the stance and heel-strike phases. Second, we investigate the limit-cycle stability, specially the relation between the heel-strike posture and the gait convergence, through numerical simulations. Throughout the gait analysis, it is shown that the gait convergence changes from the speed mode to the totter mode through the deadbeat mode with respect to the asymmetry of the heel-strike posture.

Keywords: Principles and technologies for anthropomorphic/bionic design and control, Novel materials, devices, mechanisms, energy system for humanoids, Humanoid locomotion, manipulation, perception, planning
Abstract: One of the most important features for a system capable of working in uncertain and unstructured environments is reliability. Nowadays robots are excellent machines, but are still not able to interact with their surrounding environment as humans or animals do. Recent studies highlight the role played by impedance changes in the human arm during manipulation tasks. In particular the possibility to vary the stiffness of shoulder, elbow and wrist allows humans to interact easily with fast changing environments while rejecting unpredictable noise disturbances. Several studies also showed how the capability of “cocontracting” antagonistic muscles is required to interact with noisy/unpredictable environments. Starting from these premises we recently proposed novel design principles to build actuators with the ability to actively regulate the passive noise rejection (i.e. the ability to cancel the effect of disturbances without explicitly relying on feedback). In the present paper we implement these principles in the mechanical design of a novel actuator. The actuator is composed of two electric motors in agonistantagonist configuration. The final design includes also four nonlinear springs whose force-displacement characteristic has been customized on the specific application requirements. Validation of the proposed non-linear spring design has been conducted on a prototype and results are reported in this paper. Future works foreseen the integration of the proposed actuators on a two limbed robot with six artificial muscle, (three agonist-antagonist pairs) in a simple and bi-articular configuration.

Keywords: Social interaction and acceptability, Human and humanoid skills/cognition/interaction, Others (not listed above)
Abstract: In face-to-face interaction, humans coordinate actions in their surroundings with the help of a well structured spatial representation. For example on a dinner table, everybody exactly knows which objects belong to her and where she is allowed to grasp. To have robots, e.g. receptionists, act accordingly, we conducted an on-line survey about the expectations humans have while interacting with such a robot. Results indicate that humans attribute the robot handedness and an awareness of distance and territoriality in its own peripersonal space. In order to align a robot's behavior to these expectations, we have have developed a first spatial representation of the robots peripersonal space.

Keywords: Software and hardware architecture, system integration, Human and humanoid skills/cognition/interaction, Humanoid locomotion, manipulation, perception, planning
Abstract: Even for robots delivered with a middleware, end-user implementation of reaction and deliberation remains a difficult task. We proposed Targets Drives Means (TDM), a behavior based architecture which level of abstraction allows behavior specification through declarative association of reusable components. If behavior based robotic has been broadly used in specialized architectures, TDM implemented it for end-users tools for behavior specification. In this paper TDM is evaluated as a solution for end-user robot programming and results of comparative usability tests are presented.

Keywords: Software and hardware architecture, system integration, Social interaction and acceptability, Novel materials, devices, mechanisms, energy system for humanoids
Abstract: This paper describes the development of “Mask- Bot 2i” (codenamed: Kabuto), a robotic head designed for research into Human-Robot-Interactions. The uniqueness of our new robotic head is that the appearance of its face can be alter on-the-fly. Different faces can be projected onto the active head system. The head is fully active with a 3-DOF neck, with supports for biannual hearing as well as video camera for seeing. An interchangeable face is the main feature of this new Mask-Bot, the head can be equipped with average face mask as well as for better visibility a customised individualised face can be easily exchanged. Additionally, the action of the head has been designed to match the natural head movements of an average person. Thus, enable the head and face to be activated synchronously to speech production while natural head movements matching that of the animated face. The design and realisation of this new system is presented in details in this paper.

Keywords: Software and hardware architecture, system integration, Whole-body dynamics, control, sensing, informatics, Human and humanoid skills/cognition/interaction
Abstract: Recognizing environmental contact on whole body of a humanoid robot can be very advantageous to work with people in human's environment. In the tasks with environmental contacts, it is important as an interface with the environment to detect pushing, shearing and twist on the whole body of a robot such that it gets to know its current state and what to do next. In this paper, we describe a full body soft tactile sensor suit for a humanoid robot and an algorithm to calculate pushing, shearing, and twist for each sensor unit. These sensors are small muti-axis sensors with urethane structure and they can be placed densely on the body of a humanoid robot. We arranged 347 multi-axis soft tactile sensors on a humanoid robot imitating a human tactile sense to detect contact states. Then, we calculate a deformation vector for each muti-axis soft tactile sensor and detect the three contact states using deformation moment and average of deformation vectors in the contact surface consisting of soft tactile sensors. Finally, we confirmed the validity of the full body tactile suit and contact state detector by object passing experiment between a human and a robot.

Keywords: Teleoperation, tele-experience, tele-presence using humanoids, Cyborgs, prostheses, assistive devices and sensor/motor suits, Applications: home, field, space, social, industrial, medical,health/mental care, art/entertainment, education
Abstract: Brain-controlled robots as a "surrogate presence" for humans may appear only a vision to date. However, the field of brain-robot interfacing is making rapid progress. Humanoid robots are ideal candidates for providing such a "surrogate presence", because they have the same embodiment as humans. Telepresence scenarios, such as virtual meetings, imply that not only one robot controlled by one human user is present, but that several users interact with each other mediated by their robots. Inspired by this scenario, we present a multi-user brain-robot interface, where currently two users control one humanoid robot each and interact with each other in a shared space. Brain-control is based on an asynchronous, dynamic and hybrid EEG-based brain-robot interface. We investigated two types of interaction: collaboration and competition. System performance was evaluated in a user study with 12 participants. Our results show that all users are capable of controlling the robots in the complex tasks. The complexity, however, imposes a high cognitive load that hampers focusing on the interaction with the other user.

Keywords: Whole-body dynamics, control, sensing, informatics, Humanoid locomotion, manipulation, perception, planning
Abstract: n this contribution we propose an inverse dynamics controller for a humanoid robot that exploits torque redundancy to minimize any combination of linear and quadratic costs in the contact forces and the commands. In addition the controller satisfies linear equality and inequality constraints in the contact forces and the commands such as torque limits, unilateral contacts or friction cones limits. The originality of our approach resides in the formulation of the problem as a quadratic program where we only need to solve for the control commands and where the contact forces are optimized implicitly. Furthermore, we do not need a structured representation of the dynamics of the robot (i.e. an explicit computation of the inertia matrix). It is in contrast with existing methods based on quadratic programs. The controller is then robust to uncertainty in the estimation of the dynamics model and the optimization is fast enough to be implemented in high bandwidth torque control loops that are increasingly available on humanoid platforms. We demonstrate properties of our controller with simulations of a human size humanoid robot.

Keywords: Whole-body dynamics, control, sensing, informatics, Software and hardware architecture, system integration
Abstract: We present the Dynamic Force Field Controller (DForC), a reliable and effective framework in the context of humanoid robotics for real-time reaching and tracking in presence of obstacles. It is inspired by well established works based on artificial potential fields, providing a robust basis for sidestepping a number of issues related to inverse kinematics of complex manipulators. DForC is composed of two layers organized in descending order of abstraction: (1) at the highest level potential fields are employed to outline on the fly collision-free trajectories that serve to drive the robot end-effector toward fixed or moving targets while accounting for obstacles; (2) at the bottom level an optimization algorithm is exploited in place of traditional techniques that resort to the Transposed or Pseudo-Inverse Jacobian, in order to deal with constraints specified in the joints space and additional conditions related to the robot structure. As demonstrated by experiments conducted on the iCub robot, our method reveals to be particularly flexible with respect to environmental changes allowing for a safe tracking procedure, and generating reliable paths in practically every situation.