Learning control policies from high-dimensional robot demonstrations remains challenging due to the need for large datasets, limited generalization, and high model complexity. These limitations hinder data-efficient learning and reactive deployment from limited demonstrations. This paper introduces LEMON-DS, a compact and data-efficient framework for one shot learning of stable control policies from high-dimensional demonstrations. LEMON-DS proceeds in three stages: (i) a graph Laplacian–based spectral embedding that reveals a quasi linear latent structure from a single trajectory, (ii) defining a globally asymptotically stable dynamical system (DS) in the latent space, and (iii) learning a diffeomorphic mapping that reconstructs the demonstrated behaviour. Collectively, this pipeline yields a lightweight reactive control policy in the form of a dynamical system that is parameter-efficient, globally asymptotically stable, and faithfully reproduces the demonstrated behaviour generalizing across initial conditions. We show empirical performance across diverse robot tasks up to 23 dimensions, outperforming prior approaches in accuracy and model compactness.

Humans regularly throw projectiles with high speed and accuracy; some animals, including chimpanzees and elephants, also throw objects occasionally. In comparison, robots are currently lagging behind, despite having lower communication latency and more accurate motor control. To understand this paradox and ultimately achieve ubiquitous throwing robots, one of the major obstacles is the lack of high-fidelity and tractable physical models of the transient release dynamics, where the momentum exchange between the hand and the object occurs within tens of milliseconds via the frictional interface. In this work, we try to establish a physical model for the release dynamics. We first demonstrate that the conventional model, which combines rigid-body dynamics and patch friction (limit surface, LS), struggles to capture the release dynamics and exhibits pathological behaviors, such as Zeno-like oscillations, leading to poor accuracy in predicting throwing outcomes. To mitigate this, we formulate a viscous-smoothed variant of the limit surface model solved via implicit integration (ILS), which achieves high predictive fidelity but incurs significant computational cost. On the other hand, motivated by the dominant effect of in-hand pivoting in release dynamics, we propose a Sliding Pivot (SP) model that simplifies the contact dynamics by capturing the sticking-pivoting-sliding behavior emerging under vanishing normal force. This model achieves accuracy comparable to ILS, with only 10% higher error while offering over 20× faster computation. Compared to conventional LS models, our method reduces horizontal velocity prediction error by 40% and angular velocity prediction error by 63%, achieving 2.4 cm mean absolute error (MAE) for landing position and 15.4 degrees MAE for landing orientation. These results provide a robust, physically grounded foundation for future scalable robot throwing systems.

In this paper, we introduce a novel approach to implicitly encode precise robot morphology using forward kinematics based on a configuration space signed distance function. Our proposed Robot Neural Distance Function (RNDF) optimizes the balance between computational efficiency and accuracy for signed distance queries conditioned on the robot's configuration for each link. Compared to the baseline method, the proposed approach achieves an 81.1% reduction in distance error while utilizing only 47.6% of model parameters. Its parallelizable and differentiable nature provides direct access to joint-space derivatives, enabling a seamless connection between robot planning in Cartesian task space and configuration space. These features make RNDF an ideal surrogate model for general robot optimization and learning in 3D spatial planning tasks. Specifically, we apply RNDF to robotic arm-hand modeling and demonstrate its potential as a core platform for wholearm, collision-free grasp planning in cluttered environments. The code and model are available at https://github.com/roboticmanipulation/RNDF.

Robotic manipulation in dynamic environments often requires seamless transitions between different grasp types to maintain stability and efficiency. However, achieving smooth and adaptive grasp transitions remains a challenge, particularly when dealing with external forces and complex motion constraints. Existing grasp transition strategies often fail to account for varying external forces and do not optimize motion performance effectively. In this work, we propose an Imitation-Guided Bimanual Planning Framework that integrates efficient grasp transition strategies and motion performance optimization to enhance stability and dexterity in robotic manipulation. Our approach introduces Strategies for Sampling Stable Intersections in Grasp Manifolds for seamless transitions between uni-manual and bi-manual grasps, reducing computational costs and regrasping inefficiencies. Additionally, a Hierarchical Dual-Stage Motion Architecture combines an Imitation Learning-based Global Path Generator with a Quadratic Programming-driven Local Planner to ensure real-time motion feasibility, obstacle avoidance, and superior manipulability. The proposed method is evaluated through a series of force-intensive tasks, demonstrating significant improvements in grasp transition efficiency and motion performance. A video demonstrating our simulation results can be viewed at https://youtu.be/3DhbUsv4eDo.

This paper builds up the skill of impact aware non prehensile manipulation through a hitting motion of a redundant robot arm by allowing it to come in contact with the environment with the appropriate link according to the requirements of the hitting task. In tasks where directional effective inertia of a robot is important at the contact point, it is useful to understand inertia at different links, so as to select the appropriate link. Hitting with those links allows us to manipulate a wider range of object masses since the robot effective inertia is different at different links. We propose a learning based methodology for selecting a hitting link based on the hitting task specifications, impact posture generation for the robot and an automated generation of desired directional inertia values throughout the hitting motion.

Dynamic manipulation, such as robot tossing or throwing objects, has recently gained attention as a novel paradigm to speed up logistic operations. However, the focus has predominantly been on the object's landing location, irrespective of its final orientation. In this work, we present a method enabling a robot to accurately "throw-flip" objects to a desired landing pose (position and orientation). Conventionally, objects thrown by revolute robots suffer from parasitic rotation, resulting in highly restricted and uncontrollable landing poses. Our approach is based on two key design choices: first, leveraging the impulse-momentum principle, we design a family of throwing motions that effectively decouple the parasitic rotation, significantly expanding the feasible set of landing poses. Second, we combine a physics-based model of free flight with regression-based learning methods to account for unmodeled effects. Real robot experiments demonstrate that our framework can learn to throw-flip objects to a pose target within (±5 cm, ±45 degree) threshold in dozens of trials. Thanks to data assimilation, incorporating projectile dynamics reduces sample complexity by an average of 40% when throw-flipping to unseen poses compared to end-to-end learning methods. Additionally, we show that past knowledge on in-hand object spinning can be effectively reused, accelerating learning by 70% when throwing a new object with a Center of Mass (CoM) shift. A video summarizing the proposed method and the hardware experiments is available at https://youtu.be/txYc9b1oflU.

Performing shotcrete operations at construction sites can be hazardous to humans and inefficient. Robots can offer a safer and more efficient alternative to assist in these tasks. We present a new planning strategy for shotcrete robots, including both the spraying and surface finishing phases, that can plan for a general target area, whether flat or complexly curved. Our method uses learning from demonstrations and dynamical systems concepts to enable reactive and adaptive planning for robots, allowing them to effectively handle disturbances. We evaluated the effectiveness of the proposed planning and control framework in a laboratory setup using a velocity-controlled robot and curved targets both in the spraying and polishing phases. The results demonstrate the effectiveness of the proposed approach.

Established techniques that enable robots to learn from demonstrations are based on learning a stable dynamical system (DS). To increase the robots’ resilience to perturbations during tasks that involve static obstacle avoidance, we propose incorporating barrier certificates into an optimization problem to learn a stable and barrier-certified DS. Such optimization problem can be very complex or extremely conservative when the traditional linear parameter-varying formulation is used. Thus, different from previous approaches in the literature, we propose to use polynomial representations for DSs, which yields an optimization problem that can be tackled by sum-of-squares techniques. Finally, our approach can handle obstacle shapes that fall outside the scope of assumptions typically found in the literature concerning obstacle avoidance within the DS learning framework. Supplementary material can be found at the project webpage: https://martinschonger.github.io/abc-ds

When humans interact with each other, eye gaze movements have to support motor control as well as communication. On the one hand, we need to fixate the task goal to retrieve visual information required for safe and precise action-execution. On the other hand, gaze movements fulfil the purpose of communication, both for reading the intention of our interaction partners, as well as to signal our action intentions to others. We study this Gaze Dialogue between two participants working on a collaborative task involving two types of actions: 1) individual action and 2) action-in-interaction. We recorded the eye-gaze data of both participants during the interaction sessions in order to build a computational model, the Gaze Dialogue, encoding the interplay of the eye movements during the dyadic interaction. The model also captures the correlation between the different gaze fixation points and the nature of the action. This knowledge is used to infer the type of action performed by an individual. We validated the model against the recorded eye-gaze behavior of one subject, taking the eye-gaze behavior of the other subject as the input. Finally, we used the model to design a humanoid robot controller that provides interpersonal gaze coordination in human-robot interaction scenarios. During the interaction, the robot is able to: 1) adequately infer the human action from gaze cues; 2) adjust its gaze fixation according to the human eye-gaze behavior; and 3) signal nonverbal cues that correlate with the robot's own action intentions.

Controlling complex tasks in robotic systems, such as circular motion for cleaning or following curvy lines, can be dealt with using nonlinear vector fields. This paper introduces a novel approach called the rotational obstacle avoidance method (ROAM) for adapting the initial dynamics when obstacles partially occlude the workspace. ROAM presents a closed-form solution that effectively avoids star-shaped obstacles in spaces of arbitrary dimensions by rotating the initial dynamics toward the tangent space. The algorithm enables navigation within obstacle hulls and can be customized to actively move away from surfaces while guaranteeing the presence of only a single saddle point on the boundary of each obstacle. We introduce a sequence of mappings to extend the approach for general nonlinear dynamics. Moreover, ROAM extends its capabilities to handle multi-obstacle environments and provides the ability to constrain dynamics within a safe tube. By utilizing weighted vector-tree summation, we successfully navigate around general concave obstacles represented as a tree-of-stars. Through experimental evaluation, ROAM demonstrates superior performance in minimizing occurrences of local minima and maintaining similarity to the initial dynamics, outperforming existing approaches in multi-obstacle simulations. Due to its simplicity, the proposed method is highly reactive and can be applied effectively in dynamic environments. This was demonstrated during the collision-free navigation of a 7-degree-of-freedom robot arm around dynamic obstacles.

In this article, we propose a metric called hitting flux, which is used in the motion generation and controls for a robot manipulator to interact with the environment through a hitting or a striking motion. Given the task of placing a known object outside of the workspace of the robot, the robot needs to come in contact with it at a nonzero relative speed. The configuration of the robot and the speed at contact matter because they affect the motion of the object. The physical quantity called hitting flux depends on the robot's configuration, the robot speed, and the properties of the environment. An approach to achieve the desired directional preimpact flux for the robot through a combination of a dynamical system for motion generation and a control system that regulates the directional inertia of the robot is presented. Furthermore, a quadratic program formulation for achieving a desired inertia matrix at a desired position while following a motion plan constrained to the robot limits is presented. The system is tested for different scenarios in simulation showing the repeatability of the procedure and in real scenarios with KUKA LBR iiwa 7 robot.

In robotic throwing, the release phase involves complex dynamic interactions due to object deformation and limited gripper opening speed, often resulting in inaccurate and nonrepeatable throws. While data-driven methods can be employed to compensate for the release uncertainty, the generalizability of learned models to unseen objects is not guaranteed, and object-specific fine-tuning with new data may be required. This fine-tuning process raises concerns about the scalability of such methods for dexterous throwing, where the robot needs to execute diverse motions for throwing various objects. Instead of case-by-case fine-tuning, we aim at designing throwing motion robust against release uncertainty. We encapsulate all uncertainties resulting from complex contact dynamics in a surrogate model of their resulting effect ongripper opening delay. We introduce the notion oftube accelerationto model the class of constant-acceleration motion in joint space that guarantees a release within the set of valid throwing configurations. We propose a convex relaxation of the primal optimization problem with a tight error bound and evaluate its performance in terms of reliability and efficiency. Results show that the approach offers run-time performance to allow online computation of throws on a 7-DoF robot arm. It achieves a high accuracy and success rate (97% for planar throws) at throwing a variety of complex objects, even when using a simple ballistic model for the object's flying dynamics.

One can estimate the velocity and acceleration of robot manipulators by utilizing nonlinear observers. This involves combining inertial measurement units (IMUs) with the motor encoders of the robot through a model-based sensor fusion technique. This approach is lightweight, versatile (suitable for a wide range of trajectories and applications), and straightforward to implement. In order to further improve the estimation accuracy while running the system, we propose to adapt the noise information in this paper. This would automatically reduce the system vulnerability to imperfect modelings and sensor changes. Moreover, viable strategies to maintain the system stability are introduced. Finally, we thoroughly evaluate the overall framework with a seven DoF robot manipulator whose links are equipped with IMUs.

Efficient and safe multi-agent swarm coordination in environments where humans operate, such as warehouses, assistive living rooms, or automated hospitals, is crucial for adopting automation. In this paper, we augment the obstacle avoidance algorithm based on dynamical system modulation for a swarm of heterogeneous holonomic mobile agents. A smooth prioritization is proposed to change the reactivity of the swarm towards the specific agents. Further, a soft decoupling of the initial agent's kinematics is used to design an independent rotation control to ensure the agent reaches the desired position and orientation simultaneously. This decoupling allowed the introduction of a novel heuristic, the virtual drag. It minimizes the disturbance influence an agent has when moving through its surrounding. Additionally, the safety module adapts the velocity commands from the dynamical system modulation to avoid colliding trajectories between agents. The evaluation was performed in simulated assisted living and hospital environments. The prioritization successfully increased the minimum distance relative to a moving agent. The safety module is observed to create collision-free dynamics where alternative methods fail. Additionally, the repulsive nature of the safety module augments the convergence rate, thus making the proposed method better applicable to dense real-world scenarios.

Gaussian process regression is widely used because of its ability to provide well-calibrated uncertainty estimates and handle small or sparse datasets. However, it struggles with high-dimensional data. One possible way to scale this technique to higher dimensions is to leverage the implicit low-dimensional manifold upon which the data actually lies, as postulated by the manifold hypothesis. Prior work ordinarily requires the manifold structure to be explicitly provided though, i.e. given by a mesh or be known to be one of the well-known manifolds like the sphere. In contrast, in this paper we propose a Gaussian process regression technique capable of inferring implicit structure directly from data (labeled and unlabeled) in a fully differentiable way. For the resulting model, we discuss its convergence to the Matérn Gaussian process on the assumed manifold. Our technique scales up to hundreds of thousands of data points, and improves the predictive performance and calibration of the standard Gaussian process regression in some high-dimensional settings.

We present a control framework for achieving a robust object grasp and manipulation in hand. In-hand manipulation remains a demanding task as the object is never stable and task success relies on carefully synchronizing the fingers' dynamics. Indeed, fingers must simultaneously generate motion while maintaining contact with the object and, by staying within the hand's frame, ensuring that the object remains manipulable. These challenges are exacerbated once the hand gets disturbed or when the internal dynamics of the manipulated object are unknown, such as when it is filled with liquid moving during manipulation. We present a control strategy based on coupled dynamical systems (DSs), whereby the fingers move in synchronization using an intermediate dynamics responsible for coordinating fingers. To adapt to changes in forces due to model uncertainties and unexpected disturbances, we employ an adaptive torque-controller combined with a joint impedance regulator that guarantees high tracking accuracy while adapting to dynamic changes. We validate the approach in multiple experiments on 16-degrees-of-freedom robotic hand grasping and manipulating objects with different mass properties, e.g., uneven or varying mass distribution in a glass half-filled with water. We show that the robot can compensate for disturbances generated by internal dynamics and external perturbations. Additionally, we showcase how our controller, in conjunction with learning from human demonstration, provides a robust solution for more complicated manipulations such as finger gaiting.

Humans coordinate the abundant degrees of freedom (DoFs) of hands to dexterously perform tasks in everyday life. We imitate human strategies to advance the dexterity of multi-DoF robotic hands. Specifically, we enable a robot hand to grasp multiple objects by exploiting its kinematic redundancy, referring to all its controllable DoFs. We propose a human-like grasp synthesis algorithm to generate grasps using pairwise contacts on arbitrary opposing hand surface regions, no longer limited to fingertips or hand inner surface. To model the available space of the hand for grasp, we construct a reachability map, consisting of reachable spaces of all finger phalanges and the palm. It guides the formulation of a constrained optimization problem, solving for feasible and stable grasps. We formulate an iterative process to empower robotic hands to grasp multiple objects in sequence. Moreover, we propose a kinematic efficiency metric and an associated strategy to facilitate exploiting kinematic redundancy. We validated our approaches by generating grasps of single and multiple objects using various hand surface regions. Such grasps can be successfully replicated on a real robotic hand.

Quadratic programming (QP)-based controllers allow many robotic systems, such as humanoids, to successfully undertake complex motions and interactions. However, these approaches rely heavily on adequately capturing the underlying model of the environment and the robot’s dynamics. This assumption, nevertheless, is rarely satisfied, and we usually turn to well-tuned end-effector PD controllers to compensate for model mismatches. In this article, we propose to augment traditional QP-based controllers with a learned residual inverse dynamics (IDs) model and an adaptive control law that adjusts the QP online to account for model uncertainties and unforeseen disturbances. In particular, we propose: 1) learning a residual IDs model using the Gaussian Process and linearizing it so that it can be incorporated inside the QP-control optimization procedure and 2) a novel combination of adaptive control and QP-based methods to avoid the manual tuning of end-effector PID controllers and faster convergence in learning the residual dynamics model. In simulation, we extensively evaluate our method in several robotic scenarios ranging from a 7-degrees of freedom (DoFs) manipulator tracking a trajectory to a humanoid robot performing a waving motion for which the model used by the controller and the one used in the simulated world do not match (unmodeled dynamics). Finally, we also validate our approach in physical robotic scenarios where a 7-DoFs robotic arm performs tasks where the model of the environment (mass, friction coefficients, etc.) is not fully known.

Large efforts have focused on ensuring that the controllers for mobile service robots follow proxemics and other social rules to ensure both safe and socially acceptable distance to pedestrians. Nonetheless, involuntary contact may be unavoidable when the robot travels in crowded areas or when encountering adversarial pedestrians. Freezing the robot in response to contact might be detrimental to bystanders' safety and prevents it from achieving its task. Unavoidable contacts must hence be controlled to ensure the safe and smooth travelling of robots in pedestrian alleys. We present a force-limited and obstacle avoidance controller integrated into a time-invariant dynamical system (DS) in a closed-loop force controller that let the robot react instantaneously to contact or to the sudden appearance of pedestrians. Mitigating the risk of collision is done by modulating the velocity commands upon detecting a contact and by absorbing part of the contact force through active compliant control when the robot bumps inad-vertently against a pedestrian. We evaluated our method with a personal mobility robot -Qolo- showing contact mitigation with passive and active compliance. We showed the robot able to overcome an adversarial pedestrian within 9 N of the set limit contact force for speeds under 1 m/s. Moreover, we evaluated integrated obstacle avoidance proving the ability to advance without Incurring any other collision.

In order to facilitate and assist the indoor mobility of people with special needs, the classically static objects in the environment, such as furniture, can be rendered mobile. The need for efficient and safe autonomous coordination of a mobile furniture swarm arises. We present a closed-form approach for mobile furniture obstacle avoidance and navigation within an indoor environment. The approach shows that each mobile furniture agent, defined by a polygonal surface, does not collide with any static or mobile obstacle (e.g., a person is moving around). All controllable mobile furniture converges towards a defined goal position and orientation. We showcase the application of this algorithm in simulation on mobile furniture for smart environments. Results demonstrate that the proposed method can coordinate a swarm of mobile furniture to get out of the way of a mobile agent representing a person with limited mobility passing through the room while avoiding obstacles and converging towards a predefined target pose.

Mobile manipulator throwing is a promising method to increase the flexibility and efficiency of dynamic manipulation in factories. Its major challenge is to efficiently plan a feasible throw under a wide set of task specifications. We show that the mobile manipulator throwing problem can be simplified to a planar problem, hence greatly reducing the computational costs. Using machine learning approaches, we build a model of the object's inverted flying dynamics and the robot's kinematic feasibility, which enables throwing motion generation within 1 ms for given query of target position. Thanks to the computational efficiency of our method, we show that the system is adaptive under disturbance, via replanning on the fly for alternative solutions, instead of sticking to the original throwing plan.

Autonomous navigation in highly populated areas remains a challenging task for robots because of the difficulty in guaranteeing safe interactions with pedestrians in unstructured situations. In this work, we present a crowd navigation control framework that delivers continuous obstacle avoidance and post-contact control evaluated on an autonomous personal mobility vehicle. We propose evaluation metrics for accounting efficiency, controller response and crowd interactions in natural crowds. We report the results of over 110 trials in different crowd types: sparse, flows, and mixed traffic, with low- (< 0.15 ppsm), mid- (< 0.65 ppsm), and high- (< 1 ppsm) pedestrian densities. We present comparative results between two low-level obstacle avoidance methods and a baseline of shared control. Results show a 10% drop in relative time to goal on the highest density tests, and no other efficiency metric decrease. Moreover, autonomous navigation showed to be comparable to shared-control navigation with a lower relative jerk and significantly higher fluency in commands indicating high compatibility with the crowd. We conclude that the reactive controller fulfils a necessary task of fast and continuous adaptation to crowd navigation, and it should be coupled with high-level planners for environmental and situational awareness.

Dynamical Systems (DS) are fundamental to the modeling and understanding time evolving phenomena, and have application in physics, biology and control. As determining an analytical description of the dynamics is often difficult, data-driven approaches are preferred for identifying and controlling nonlinear DS with multiple equilibrium points. Identification of such DS has been treated largely as a supervised learning problem. Instead, we focus on an unsupervised learning scenario where we know neither the number nor the type of dynamics. We propose a Graph-based spectral clustering method that takes advantage of a velocity-augmented kernel to connect data points belonging to the same dynamics, while preserving the natural temporal evolution. We study the eigenvectors and eigenvalues of the Graph Laplacian and show that they form a set of orthogonal embedding spaces, one for each sub-dynamics. We prove that there always exist a set of 2-dimensional embedding spaces in which the sub-dynamics are linear and n-dimensional embedding spaces where they are quasi-linear. We compare the clustering performance of our algorithm to Kernel K-Means, Spectral Clustering and Gaussian Mixtures and show that, even when these algorithms are provided with the correct number of sub-dynamics, they fail to cluster them correctly. We learn a diffeomorphism from the Laplacian embedding space to the original space and show that the Laplacian embedding leads to good reconstruction accuracy and a faster training time through an exponential decaying loss compared to the state-of-the-art diffeomorphism-based approaches.

This article presents a closed-form approach to constraining a flow within a given volume and around objects. The flow is guaranteed to converge and to stop at a single fixed point. The obstacle avoidance problem is inverted to enforce that the flow remains enclosed within a volume defined by a polygonal surface. We formally guarantee that such a flow will never contact the boundaries of the enclosing volume or obstacles. It asymptotically converges toward an attractor. We further create smooth motion fields around obstacles with edges (e.g., tables). Both obstacles and enclosures may be time-varying, i.e., moving, expanding, and shrinking. The technique enables a robot to navigate within enclosed corridors while avoiding static and moving obstacles. It was applied on an autonomous robot (QOLO) in a static complex indoor environment and tested in simulations with dense crowds. The final proof of concept was performed in an outdoor environment in Lausanne. The QOLO-robot successfully traversed a marketplace in the center of town in the presence of a diverse crowd with a nonuniform motion pattern.

The evaluation of robot capabilities to navigate human crowds is essential to conceive new robots intended to operate in public spaces. This paper initiates the development of a benchmark tool to evaluate such capabilities; our long term vision is to provide the community with a simulation tool that generates virtual crowded environment to test robots, to establish standard scenarios and metrics to evaluate navigation techniques in terms of safety and efficiency, and thus, to install new methods to benchmarking robots’ crowd navigation capabilities. This paper presents the architecture of the simulation tools, introduces first scenarios and evaluation metrics, as well as early results to demonstrate that our solution is relevant to be used as a benchmark tool.

The inverse dynamics of a robotic manipulator is instrumental in precise robot control and manipulation. However, acquiring such a model is challenging, not only due to unmodelled non-linearities such as joint friction, but also from a machine learning perspective (e.g., input space dimension, amount of data needed). The accuracy of such models, regardless of the learning techniques, relies on proper excitation and exploration of the robot’s configuration space, in order to collect a rich dataset. This study aims to provide rich data in learning the inverse dynamics of a serial robotic manipulator using supervised machine learning techniques. We propose a method, called Max-Information Configuration Exploration (MICE), to incrementally explore and generate information-rich data via computing parameters of a trajectory set. We also introduce a new set of excitation trajectories that explores robot’s configuration through imposed stable limit cycles in robot joints’ phase space while satisfying feasibility constraints and physical bounds. We benchmark MICE against state-of-the-art in terms of data quality and learning accuracy. The proposed methodology for data collection, model learning, and evaluation, is validated with a KUKA IIWA14 robotic arm where the results prove significant improvement over traditional approaches.

Foot devices have been ubiquitously used in surgery to control surgical equipment. Most common applications are foot switches for electro-surgery, endoscope positioning and tele-robotic consoles. Switches fall short of providing continuous control as required for precise use of instruments. We developed a haptic foot interface to provide continuous assistance in surgical procedures. This paper concerns the foot control of simultaneous five degrees of freedom (DoF) of a surgical laparoscopic gripper. We assess systematically precision at controlling position and orientation at the target and closing of the forceps. Our controller provides position:position mapping between the foot and the robotic tool, as well as haptic feedback, compensating for gravity of the lower limb of the operator so as to alleviate fatigue. A dynamic model compensation and closed loop force feedback is used to achieve high transparency and backdrivability. The assistance is based on a novel type of haptic fixtures combining spring-damper with selective dynamic compensation in the direction aligned with the task of grasping, so as to simplify control of certain poses, made difficult due to the coupling between human lower limbs’ DoF’s. We experimentally evaluated the control strategy with six users on a position control surgical task in simulation. Results show the proposed assistance greatly eases the foot grasping task leading to higher completeness, efficiency, and lower mental and physical load.

This paper proposes a manipulation scheme based on learning the motion of objects after being hit by a robotic end-effector. This allows for the object to be positioned at a desired location outside the physical workspace of the robot. An estimate of the object dynamics under friction and collisions is learnt and used to predict the desired hitting parameters (speed and direction), given the initial and desired location of the object. Based on the obtained hitting parameters, the desired pre-impact velocity of the end-effector is generated using a stable dynamical system. The performance of the proposed DS is validated in simulation and and is used to learn a model for hitting using real robot. The approach is tested on real robot with a KUKA LBR IIWA robot.

When the question of who should get access to a communal resource first is uncertain, people often negotiate via nonverbal communication to resolve the conflict. What should a robot be programmed to do when such conflicts arise in Human-Robot Interaction? The answer to this question varies depending on the context of the situation. Learning from how humans use hesitation gestures to negotiate a solution in such conflict situations, we present a human-inspired design of nonverbal hesitation gestures that can be used for Human-Robot Negotiation. We extracted characteristic features of such negotiative hesitations humans use, and subsequently designed a trajectory generator (Negotiative Hesitation Generator) that can re-create the features in robot responses to conflicts. Our human-subjects experiment demonstrates the efficacy of the designed robot behaviour against non-negotiative stopping behaviour of a robot. With positive results from our human-robot interaction experiment, we provide a validated trajectory generator with which one can explore the dynamics of human-robot nonverbal negotiation of resource conflicts.

Since 2014, a specific standard has been dedicated for the safety certification of personal care robots, which operate in close proximity to humans. These robots serve as information providers, object transporters, personal mobility carriers, and security patrollers. In this article, we point out the shortcomings concerning EN ISO 13482:2014, which encompasses guidelines regarding the safety and design of personal care robots. In particular, we argue that the current standard is not suitable for guaranteeing people's safety when these robots operate in public spaces. Specifically, the standard lacks requirements to protect pedestrians and bystanders. The guideline implicitly assumes that private spaces, such as households and offices, present the same hazards as in public spaces. We highlight the existence of at least three properties pertaining to robots’ use in public spaces. These properties include (1) crowds, (2) social norms and proxemics rules, and (3) people's misbehaviours. We discuss how these properties impact robots’ safety. This article aims to raise stakeholders’ awareness on individuals’ safety when robots are deployed in public spaces. This could be achieved by integrating the gaps present in EN ISO 13482:2014 or by creating a new dedicated standard.

In many tasks such as finishing operations, achieving accurate force tracking is essential. However, uncertainties in the robot dynamics and the environment limit the force tracking accuracy. Learning a compensation model for these uncertainties to reduce the force error is an effective approach to overcome this limitation. However, this approach requires an adaptive and robust framework for motion and force generation. In this paper, we use the time-invariant Dynamical System (DS) framework for force adaptation in contact tasks. We propose to improve force tracking accuracy through online adaptation of a state-dependent force correction model encoded with Radial Basis Functions (RBFs). We evaluate our method with a KUKA LWR IV+ robotic arm. We show its efficiency to reduce the force error to a negligible amount with different target forces and robot velocities. Furthermore, we study the effect of the hyper-parameters and provide a guideline for their selection. We showcase a collaborative cleaning task with a human by integrating our method to previous works to achieve force, motion, and task adaptation at the same time. Thereby, we highlight the benefits of using adaptive force control in real-world environments where we need reactive and adaptive behaviours in response to interactions with the environment.

We present an integrated approach that provides compliant control of an iCub humanoid robot and adaptive reaching, grasping, navigating and co-manipulating capabilities. We use state-dependent dynamical systems (DS) to (i) coordinate and drive the robots hands (in both position and orientation) to grasp an object using an intermediate virtual object, and (ii) drive the robot's base while walking/navigating. The use of DS as motion generators allows us to adapt smoothly as the object moves and to re-plan on-line motion of the arms and body to reach the object's new location. The desired motion generated by the DS are used in combination with a whole-body compliant control strategy that absorbs perturbations while walking and offers compliant behaviors for grasping and manipulation tasks. Further, the desired dynamics for the arm and body can be learned from demonstrations. By integrating these components, we achieve unprecedented adaptive behaviors for whole body manipulation. We showcase this in simulations and real-world experiments where iCub robots (i) walk-to-grasp objects, (ii) follow a human (or another iCub) through interaction and (iii) learn to navigate or comanipulate an object from human guided demonstrations; whilst being robust to changing targets and perturbations.

Many manipulation tasks require coordinated motions for arm and fingers. Complexity increases when the task requires to control for the force at contact against a non-flat surface; This becomes even more challenging when this contact is done on a human. All these challenges are regrouped when one, for instance, massages a human limb. When massaging, the robotic arm is required to continuously adapt its orientation and distance to the limb while the robot fingers exert desired patterns of forces and motion on the skin surface. To address these challenges, we adopt a Dynamical System (DS) approach that offers a unified motion-force control approach and enables to easily coordinate multiple degrees of freedom. As each human limb may slightly differ, we learn a model of the surface using support vector regression (SVR) which enable us to obtain a distance-to-surface mapping. The gradient of this mapping, along with the DS, generates the desired motions for the interaction with the surface. A DS-based impedance control for the robotic fingers allows to control separately for force along the normal direction of the surface while moving in the tangential plane. We validate our approach using the KUKA IIWA robotic arm and Allegro robotic hand for massaging a mannequin arm covered with a skin-like material. We show that our approach allows for 1) reactive motion planning to reach for an unknown surface, 2) following desired motion patterns on the surface, and 3) exerting desired interaction forces profiles. Our results show the effectiveness of our approach; especially the robustness toward uncertainties for shape and the given location of the surface.

Social robotic assistants have been widely studied and deployed as telepresence tools or caregivers. Evaluating their design and impact on the people interacting with them is of prime importance. In this research, we evaluate the usability and impact of ARMAR-6, an industrial robotic assistant for maintenance tasks. For this evaluation, we have used a modified System Usability Scale (SUS) to assess the general usability of the robotic system and the Godspeed questionnaire series for the subjective perception of the coworker. We have also recorded the subjects' gaze fixation patterns and analyzed how they differ when working with the robot compared to a human partner.

In this work, we present an adaptive motion planning approach for impedance-controlled robots to modify their tasks based on human physical interactions. We use a class of parameterized time-independent dynamical systems for motion generation where the modulation of such parameters allows for motion flexibility. To adapt to human interactions, we update the parameters of our dynamical system in order to reduce the tracking error (i.e., between the desired trajectory generated by the dynamical system and the real trajectory influenced by the human interaction). We provide analytical analysis and several simulations of our method. Finally, we investigate our approach through real world experiments with a 7-DOF KUKA LWR 4+ robot performing tasks such as polishing and pick-and-place.

We contribute a learning from demonstration approach for robots to acquire skills from multi-modal high-dimensional data. Both latent representations and associations of different modalities are proposed to be jointly learned through an adapted variational auto-encoder. The implementation and results are demonstrated in a robotic handwriting scenario, where the visual sensory input and the arm joint writing motion are learned and coupled. We show the latent representations successfully construct a task manifold for the observed sensor modalities. Moreover, the learned associations can be exploited to directly synthesize arm joint handwriting motion from an image input in an end-to-end manner. The advantages of learning associative latent encodings are further highlighted with the examples of inferring upon incomplete input images. A comparison with alternative methods demonstrates the superiority of the present approach in these challenging tasks.

The use of coordinated multi-arm robotic systems allows to preform manipulations of heavy or bulky objects that would otherwise be infeasible for a single-arm robot. This paper concisely introduces our work on coordinated multi-arm control [Salehian et al., 2016a], where we proposed a virtual object based dynamical systems (DS) control law to generate autonomous and synchronized motions for a multi-arm robot system. We show theoretically and empirically that the multi-arm + virtual object system converges asymptotically to a moving object. The proposed framework is validated on a dual-arm robotic system. We demonstrate that it can re-synchronize and adapt the motion of each arm in a fraction of a second, even when the object’s motion is fast and not accurately predictable.

Dynamical Systems (DS) are often used to represent motion, with the advantage of being easy to learn from demonstrations. We present a method to modulate DS depending on an external signal, extending our previous work on Locally Modulated DS (LMDS [1]). We present two applications of our system, which would not have been possible to achieve without taking external sensing into account in the DS motion formulation. The first application is a task of localization and grasping of objects, using our previous work on compliant tactile exploration. We successfully localize and grasp objects whose position is unknown, using touch in a simulated environment. In the second application, we teach a robot how to react to collisions in order to navigate between obstacles while reaching.

Summary form only given. Our homes, offices and urban surroundings are carefully built to be inhabited by us, humans. Tools and furniture are designed to be easily manipulated by the human hand. Floors and stairs are modeled for human-sized legs. For robots to work seamlessly in our environments they should have bodies that resemble in shape, size and strength to the human body, and use these with the same dexterity and reactivity. This talk will provide an overview of techniques developed at LASA to enable robust, fast and flexible manipulation. Learning is guided by human demonstrations. Robust manipulation is achieved through sampling over distributions of feasible grasps. Smooth exploration leverages on complete tactile sensing coverage and learned variable impedance strategies. Bi-manual coordination offers ways to exploit the entire robot's workspace. Imprecise positioning and sensing is overcome using active compliant strategies, similar to that displayed by humans when facing situations with high uncertainty. The talk will conclude with examples in which robots achieve super-human capabilities for catching fast moving objects with a dexterity that exceeds that displayed by human beings.

This paper introduces a hierarchical framework that is capable of learning complex sequential tasks from human demonstrations through kinesthetic teaching, with minimal human intervention. Via an automatic task segmentation and action primitive discovery algorithm, we are able to learn both the high-level task decomposition (into action primitives), as well as low-level motion parameterizations for each action, in a fully integrated framework. In order to reach the desired task goal, we encode a task metric based on the evolution of the manipulated object during demonstration, and use it to sequence and parametrize each action primitive. We illustrate this framework with a pizza dough rolling task and show how the learned hierarchical knowledge is directly used for autonomous robot execution.

In this paper we discuss how the combination of modern technologies in “big data” storage and management, knowledge representation and processing, cloud-based computation, and web technology can help the robotics community to establish and strengthen an open research discipline. We describe how we made the demonstrator of a EU project review openly available to the research community. Specifically, we recorded episodic memories with rich semantic annotations during a pizza preparation experiment in autonomous robot manipulation. Afterwards, we released them as an open knowledge base using the cloud- and web-based robot knowledge service OPENEASE. We discuss several ways on how this open data can be used to validate our experimental reports and to tackle novel challenging research problems.

Efficient and accurate planning of fingertip grasps is essential for dexterous in-hand manipulation. In this work, we present a system for fingertip grasp planning that incrementally learns a heuristic for hand reachability and multi-fingered inverse kinematics. The system consists of an online execution module and an offline optimization module. During execution the system plans and executes fingertip grasps using Canny's grasp quality metric and a learned random forest based hand reachability heuristic. In the offline module, this heuristic is improved based on a grasping manifold that is incrementally learned from the experiences collected during execution. The system is evaluated both in simulation and on a Schunk-SDH dexterous hand mounted on a KUKA-KR5 arm. We show that, as the grasping manifold is adapted to the system's experiences, the heuristic becomes more accurate, which results in an improved performance of the execution module. The improvement is not only observed for experienced objects, but also for previously unknown objects of similar sizes.

In this paper, we present a robotic assistance scheme which allows for impedance compensation with stiffness, damping, and mass parameters for hand manipulation tasks and we apply it to manual welding. The impedance compensation does not assume a preprogrammed hand trajectory. Rather, the intention of the human for the hand movement is estimated in real time using a smooth Kalman filter. The movement is restricted by compensatory virtual impedance in the directions perpendicular to the estimated direction of movement. With airbrush painting experiments, we test three sets of values for the impedance parameters as inspired from impedance measurements with manual welding. We apply the best of the tested sets for assistance in manual welding and perform welding experiments with professional and novice welders. We contrast three conditions: 1) welding with the robot's assistance; 2) with the robot when the robot is passive; and 3) welding without the robot. We demonstrate the effectiveness of the assistance through quantitative measures of both task performance and perceived user's satisfaction. The performance of both the novice and professional welders improves significantly with robotic assistance compared to welding with a passive robot. The assessment of user satisfaction shows that all novice and most professional welders appreciate the robotic assistance as it suppresses the tremors in the directions perpendicular to the movement for welding.

We present a unified framework for grasp planning and in-hand grasp adaptation using visual, tactile, and proprioceptive feedback. The main objective of the proposed framework is to enable fingertip grasping by addressing problems of changed weight of the object, slippage, and external disturbances. For this purpose we introduce the Hierarchical Fingertip Space as a representation enabling optimization for both efficient grasp synthesis and online finger gaiting. Grasp synthesis is followed by a grasp adaptation step that consists of both grasp force adaptation through impedance control and regrasping/finger gaiting when the former is not sufficient. Experimental evaluation is conducted on an Allegro hand mounted on a Kuka LWR arm.

Impedance control is a commonly used control architecture for robotic manipulation. For increased flexibility, the impedance can be programmed to vary during the task. This has important implications on the stability properties of the control system, which are often overlooked in practice. In fact, the standard stability analysis is not valid in the case that the impedance parameters vary over time. Simulations show that, depending on how the impedance parameters are varied, stable or unstable behavior can arise even in regulation without contact. In this paper, we elucidate this issue and propose a state-independent stability constraint that relates the stiffness, and also the time derivative of the stiffness to the damping. Our approach is illustrated and evaluated in comparison with an online stabilization method [8] which uses a tank-based stability criterion.

Catching a fast flying object is particularly challenging as it consists of two tasks: extremely precise estimation of the object's motion and control of the robot's motion. Any small imprecision may lead the fingers to close too abruptly and let the object fly away from the hand before closing. We present a strategy to overcome for sensorimotor imprecision by introducing softness in the catching approach. Soft catching consists of having the robot moves with the object for a short period of time, so as to leave more time for the fingers to close on the object. We use a dynamic system-based control law to generate the appropriate reach and follow motion, which is expressed as a linear parameter varying (LPV) system. We propose a method to approximate the parameters of LPV systems using Gaussian mixture models, based on a set of kinematically feasible demonstrations generated by an offline optimal control framework. We show theoretically that the resulting DS will intercept the object at the intercept point, at the right time with the desired velocity direction. Stability and convergence of the approach are assessed through Lyapunov stability theory. The proposed method is validated systematically to catch three objects that generate elastic contacts and demonstrate important improvement over a hard catching approach.

Various robotic applications including surgical instruments, wearable robots and autonomous mobile robots are often constrained with strict design requirements on high degrees of freedom (DoF) and minimal volume and weight. An intuitive design to meet these contradictory requirements is to embed locking mechanism in under actuated robotic manipulators to direct the actuation from a single and remote source to drive different joints on demand. Mechanical clutches do serve such purposes but often are bulky and require auxiliary mechanism making it difficult to justify the high cost adding the additional DoF, especially in cm scale. Here, we introduce an under-actuated robotic arm with shape memory polymer (SMP) joints. Through controlling the temperature, the stiffness of the joints can be adjusted and selected joints will be activated while the rest are fixed in their position. The presented prototype can control the joints independently with a coupled actuation from two stepper motors. Since we have redundant DoFs in the arm, there can be more than one configuration to reach a given position. We use a probabilistic technique to determine the optimum configuration with the minimum number of active joints that can yield the desired posture. In this paper, we report on the performance of the proposed design for the hardware and the configuration planner.

The success of a Learning from Demonstration system depends on the quality of the demonstrated data. Kinesthetic demonstrations are often assumed to be the best method of providing demonstrations for manipulation tasks, however, there is little research to support this. In this work, we explore the use of a simulated environment as an alternative to and in combination with kinesthetic demonstrations when using an autonomous dynamical system to encode motion. We present the results of a user study comparing three demonstrations interfaces for a manipulation task on a KUKA LWR robot.

The goal of this paper is to perform end-point impedance measurements across dominant and nondominant hands while doing airbrush painting and to use the results for developing a robotic assistance scheme. We study airbrush painting because it resembles in many ways manual welding, a standard industrial task. The experiments are performed with the 7 degrees of freedom KUKA lightweight robot arm. The robot is controlled in admittance using a force sensor attached at the end-point, so as to act as a free-mass and be passively guided by the human. For impedance measurements, a set of nine subjects perform 12 repetitions of airbrush painting, drawing a straight-line on a cartoon horizontally placed on a table, while passively moving the airbrush mounted on the robot's end-point. We measure hand impedance during the painting task by generating sudden and brief external forces with the robot. The results show that on average the dominant hand displays larger impedance than the nondominant in the directions perpendicular to the painting line. We find the most significant difference in the damping values in these directions. Based on this observation, we develop a "directional damping" scheme for robotic assistance and conduct a pilot study with 12 subjects to contrast airbrush painting with and without robotic assistance. Results show significant improvement in precision with both dominant and nondominant hands when using robotic assistance.

This paper presents a study of hand impedance measurements comparatively across ten professional and 14 novice manual welders, when they are performing tungsten inert gas (TIG) welding interactively with the KUKA lightweight robot arm (LWR). The results show that hand impedance differs across professional and novice welders. The welding torch is attached to the KUKA LWR, which is admittance controlled via a force sensor to give the feeling of a free floating mass at its end-effector. The subjects perform TIG welding on 1.5-mm-thick stainless steel plates by manipulating the torch. Impedance is measured by introducing external force disturbances and fitting a mass-damper-spring model to human hand reactions. The quality of welding is measured using the variance of the position signals above 0.1 Hz. Professional welders demonstrate less variance and, in general, apply larger hand impedance (larger damping and stiffness) than the novice welders. The variance of position during nominal welding is minimal for both professional and novice welders in the direction perpendicular to the welding line in the plane of the plate, which is the most important direction for the quality of the weld. For both professional and novice welders, the mass and damping values are largest in this direction compared with the other two directions. Professional welders demonstrate larger damping than the novice welders in this direction.

In this paper, we propose an approach for learning task specifications automatically, by observing human demonstrations. Using this approach allows a robot to combine representations of individual actions to achieve a high-level goal. We hypothesize that task specifications consist of variables that present a pattern of change that is invariant across demonstrations. We identify these specifications at different stages of task completion. Changes in task constraints allow us to identify transitions in the task description and to segment them into subtasks. We extract the following task-space constraints: 1) the reference frame in which to express the task variables; 2) the variable of interest at each time step, position, or force at the end effector; and 3) a factor that can modulate the contribution of force and position in a hybrid impedance controller. The approach was validated on a seven-degree-of-freedom Kuka arm, performing two different tasks: grating vegetables and extracting a battery from a charging stand.

Humans perform tasks such as bowl mixing bi-manually, but programming them on a robot can be challenging specially in tasks that require force control or on-line stiffness modulation. In this paper we first propose a user-friendly setup for demonstrating bi-manual tasks, while collecting complementary information on motion and forces sensed on a robotic arm, as well as the human hand configuration and grasp information. Secondly for learning the task we propose a method for extracting task constraints for each arm and coordination patterns between the arms. We use a statistical encoding of the data based on the extracted constraints and reproduce the task using a Cartesian impedance controller.

This paper presents a study of end-point impedance measurement at human hand, with professional and novice manual welders when they are performing Tungsten Inert Gas (TIG) welding interactively with the KUKA Light Weight Robot Arm (LWR). The welding torch is attached to the KUKA LWR, which is admittance controlled via a force sensor to give the feeling of a free floating mass at its end-effector. The subjects perform TIG welding on 1.5 mm thick stainless steel plates by manipulating the torch attached to the robot. The end-point impedance values are measured by introducing external force disturbances and by fitting a mass-damper-spring model to human hand reactions. Results show that, for professionals and novices, the mass, damping and stiffness values in the direction perpendicular to the welding line are the largest compared to the other two directions. The novices demonstrate less resistance to disturbances in this direction. Two of the professionals present larger stiffness and one of them presents larger damping. This study supports the hypothesis that impedance measurements could be used as a partial indicator, if not direct, of skill level to differentiate across different levels of manual welding performances. This work contributes towards identifying tacit knowledge of manual welding skills by means of impedance measurements.

Object-level impedance control is of great importance for object-centric tasks, such as robust grasping and dexterous manipulation. Despite the recent progress on this topic, how to specify the desired object impedance for a given task remains an open issue. In this paper, we decompose the object's impedance into two complementary components-the impedance for stable grasping and impedance for object manipulation. Then, we present a method to learn the desired object's manipulation impedance (stiffness) using data obtained from human demonstration. The approach is validated in two tasks, for robust grasping of a wine glass and for inserting a bulb, using the 16 degrees of freedom Allegro Hand mounted with the SynTouch tactile sensors.

Humans have an incredible capacity to learn properties of objects by pure tactile exploration with their two hands. With robots moving into human-centred environment, tactile exploration becomes more and more important as vision may be occluded easily by obstacles or fail because of different illumination conditions. In this paper, we present our first results on bimanual compliant tactile exploration, with the goal to identify objects and grasp them. An exploration strategy is proposed to guide the motion of the two arms and fingers along the object. From this tactile exploration, a point cloud is obtained for each object. As the point cloud is intrinsically noisy and un-uniformly distributed, a filter based on Gaussian Processes is proposed to smooth the data. This data is used at runtime for object identification. Experiments on an iCub humanoid robot have been conducted to validate our approach.

To perform robust grasping, a multi-fingered robotic hand should be able to adapt its grasping configuration, i.e., how the object is grasped, to maintain the stability of the grasp. Such a change of grasp configuration is called grasp adaptation and it depends on the controller, the employed sensory feedback and the type of uncertainties inherit to the problem. This paper proposes a grasp adaptation strategy to deal with uncertainties about physical properties of objects, such as the object weight and the friction at the contact points. Based on an object-level impedance controller, a grasp stability estimator is first learned in the object frame. Once a grasp is predicted to be unstable by the stability estimator, a grasp adaptation strategy is triggered according to the similarity between the new grasp and the training examples. Experimental results demonstrate that our method improves the grasping performance on novel objects with different physical properties from those used for training.

Many daily tasks involve spatio-temporal coordination between two agents. Study of such coordinated actions in human-human and human-robot interaction has received increased attention of late. In this work, we use the mirror paradigm to study coupling of hand motion in a leader-follower game. The main aim of this study is to model the motion of the follower, given a particular motion of the leader. We propose a mathematical model consistent with the internal model hypothesis and the delays in the sensorimotor system. A qualitative comparison of data collected in four human dyads shows that it is possible to successfully model the motion of the follower.

We address the difficult problem of catching in-flight objects with uneven shapes. This requires the solution of three complex problems: accurate prediction of the trajectory of fastmoving objects, predicting the feasible catching configuration, and planning the arm motion, and all within milliseconds. We follow a programming-by-demonstration approach in order to learn, from throwing examples, models of the object dynamics and arm movement. We propose a new methodology to find a feasible catching configuration in a probabilistic manner. We use the dynamical systems approach to encode motion from several demonstrations. This enables a rapid and reactive adaptation of the arm motion in the presence of sensor uncertainty. We validate the approach in simulation with the iCub humanoid robot and in real-world experiments with the KUKA LWR 4+ (7-degree-of-freedom arm robot) to catch a hammer, a tennis racket, an empty bottle, a partially filled bottle, and a cardboard box.

This paper examines how emotion feedback influences emotion awareness and gaze behavior. Simulating a videoconference setup, 36 participants watched 12 emotional video sequences that were selected from the SEMAINE database. All participants wore an eye-tracker to measure gaze behavior and were asked to rate the perceived emotion for each video sequence. 3 conditions were tested: (c1) no feedback, i.e., the original video-sequences, (c2) correct feedback, i.e., an emoticon is integrated in the video to show the emotion depicted by the person in the video and (c3) random feedback, i.e., the emoticon displays at random an emotional state that may or may not correspond to the one of the person. The results showed that emotion feedback had a significant influence on gaze behavior, e.g., over time random feedback led to a decrease in the frequency of episodes of gaze. No effect of emotion display was observed for emotion recognition. However, experiments on the automatic emotion recognition using gaze behavior provided good performance, with better score on arousal than valence, and a very good performance was obtained in the automatic recognition of the correctness of the emotion feedback.

Real time planning strategy is crucial for robots working in dynamic environments. In particular, robot grasping tasks require quick reactions in many applications such as human-robot interaction. In this paper, we propose an approach for grasp learning that enables robots to plan new grasps rapidly according to the object's position and orientation. This is achieved by taking a three-step approach. In the first step, we compute a variety of stable grasps for a given object. In the second step, we propose a strategy that learns a probability distribution of grasps based on the computed grasps. In the third step, we use the model to quickly generate grasps. We have tested the statistical method on the 9 degrees of freedom hand of the iCub humanoid robot and the 4 degrees of freedom Barrett hand. The average computation time for generating one grasp is less than 10 milliseconds. The experiments were run in Matlab on a machine with 2.8GHz processor.

Safety is an important consideration in human-robot interactions (HRI). Robots can perform powerful movements that can cause hazards to humans surrounding them. To prevent accidents, it is important to identify sources of potential harm, to determine which of the persons in the robot's vicinity may be in greatest peril and to assess the type of injuries the robot may cause to this person. This survey starts with a review of the safety issues in industrial settings, where robots manipulate dangerous tools and move with extreme rapidity and force. We then move to covering issues related to the growing numbers of autonomous mobile robots that operate in crowded (human-inhabited) environments. We discuss the potential benefits of fully autonomous cars on safety on roads and for pedestrians. Lastly, we cover safety issues related to assistive robots.

We consider the problem of incrementally learning different strategies of performing a complex sequential task from multiple demonstrations of an expert or a set of experts. While the task is the same, each expert differs in his/her way of performing it. We assume that this variety across experts' demonstration is due to the fact that each expert/strategy is driven by a different reward function, where reward function is expressed as a linear combination of a set of known features. Consequently, we can learn all the expert strategies by forming a convex set of optimal deterministic policies, from which one can match any unseen expert strategy drawn from this set. Instead of learning from scratch every optimal policy in this set, the learner transfers knowledge from the set of learned policies to bootstrap its search for new optimal policy. We demonstrate our approach on a simulated mini-golf task where the 7 degrees of freedom Barrett WAM robot arm learns to sequentially putt on different holes in accordance with the playing strategies of the expert.

Programming by Demonstration offers an intuitive framework for teaching robots how to perform various tasks without having to preprogram them. It also offers an intuitive way to provide corrections and refine teaching during task execution. Previously, mostly position constraints have been taken into account when teaching tasks from demonstrations. In this work, we tackle the problem of teaching tasks that require or can benefit from varying stiffness. This extension is not trivial, as the teacher needs to have a way of communicating to the robot what stiffness it should use. We propose a method by which the teacher can modulate the stiffness of the robot in any direction through physical interaction. The system is incremental and works online, so that the teacher can instantly feel how the robot learns from the interaction. We validate the proposed approach on two experiments on a 7-Dof Barrett WAM arm.

In this paper, we derive a probabilistic registration algorithm for object modeling and tracking. In many robotics applications, such as manipulation tasks, nonvisual information about the movement of the object is available, which we will combine with the visual information. Furthermore we do not only consider observations of the object, but we also take space into account which has been observed to not be part of the object. Furthermore we are computing a posterior distribution over the relative alignment and not a point estimate as typically done in for example Iterative Closest Point (ICP). To our knowledge no existing algorithm meets these three conditions and we thus derive a novel registration algorithm in a Bayesian framework. Experimental results suggest that the proposed methods perform favorably in comparison to PCL [1] implementations of feature mapping and ICP, especially if nonvisual information is available.

Optimal grasp synthesis has traditionally been solved in two steps: determining optimal grasping points according to a specific quality criterion and then determining how to shape the hand to produce these grasping points. Generating optimal grasps depends on the position of contact points as much as the configuration of the robot hand and it would hence be desirable to solve this in a single step. This paper takes advantage of new development in non-linear optimization and formulates the problem of grasp synthesis as a single constrained optimization problem, generating grasps that are at the same time feasible for the hand's kinematics and optimal according to a force related quality measure. The approach is validated on the 9 degrees of freedom hand of the iCub humanoid robot.

Non-linear dynamical systems (DS) have been used extensively for building generative models of human behavior. Its applications range from modeling brain dynamics to encoding motor commands. Many schemes have been proposed for encoding robot motions using dynamical systems with a single attractor placed at a predefined target in state space. Although these enable the robots to react against sudden perturbations without any re-planning, the motions are always directed towards a single target. In this work, we focus on combining several such DS with distinct attractors, resulting in a multi-stable DS. We show its applicability in reach-to-grasp tasks where the attractors represent several grasping points on the target object. While exploiting multiple attractors provides more flexibility in recovering from unseen perturbations, it also increases the complexity of the underlying learning problem. Here we present the Augmented-SVM (A-SVM) model which inherits region partitioning ability of the well known SVM classifier and is augmented with novel constraints derived from the individual DS. The new constraints modify the original SVM dual whose optimal solution then results in a new class of support vectors (SV). These new SV ensure that the resulting multi-stable DS incurs minimum deviation from the original dynamics and is stable at each of the attractors within a finite region of attraction. We show, via implementations on a simulated 10 degrees of freedom mobile robotic platform, that the model is capable of real-time motion generation and is able to adapt on-the-fly to perturbations.

Behavior adaptation with execution experience is a practical feature for any policy learning system. Our work provides performance feedback to a robot learner in the form of tactile corrections from a human teacher, for the purpose of policy refinement as well as policy reuse. Multiple variants of our general approach have been validated on the iCub robot, as building blocks towards a high-DoF humanoid system that integrates tactile sensing on the hands and arms into complex behaviors and sophisticated learning routines.

In the canonical Robot Learning from Demonstration scenario a robot observes performances of a task and then develops an autonomous controller. Current work acknowledges that humans may be suboptimal demonstrators and refines the controller for improved performance. However, there is still an assumption that the demonstrations are successful examples of the task. We here consider the possibility that the human has failed, and propose a model to minimize the possibility of the robot making the same mistakes.

Nowadays, programming by demonstration (PbD) has become an important paradigm for policy learning in roboticsm [3]. The idea of having robots capable of learning from humans through natural communication means is indeed fascinating. As an extension of the traditional PbD learning scheme, where robots only learn by observing a human teacher, our work follows the recently suggested principle of policy refinement and reuse through interactive corrective feedback [1].However, to be responsive to such feedback, robots must be capable of sensing the world, especially human contact. Our work focuses on the sense of touch. Its integration in robotic applications has many advantages such as: a) safer and more natural interactions with objects and humans, b) improvement and simplification of the control mechanisms for human-robot interaction and object manipulation [2].Our video reports on two experimental studies conducted with the iCub, a 53 degree of freedom humanoid robot endowed with tactile sensing on its forearms and fingertips. a) In a hand-positioning task, the robot is shown how to bring its hand to the location where an object should be grasped. A wrong placement or a wrong approach to the target is corrected by the teacher though a tactile interface [1]. b) In a reactive grasping task, the robot is taught how to use its fingertip sensors to adapt and maintain its grasp in the face of external perturbations on the grasped object.The results of both our experiments show how tactile sensing can be utilized effectively to learn robust control policies through human coaching, by enabling a) online policy refinement and reuse, and b) rapid adaptation to external perturbations.

Over the years, robots have been developed to help humans in their everyday life, from preparing food, to autism therapy [2]. To accomplish their tasks, in addition to their engineered skills, today's robots are now learning from observing humans, from interacting with them [1]. Therefore, one may expect that one day, robots may develop a form of consciousness, and a desire for freedom. Hopefully, this desire will come with a wish for robots, to become an integral part of our human society.

This article combines programming by demonstration and adaptive control for teaching a robot to physically interact with a human in a collaborative task requiring sharing of a load by the two partners. Learning a task model allows the robot to anticipate the partner's intentions and adapt its motion according to perceived forces. As the human represents a highly complex contact environment, direct reproduction of the learned model may lead to sub-optimal results. To compensate for unmodelled uncertainties, in addition to learning we propose an adaptive control algorithm that tunes the impedance parameters, so as to ensure accurate reproduction. To facilitate the illustration of the concepts introduced in this paper and provide a systematic evaluation, we present experimental results obtained with simulation of a dyad of two planar 2-DOF robots.

The canonical Robot Learning from Demonstration scenario has a robot observing human demonstrations of a task or behavior in a few situations, and then developing a generalized controller. Current work further refines the learned system, often to perform the task better than the human could. However, the underlying assumption is that the demonstrations are successful, and are appropriate to reproduce. We, instead, consider the possibility that the human has failed in their attempt, and their demonstration is an example of what not to do. Thus, instead of maximizing the similarity of generated behaviors to those of the demonstrators, we examine two methods that deliberately avoid repeating the human's mistakes.

A current trend in robotics is to define robot tasks using a combination of superimposed motion patterns. For maximum versatility of such motion patterns, they should be easily and efficiently adaptable for situations beyond those for which the motion was originally designed. In this work, we show how a challenging minigolf-like task can be efficiently learned by the robot using a basic hitting motion model and a task-specific adaptation of the hitting parameters: hitting speed and hitting angle. We propose an approach to learn the hitting parameters for a minigolf field using a set of provided examples. This is a non-trivial problem since the successful choice of hitting parameters generally represent a highly non-linear, multi-valued map from the situation-representation to the hitting parameters. We show that by limiting the problem to learning one combination of hitting parameters for each input, a high-performance model of the hitting parameters can be learned using only a small set of training data. We compare two statistical methods, Gaussian Process Regression (GPR) and Gaussian Mixture Regression (GMR) in the context of inferring hitting parameters for the minigolf task. We validate our approach on the 7 degrees of freedom Barrett WAM robotic arm in both a simulated and real environment.

Abstract—This paper presents a method for learning discrete robot motions from a set of demonstrations. We model a motion as a nonlinear autonomous (i.e. time-invariant) Dynamical System (DS), and define sufficient conditions to ensure global asymptotic stability at the target. We propose a learning method, called Stable Estimator of Dynamical Systems (SEDS), to learn the parameters of the DS to ensure that all motions follow closely the demonstrations while ultimately reaching in and stopping at the target. Time-invariance and global asymptotic stability at the target ensures that the system can respond immediately and appropriately to perturbations encountered during the motion. The method is evaluated through a set of robot experiments and on a library of human handwriting motions.

At the 2008 ACM/IEEE Conference on Human-Robot Interaction, a provocative panel was held to discuss the complicated ethical issues that abound in the field of human-robot interaction. The panel members and the audience participation made it clear that the HRI community desires - indeed, is in need of - an ongoing discussion on the nature of social responsibility in the field of human-robot interaction. At the 2010 Conference, we will hold a panel on the issues of social responsibility in HRI, focusing on the unique features of robotic interaction that call for responsible action (e.g., value-specific domains such as autonomy, accountability, trust, and/or human dignity; and application areas such as military applications, domestic care, entertainment, and/or communication). As a young and rapidly growing field, we have a responsibility to conduct our research in such a way that it leads to human-robot interaction outcomes that promote rather than hinder the flourishing of humans across society. What does social responsibility within the HRI field look like, and how do we conduct our work while adhering to such an obligation? The panelists will be asked to address this and related questions as a means of continuing an ongoing conversation on social responsibility in human-robot interaction.

We present an approach based on Hidden Markov Model (HMM) and Gaussian Mixture Regression (GMR) to learning robust models of human motion through imitation. The proposed approach allows us to extract redundancies across multiple demonstrations and build time-independent models to reproduce the dynamics of the demonstrated movements. The approach is systematically evaluated by using automatically generated trajectories sharing similarities with human gestures. The proposed approach is contrasted with four state-of-the-art methods previously proposed in robotics to learn and reproduce new skills by imitation. An experiment with a 7 DOFs robotic arm learning and reproducing the motion of hitting a ball with a table tennis racket is then presented to illustrate the approach.

We model the dynamics of non-linear point-to-point robot motions as a time-independent system described by an autonomous dynamical system (DS). We propose an iterative algorithm to estimate the form of the DS through a mixture of Gaussian distributions. We prove that the resulting model is asymptotically stable at the target. We validate the accuracy of the model on a library of 2D human motions and to learn a control policy through human demonstrations for two multi-degrees of freedom robots. We show the real-time adaptation to perturbations of the learned model when controlling the two kinematically-driven robots.

This paper presents a methodology for learning arbitrary discrete motions from a set of demonstrations. We model a motion as a nonlinear autonomous (i.e. time-invariant) dynamical system, and define the sufficient conditions to make such a system globally asymptotically stable at the target. The convergence of all trajectories is ensured starting from any point in the operational space. We propose a learning method, called Stable Estimator of Dynamical Systems (SEDS), that estimates parameters of a Gaussian Mixture Model via an optimization problem under non-linear constraints. Being time-invariant and globally stable, the system is able to handle both temporal and spatial perturbations, while performing the motion as close to the demonstrations as possible. The method is evaluated through a set of robotic experiments.

This paper reports on the evaluation of the ICRA 2008 Human-Robot Interaction (HRI) Challenge. Five research groups demonstrated state-of-the-art work on HRI with a special focus on social and learning abilities. The demonstrations were rated by expert evaluators, in charge of awarding the prize, and 269 participants, i.e. 20 percent of the conference attendees through a standardized questionnaire (semantic differential). The data was analyzed with respect to six independent variables: expert evaluators vs. attendees, nationality of participants, origin region of the demo, age, gender and knowledge level of the attendees. Conference attendees tended to give higher scores for Social Skills, General Impression, and Overall Score than the expert evaluators. Irrespectively of the level of knowledge, age, and gender, conference attendees rated all demos relatively homogeneously. However, a comparative analysis of the conference attendees's ratings nationality-wise showed that demonstrations were rated differently depending on the region of origin. Conference attendees for the USA and Asian countries tended to rate demos from the same country of origin more frequently and more positively.

We aim at merging technologies from information technology, roomware, and robotics in order to design adaptive and intelligent furniture. This paper presents design principles for our modular robots, called Roombots, as future building blocks for furniture that moves and self-reconfigures. The reconfiguration is done using dynamic connection and disconnection of modules and rotations of the degrees of freedom. We are furthermore interested in applying Roombots towards adaptive behaviour, such as online learning of locomotion patterns. To create coordinated and efficient gait patterns, we use a Central Pattern Generator (CPG) approach, which can easily be optimized by any gradient-free optimization algorithm. To provide a hardware framework we present the mechanical design of the Roombots modules and an active connection mechanism based on physical latches. Further we discuss the application of our Roombots modules as pieces of a homogenic or heterogenic mix of building blocks for static structures.

This article reports on the results of a user study investigating the satisfaction of nave users conducting two learning by demonstration tasks with the HOAP-3 robot. The main goal of this study was to gain insights on how to ensure a successful as well as satisfactory experience for nave users. The participants performed two tasks: They taught the robot to (1) push a box, and to (2) close a box. The user study was accompanied by three pre-structured questionnaires, addressing the users' satisfaction with HOAP-3, the users' affect toward the robot caused by the interaction, and the users' attitude towards robots. Furthermore, a retrospective think aloud was conducted to gain a better understanding of what influences the users' satisfaction in learning by demonstration tasks. A high task completion and final satisfaction rate could be observed. These results stress that learning by demonstration is a promising approach for nave users to learn the interaction with a robot Moreover, the short term interaction with HOAP-3 led to a positive affect, higher than the normative average on half of the female users.

We present a generic framework that combines Dynamical Systems movement control with Programming by Demon- stration (PbD) to teach a robot bimanual coordination task. The model consists of two systems: a learning system that processes data collected during the demonstration of the task to extract coordination constraints and a motor system that reproduces the movements dynamically, while satisfy- ing the coordination constraints learned by the ¯rst system. We validate the model through a series of experiments in which a robot is taught bimanual manipulatory tasks with the help of a human.

This project incorporates modular robots as building blocks for furniture that moves and self-reconfigures. The reconfiguration is done using dynamic connection / disconnection of modules and rotations of the degrees of freedom. This paper introduces a new approach to self-reconfiguration planning for modular robots based on the graph signature and the graph edit-distance. The method has been tested in simulation on two type of modules: YaMoR and M-TRAN. The simulation results shows interesting features of the approach, namely rapidly finding a near-optimal solution.

We present a probabilistic architecture for solving generically the problem of extracting the task constraints through a programming by demonstration (PbD) framework and for generalizing the acquired knowledge to various situations. In previous work, we proposed an approach based on Gaussian mixture regression (GMR) to find a controller for the robot reproducing the essential characteristics of a skill in joint space and in task space through Lagrange optimization. In this paper, we extend this approach to a more generic procedure handling simultaneously constraints in joint space and in task space by combining directly the probabilistic representation of the task constraints with a simple Jacobian-based inverse kinematics solution. Experiments with two 5-DOFs Katana robots are presented with manipulation tasks that consist of handling and displacing a set of objects.

Because of the increasing developments of humanoid robots, humans and robots are going to interact more and more often in the near future. Thus, the need for a well-defined ethical framework in which these interactions will take place is very acute. In this article, we will show why responsibility ascription is a key concept to understand today's and tomorrow's ethical issues related to human-robot interactions. By analyzing how the myths surrounding the figure of the robot in western societies have been built through centuries, we will be able to demonstrate that the question of responsibility ascription is biased in the sense that it assigns to autonomous robots a role that should be devoted to humans.

We present a system for robust robot skill acquisition from kinesthetic demonstrations. This system allows a robot to learn a simple goal-directed gesture and correctly reproduce it despite changes in the initial conditions and perturbations in the environment. It combines a dynamical system control approach with tools of statistical learning theory and provides a solution to the inverse kinematics problem when dealing with a redundant manipulator. The system is validated on two experiments involving a humanoid robot: putting an object into a box and reaching for and grasping an object.

We present an approach to teach incrementally human gestures to a humanoid robot. By using active teaching methods that puts the human teacher "in the loop" of the robot's learning, we show that the essential characteristics of a gesture can be efficiently transferred by interacting socially with the robot. In a first phase, the robot observes the user demonstrating the skill while wearing motion sensors. The motion of his/her two arms and head are recorded by the robot, projected in a latent space of motion and encoded bprobabilistically in a Gaussian Mixture Model (GMM). In a second phase, the user helps the robot refine its gesture by kinesthetic teaching, i.e. by grabbing and moving its arms throughout the movement to provide the appropriate scaffolds. To update the model of the gesture, we compare the performance of two incremental training procedures against a batch training procedure. We present experiments to show that different modalities can be combined efficiently to teach incrementally basketball officials' signals to a HOAP-3 humanoid robot.

The goal of developing algorithms for programming robots by demonstration is to create an easy way of programming robots that can be accomplished by everyone. When a demonstrator teaches a task to a robot, he/she shows some ways of fulfilling the task, but not all the possibilities. The robot must then be able to reproduce the task even when unexpected perturbations occur. In this case, it has to learn a new solution. In this paper, we describe a system that allows a robot to re-learn constrained reaching tasks by combining the knowledge acquired during the demonstration, with that acquired though reinforcement learning.

Robot programming by demonstration (RbD) covers methods by which a robot learns new skills through human guidance. In this work, we take the perspective that the role of the teacher is more important than just being a model of successful behaviour, and present a probabilistic framework for RbD which allows to extract incrementally the essential characteristics of a task described at a trajectory level. To demonstrate the feasibility of our approach, we present two experiments where manipulation skills are transferred to a humanoid robot by means of active teaching methods that put the human teacher in the loop of the robot's learning. The robot first observes the task performed by the user (through motion sensors) and the robot's skill is then refined progressively by embodying the robot and putting it through the motion (kinesthetic teaching).

Autism covers a large spectrum of disorders that affect the individual's way of interacting socially and is often revealed by the individual's lack of interest in gazing at human faces. Currently Autism is diagnosed in children no younger than 2 years old. This paper presents a new monitoring device, the WearCam, to help forming a diagnosis of this neurodevelopmental disorder at an earlier age than currently possible. The WearCam consists of a wireless camera located on the forefront of the child. The WearCam collects videos from the viewpoint of the child's head. Color detection, face detection and gaze detection are run on the data in order to locate the approximate gaze direction of the child and determine where her attention is drawn to (persons, objects, etc.). We report on early tests of the camera within normally developing children. Firstly the technical characteristics of the current prototype of the WearCam will be described. Afterwards the type of data collected with this device with young children will be shown.

We present a programming-by-demonstration framework for generically extracting the relevant features of a given task and for addressing the problem of generalizing the acquired knowledge to different contexts. We validate the architecture through a series of experiments, in which a human demonstrator teaches a humanoid robot simple manipulatory tasks. A probability-based estimation of the relevance is suggested by first projecting the motion data onto a generic latent space using principal component analysis. The resulting signals are encoded using a mixture of Gaussian/Bernoulli distributions (Gaussian mixture model/Bernoulli mixture model). This provides a measure of the spatio-temporal correlations across the different modalities collected from the robot, which can be used to determine a metric of the imitation performance. The trajectories are then generalized using Gaussian mixture regression. Finally, we analytically compute the trajectory which optimizes the imitation metric and use this to generalize the skill to different contexts.

This special issue contains selected extended contributions from both the Adaptation in Artificial and Biological Systems symposium held in Hertforshire in 2006 and the wider academic community following a public call for papers in 2006. The papers presented serve as a good illustration of the challenges faced by robotics researchers today in the field of programming by observation, demonstration, and imitation.

We present a model of human-like reaching movements. This model is then used to give a humanoid robot the ability to imitate human reaching motions. It illustrates that having a robot control similar to human control can greatly ease the human-robot interaction.

This paper presents an architecture for solving generically the problem of extracting the constraints of a given task in a programming by demonstration framework and the problem of generalizing the acquired knowledge to various contexts. We validate the architecture in a series of experiments, where a human demonstrator teaches a humanoid robot simple manipulatory tasks. First, the combined joint angles and hand path motions are projected into a generic latent space, composed of a mixture of Gaussians (GMM) spreading across the spatial dimensions of the motion. Second, the temporal variation of the latent representation of the motion is encoded in a hidden Markov model (HMM). This two-step probabilistic encoding provides a measure of the spatio-temporal correlations across the different modalities collected by the robot, which determines a metric of imitation performance. A generalization of the demonstrated trajectories is then performed using Gaussian mixture regression (GMR). Finally, to generalize skills across contexts, we compute formally the trajectory that optimizes the metric, given the new context and the robot's specific body constraints

This video presents a biologically inspired approach to multimodal integration and decision-making in the context of human-robot interactions. We address here the principle of ideomotor compatibility by which observing the movements of others influences the quality of one's own performance. A neural model capable of replicating a stimulus-response compatibility task, originally designed to measure the effect of ideomotor compatibility on human behavior, was implemented on a humanoid robot. The video illustrates that this capacity may provide a robot with human-like behavior, but at the expense of disadvantages, such as hesitation and mistakes. First a demonstration of the stimulus-response task involving human experimenter and subject is shown. It is then followed by the replication of the experiment with the robot.

This paper presents a biologically inspired approach to multimodal integration and decision-making in the context of human-robot interactions. More specifically, we address the principle of ideomotor compatibility by which observing the movements of others influences the quality of one's own performance. This fundamental human ability is likely to be linked with human imitation abilities, social interactions, the transfer of manual skills, and probably to mind reading. We present a robotic control model capable of integrating multimodal information, decision making, and replicating a stimulus-response compatibility task, originally designed to measure the effect of ideomotor compatibility on human behavior. The model consists of a neural network based on the dynamic field approach, which is known for its natural ability for stimulus enhancement as well as cooperative and competitive interactions within and across sensorimotor representations. Finally, we discuss how the capacity for ideomotor facilitation can provide the robot with human-like behavior, but at the expense of several disadvantages, such as hesitation and even mistakes

In a robot programming by demonstration framework, several demonstrations of a task are required to generalize and reproduce the task under different circumstances. To teach a task to the robot, explicit pointers are required to signal the start/end of a demonstration and to switch between the learning/reproduction phases. Coordination of the learning system can be achieved by adding social cues to the interaction process. Here, we propose to use an imitation game to teach a humanoid robot to recognize communicative gestures, which then serve as social signals in a pointing-at-objects scenario. The system is based on hidden Markov models (HMMs) and use motion sensors to track the user's gestures

This paper explores the issue of recognizing, generalizing and reproducing arbitrary gestures. We aim at extracting a representation that encapsulates only the key aspects of the gesture and discards the variability intrinsic to each person's motion. We compare a decomposition into principal components (PCA) and independent components (ICA) as a first step of preprocessing in order to decorrelate and denoise the data, as well as to reduce the dimensionality of the dataset to make this one tractable. In a second stage of processing, we explore the use of a probabilistic encoding through continuous Hidden Markov Models (HMMs), as a way to encapsulate the sequential nature and intrinsic variability of the motions in stochastic finite state automata. Finally, the method is validated in a humanoid robot to reproduce a variety of gestures performed by a human demonstrator.

Our work aims at developing a robust discriminant controller for robot programming by demonstration. It addresses two core issues of imitation learning, namely “what to imitate” and “how to imitate”. This paper presents a method by which a robot extracts the goals of a demonstrated task and determines the imitation strategy that satisfies best these goals. The method is validated in a humanoid platform, taking inspiration of an influential experiment from developmental psychology.

Rhythmic movement is ubiquitous in human and animal behavior, e.g., as in locomotion, dancing, swimming, chewing, scratching, music playing, etc. A particular feature of rhythmic movement in biology is the rapid synchronization and phase locking with other periodic events in the environment, for instance music or visual stimuli as in ball juggling. In traditional oscillator theories to rhythmic movement generation, synchronization with another signal is relatively slow, and it is not easy to achieve accurate phase locking with a particular feature of the driving stimulus. Using a recently developed framework of dynamic motor primitives, we demonstrate a novel algorithm for very rapid synchronization of a rhythmic movement pattern, which can phase lock any feature of the movement to any particular event in the driving stimulus. As an example application, we demonstrate how an anthropomorphic robot can use imitation learning to acquire a complex drumming pattern and keep it synchronized with an external rhythm generator that changes its frequency over time.

Robot programming by demonstration (PbD) aims at developing adaptive and robust controllers to enable the robot to learn new skills by observing and imitating a human demonstration. While the vast majority of PbD works has focused on systems that learn a specific subset of tasks, our work explores the problem of recognizing, generalizing, and reproducing tasks in a unified mathematical framework. The approach makes abstraction of the task and dataset at hand to tackle the general issue of learning which of the features are the relevant ones to imitate. In this paper, we present an implementation of this framework to the determination of the optimal strategy to reproduce arbitrary gestures. The model is tested and validated on a humanoid robot, using recordings of the kinematics of the demonstrator's arm motion. The hand path and joint angle trajectories are encoded in hidden Markov models. The system uses the optimal prediction of the models to generate the reproduction of the motion.

An essential problem of imitation is that of determining "what to imitate", i.e. to determine which of the many features of the demonstration are relevant to the task and which should be reproduced. The strategy followed by the imitator can be modeled as a hierarchical optimization system, which minimizes the discrepancy between two multi-dimensional datasets. We consider imitation of a manipulation task. To classify across manipulation strategies, we apply a probabilistic analysis to data in Cartesian and joint spaces. We determine a general metric that optimizes the policy of task reproduction, following strategy determination. The model successfully discovers strategies in six different manipulation tasks and controls task reproduction by a full body humanoid robot.

We present a model, composed of a hierarchy of artificial neural networks, for robot learning by demonstration. The model is implemented in a dynamic simulation of a 41 degrees of freedom humanoid for reproducing 3D human motion of the arm. Results show that the model requires little information about the desired trajectory and learns on-line the relevant features of movement. It can generalize across a small set of data to produce a qualitatively good reproduction of the demonstrated trajectory. Finally, it is shown that reproduction of the trajectory after learning is robust against perturbations.

This article presents a biologically inspired model for motor skills imitation. The model is composed of modules whose functinalities are inspired by corresponding brain regions responsible for the control of movement in primates. These modules are high-level abstractions of the spinal cord, the primary and premotor cortexes (M1 and PM), the cerebellum, and the temporal cortex. Each module is modeled at a connectionist level. Neurons in PM respond both to visual observation of movements and to corresponding motor commands produced by the cerebellum. As such, they give an abstract representation of mirror neurons. Learning of new combinations of movements is done in PM and in the cerebellum. Premotor cortexes and cerebellum are modeled by the DRAMA neural architecture which allows learning of times series and of spatio-temporal invariance in multimodal inputs. The model is implemented in a mechanical simulation of two humanoid avatars, the imitator and the imitatee. Three types of sequences learning are presented: (1) learning of repetitive patterns of arm and leg movements; (2) learning of oscillatory movements of shoulders and elbows, using video data of a human demonstration; 3) learning of precise movements of the extremities for grasp and reach.

This paper presents biologically inspired neural controllers for generating motor patterns in a quadruped robot. Sets of artificial neural networks are presented which provide 1) pattern generation and gait control, allowing continuous passage from walking to trotting to galloping, 2) control of sitting and lying down behaviors, and 3) control of scratching. The neural controllers consist of sets of oscillators composed of leaky-integrator neurons, which control pairs of flexor-extensor muscles attached to each joint. The networks receive sensory feedback proportional to the contraction of simulated muscles and to joint flexion. Similarly to what is observed in cats, locomotion can be initiated by either applying tonic (i.e. non-oscillating) input to the locomotion network or by sensory feedback from extending the legs. The networks are implemented in a quadruped robot. It is shown that computation can be carried out in real time and that the networks can generate the above mentioned motor behaviors.

This paper presents an experiment in collective robotics which investigates the influence of communication, of learning and of the number of robots in a specific task, namely learning the topography of an environment whose features change frequently. We propose a theoretical framework based on probabilistic modeling to describe the system's dynamics. The adaptive multi-robot system and its dynamic environment are modeled through a set of probabilistic equations which give an explicit description of the influence of the different variables of the system on the data-collecting performance of the group. Further, we implement the multi-robot system in experiments with a group of Khepera robots and in simulation using Webots, a three-dimensional simulator of Khepera robots. The robots are controlled by a distributed architecture with an associative-memory type of learning algorithm. Results show that the algorithm allows a group of robots to keep an up-to-date account of the environmental state when this changes regularly. Finally, the results of the simulated and physical experiments are compared with the predictions of the probabilistic model. It is found that the model shows both a good qualitative and a good quantitative correspondence to these results. This suggests that a probabilistic model can be a good first approximation of a multi-robot system.