Generating intelligent robot behavior in contact-rich settings is a research problem where zeroth-order methods currently prevail. A major contributor to the success of such methods is their robustness in the face of non-smooth and discontinuous optimization landscapes that are characteristic of contact interactions, yet zeroth-order methods remain computationally inefficient. It is therefore desirable to develop methods for perception, planning and control in contact-rich settings that can achieve further efficiency by making use of first and second order information (i.e., gradients and Hessians). To facilitate this, we present a joint formulation of collision detection and contact modelling which, compared to existing differentiable simulation approaches, provides the following benefits: i) it results in forward and inverse dynamics that are entirely analytical (i.e. do not require solving optimization or root-finding problems with iterative methods) and smooth (i.e. twice differentiable), ii) it supports arbitrary collision geometries without needing a convex decomposition, and iii) its runtime is independent of the number of contacts. Through simulation experiments, we demonstrate the validity of the proposed formulation as a "physics for inference" that can facilitate future development of efficient methods to generate intelligent contact-rich behavior.

Generating motions for robots interacting with objects of various shapes is a complex challenge, further complicated by the robot’s geometry and multiple desired behaviors. While current robot programming tools (such as inverse kinematics, collision avoidance, and manipulation planning) often treat these problems as constrained optimization, many existing solvers focus on specific problem domains or do not exploit geometric constraints effectively. We propose an efficient first-order method, Augmented Lagrangian Spectral Projected Gradient Descent (ALSPG), which leverages geometric projections via Euclidean projections, Minkowski sums, and basis functions. We show that by using geometric constraints rather than full constraints and gradients, ALSPG significantly improves real-time performance. Compared to second-order methods like iLQR, ALSPG remains competitive in the unconstrained case. We validate our method through toy examples and extensive simulations, and demonstrate its effectiveness on a 7-axis Franka robot, a 6-axis P-Rob robot and a 1:10 scale car in real-world experiments. Source codes, experimental data and videos are available on the project webpage: https://sites.google.com/view/alspg-oc

Recent work has demonstrated the potential of diffusion models in robot bimanual skill learning. However, existing methods ignore the learning of posture-dependent task features, which are crucial for adapting dual-arm configurations to meet specific force and velocity requirements in dexterous bimanual manipulation. To address this limitation, we propose Manipulability-Aware Diffusion Policy (ManiDP), a novel imitation learning method that not only generates plausible bimanual trajectories, but also optimizes dual-arm configurations to better satisfy posture-dependent task requirements. ManiDP achieves this by extracting bimanual manipulability from expert demonstrations and encoding the encapsulated posture features using Riemannian-based probabilistic models. These encoded posture features are then incorporated into a conditional diffusion process to guide the generation of task-compatible bimanual motion sequences. We evaluate ManiDP on six real-world bimanual tasks, where the experimental results demonstrate a 39.33% increase in average manipulation success rate and a 0.45 improvement in task compatibility compared to baseline methods. This work highlights the importance of integrating posture-relevant robotic priors into bimanual skill diffusion to enable human-like adaptability and dexterity.

This paper proposes a novel whole-body impedance control method for the Collaborative dUal-arm Robot manIpulator (CURI) in Human-Robot Collaboration (HRC). The method enables CURI to adapt its physical behavior to human motion while following trajectories learned from human-human demonstrations. A whole-body impedance controller coordinates the robot joints to achieve desired Cartesian space impedance. Collaborative tasks are captured from human-human demonstrations and represented using a Task-parameterized Gaussian Mixture Model (TP-GMM). Electromyography (EMG) sensors record muscle activities to estimate human impedance profiles, which are then mimicked by a variable impedance controller. An adaptive parameter is introduced to adjust robot stiffness based on spatial displacement between the robot and human, ensuring safe and efficient interaction. Experimental validation through confrontational Tai Chi pulling/pushing tasks demonstrates the superiority of the proposed adaptive impedance method over the fixed impedance controller.

Imitation Learning offers a promising approach to learn directly from data without requiring explicit models, simulations, or detailed task definitions. During inference, actions are sampled from the learned distribution and executed on the robot. However, sampled actions may fail for various reasons, and simply repeating the sampling step until a successful action is obtained can be inefficient. In this work, we propose an enhanced sampling strategy that refines the sampling distribution to avoid previously unsuccessful actions. We demonstrate that by solely utilizing data from successful demonstrations, our method can infer recovery actions without the need for additional exploratory behavior or a high-level controller. Furthermore, we leverage the concept of diffusion model decomposition to break down the primary problem—which may require long-horizon history to manage failures—into multiple smaller, more manageable sub-problems in learning, data collection, and inference, thereby enabling the system to adapt to variable failure counts. Our approach yields a low-level controller that dynamically adjusts its sampling space to improve efficiency when prior samples fall short. We validate our method across several tasks, including door opening with unknown directions, object manipulation, and button-searching scenarios, demonstrating that our approach outperforms traditional baselines. Supplementary materials for this paper are available on our website: https://hri-eu.github.io/ccdp/.

Large image generation and vision models, combined with differentiable rendering technologies, have become powerful tools for generating paths that can be drawn or painted by a robot. However, these tools often overlook the intrinsic physicality of the human drawing/writing act, which is usually executed with skillful hand/arm gestures. Taking this into account is important for the visual aesthetics of the results and for the development of closer and more intuitive artist-robot collaboration scenarios. We present a method that bridges this gap by enabling gradient-based optimization of natural human-like motions guided by cost functions defined in image space. To this end, we use the sigma-lognormal model of human hand/arm movements, with an adaptation that enables its use in conjunction with a differentiable vector graphics (DiffVG) renderer. We demonstrate how this pipeline can be used to generate feasible trajectories for a robot by combining image-driven objectives with a minimum-time smoothing criterion. We demonstrate applications with generation and robotic reproduction of synthetic graffiti as well as image abstraction.

We integrate smoothing B-splines into a standard differentiable vector graphics (DiffVG) pipeline through linear mapping, and show how this can be used to generate smooth and arbitrarily long paths within image-based deep learning systems. We take advantage of derivative-based smoothing costs for parametric control of fidelity vs. simplicity tradeoffs, while also enabling stylization control in geometric and image spaces. The proposed pipeline is compatible with recent vector graphics generation and vectorization methods. We demonstrate the versatility of our approach with four applications aimed at the generation of stylized vector graphics: stylized space-filling path generation, stroke-based image abstraction, closed-area image abstraction, and stylized text generation.

In this article, we present a feedback control method for tactile coverage tasks such as cleaning or surface inspection. Although these tasks are challenging to plan due to the complexity of continuous physical interactions, the coverage target and progress can be effectively measured using a camera and encoded in a point cloud. We propose an ergodic coverage method that operates directly on point clouds, guiding the robot to spend more time on regions requiring more coverage. For robot control and contact behavior, we use geometric algebra to formulate a task-space impedance controller that tracks a line while simultaneously exerting a desired force along that line. We evaluate the performance of our method in kinematic simulations and demonstrate its applicability in real-world experiments on kitchenware.

Control of dynamic systems involving hybrid actions is a challenging task in robotics. To address this, we present a novel algorithm called Generalized Policy Iteration using Tensor Train (TTPI) that belongs to the class of Approximate Dynamic Programming (ADP). We use a low-rank tensor approximation technique called Tensor Train (TT) to approximate the state-value and advantage function which enables us to efficiently handle hybrid systems. We demonstrate the superiority of our approach over previous baselines for some benchmark problems with hybrid action spaces. Additionally, the robustness and generalization of the policy for hybrid systems are showcased through a real-world robotics experiment involving a non-prehensile manipulation task which is considered to be a highly challenging control problem.

Traditional strength and conditioning training relies on the utilization of free weights, such as weighted implements, to elicit external stimuli. However, this approach poses a significant challenge when attempting to modify or adjust the loads within a single training set. This paper introduces an innovative method for achieving adjustable loads during resistance training by leveraging physical Human-Robot Interaction (pHRI). The primary objective is to regulate targeted muscle activation through the use of Robo-Coach (robotic coach system). We first utilize a Task-Parameterized Gaussian Mixture Model (TP-GMM) to learn the motion of coach demonstration, which can be generalized for the trainees. The 3D path extracted from the generated trajectory is then projected onto a 2D plane with respect to the direction of the load. Furthermore, we propose a hybrid stiffness/position generator for online task execution. This generator determines the desired positions in the 2D plane according to the contact point displacements in the stimuli direction and, simultaneously, sets the desired stiffness based on the muscle activation feedback. Finally, the Robo-Coach is implemented with a variable impedance controller to achieve load-adjustable resistance training with the trainee. The biceps curl exercises were conducted and the results showed favorable performance, indicating the effectiveness of this approach.

In this work, we propose a novel approach to represent robot geometry as distance fields (RDF) that extends the principle of signed distance fields (SDFs) to articulated kinematic chains. Our method employs a combination of Bernstein polynomials to encode the signed distance for each robot link with high accuracy and efficiency while ensuring the mathematical continuity and differentiability of SDFs. We further leverage the kinematics chain of the robot to produce the SDF representation in joint space, allowing robust distance queries in arbitrary joint configurations. The proposed RDF representation is differentiable and smooth in both task and joint spaces, enabling its direct integration to optimization problems. Additionally, the 0-level set of the robot corresponds to the robot surface, which can be seamlessly integrated into whole-body manipulation tasks. We conduct various experiments in both simulations and with 7-axis Franka Emika robots, comparing against baseline methods, and demonstrating its effectiveness in collision avoidance and whole-body manipulation tasks. Project page: https://sites.google.com/view/lrdf/home

In this work, we are presenting an extension of the cooperative dual-task space (CDTS) in conformal geometric algebra. The CDTS was first defined using dual quaternion algebra and is a well established framework for the simplified definition of tasks using two manipulators. By integrating conformal geometric algebra, we aim to further enhance the geometric expressiveness and thus simplify the modeling of various tasks. We show this formulation by first presenting the CDTS and then its extension that is based around a cooperative pointpair. This extension keeps all the benefits of the original formulation that is based on dual quaternions, but adds more tools for geometric modeling of the dual-arm tasks. We also present how this CGACDTS can be seamlessly integrated with an optimal control framework in geometric algebra that was derived in previous work. In the experiments, we demonstrate how to model different objectives and constraints using the CGA-CDTS. Using a setup of two Franka Emika robots we then show the effectiveness of our approach using model predictive control in real world experiments.

Many real-world sequential manipulation tasks involve a combination of discrete symbolic search and continuous motion planning, collectively known as combined task and motion planning (TAMP). However, prevailing methods often struggle with the computational burden and intricate combinatorial challenges, limiting their applications for online replanning in the real world. To address this, we propose Dynamic Logic-Geometric Program (D-LGP), a novel approach integrating Dynamic Tree Search and global optimization for efficient hybrid planning. Through empirical evaluation on three benchmarks, we demonstrate the efficacy of our approach, showcasing superior performance in comparison to state-of-the-art techniques. We validate our approach through simulation and demonstrate its reactive capability to cope with online uncertainty and external disturbances in the real world. Project webpage: https://sites.google.com/view/dyn-lgp.

Machine learning methods have a groundbreaking impact in many application domains, but their application on real robotic platforms is still limited.Despite the many challenges associated with combining machine learning technology with robotics, robot learning remains one of the most promising directions for enhancing the capabilities of robots. When deploying learning-based approaches on real robots, extra effort is required to address the challenges posed by various real-world factors. To investigate the key factors influencing real-world deployment and to encourage original solutions from different researchers, we organized the Robot Air Hockey Challenge at the NeurIPS 2023 conference. We selected the air hockey task as a benchmark, encompassing low-level robotics problems and high-level tactics. Different from other machine learning-centric benchmarks, participants need to tackle practical challenges in robotics, such as the sim-to-real gap, low-level control issues, safety problems, real-time requirements, and the limited availability of real-world data. Furthermore, we focus on a dynamic environment, removing the typical assumption of quasi-static motions of other real-world benchmarks.The competition's results show that solutions combining learning-based approaches with prior knowledge outperform those relying solely on data when real-world deployment is challenging.Our ablation study reveals which real-world factors may be overlooked when building a learning-based solution.The successful real-world air hockey deployment of best-performing agents sets the foundation for future competitions and follow-up research directions.

Humans use tools to complete impact-aware tasks such as hammering a nail or playing tennis. The postures adopted to use these tools can significantly influence the performance of these tasks, where the force or velocity of the hand holding a tool plays a crucial role. The underlying motion planning challenge consists of grabbing the tool in preparation for the use of this tool with an optimal body posture. Directional manipulability describes the dexterity of force and velocity in a joint configuration along a specific direction. In order to take directional manipulability and tool affordances into account, we apply an optimal control method combining iterative linear quadratic regulator (iLQR) with the alternating direction method of multipliers (ADMM). Our approach considers the notion of tool affordances to solve motion planning problems, by introducing a cost based on directional velocity manipulability. The proposed approach is applied to impact tasks in simulation and on a real 7-axis robot, specifically in a nail-hammering task with the assistance of a pilot hole. Our comparison study demonstrates the importance of maximizing directional manipulability in impact-aware tasks.

Planning multi-contact motions in a receding horizon fashion requires a value function to guide the planning with respect to the future, e.g., building momentum to traverse large obstacles. Traditionally, the value function is approximated by computing trajectories in a prediction horizon (never executed) that foresees the future beyond the execution horizon. However, given the non-convex dynamics of multi-contact motions, this approach is computationally expensive. To enable online Receding Horizon Planning (RHP) of multi-contact motions, we find efficient approximations of the value function. Specifically, we propose a trajectory-based and a learning-based approach. In the former, namely RHP with Multiple Levels of Model Fidelity, we approximate the value function by computing the prediction horizon with a convex relaxed model. In the latter, namely Locally-Guided RHP, we learn an oracle to predict local objectives for locomotion tasks, and we use these local objectives to construct local value functions for guiding a short-horizon RHP. We evaluate both approaches in simulation by planning centroidal trajectories of a humanoid robot walking on moderate slopes, and on large slopes where the robot cannot maintain static balance. Our results show that locally-guided RHP achieves the best computation efficiency (95%-98.6% cycles converge online). This computation advantage enables us to demonstrate online receding horizon planning of our real-world humanoid robot Talos walking in dynamic environments that change on-the-fly.

Achieving reactive robot behavior in complex dynamic environments is still challenging as it relies on being able to solve trajectory optimization problems quickly enough, such that we can replan the future motion at frequencies which are sufficiently high for the task at hand. We argue that current limitations in Model Predictive Control (MPC) for robot manipulators arise from inefficient, high-dimensional trajectory representations and the negligence of time-optimality in the trajectory optimization process. Therefore, we propose a motion optimization framework that optimizes jointly over space and time, generating smooth and timing-optimal robot trajectories in joint-space. While being task-agnostic, our formulation can incorporate additional task-specific requirements, such as collision avoidance, and yet maintain real-time control rates, demonstrated in simulation and real-world robot experiments on closed-loop manipulation. For additional material, please visit https://sites.google.com/oxfordrobotics.institute/vp-sto.

Long-term non-prehensile planar manipulation is a challenging task for robot planning and feedback control. It is characterized by underactuation, hybrid control, and contact uncertainty. One main difficulty is to determine both the continuous and discrete contact configurations, e.g., contact points and modes, which requires joint logical and geometrical reasoning. To tackle this issue, we propose a demonstration-guided hierarchical optimization framework to achieve offline task and motion planning (TAMP). Our work extends the formulation of the dynamics model of the pusher-slider system to include separation mode with face switching mechanism, and solves a warm-started TAMP problem by exploiting human demonstrations. We show that our approach can cope well with the local minima problems currently present in the state-of-the-art solvers and determine a valid solution to the task. We validate our results in simulation and demonstrate its applicability on a pusher-slider system with a real Franka Emika robot in the presence of external disturbances. Project webpage: https://sites.google.com/view/dg-oc/.

Pushing is an essential motor skill involved in several manipulation tasks, and has been an important research topic in robotics. Recent works have shown that Deep Q-Networks (DQNs) can learn pushing policies (when, where to push, and how) to solve manipulation tasks, potentially in synergy with other skills (e.g. grasping). Nevertheless, DQNs often assume a fixed setting and task, which may limit their deployment in practice. Furthermore, they suffer from sparse-gradient backpropagation when the action space is very large, a problem exacerbated by the fact that they are trained to predict state-action values based on a single reward function aggregating several facets of the task, rendering the model training challenging. To address these issues, we propose a multi-head target-parameterized DQN to learn robotic manipulation tasks, in particular pushing policies, and make the following contributions: i) we show that learning to predict different reward and task aspects can be beneficial compared to predicting a single value function where reward factors are not disentangled; ii) we study several alternatives to generalize a policy by encoding the target parameters either into the network layers or visually in the input; iii) we propose a kernelized version of the loss function, allowing to obtain better, faster and more stable training performance. Extensive experiments on simulations validate our design choices, and we show that our architecture learned on simulated data can achieve high performance in a real-robot setup involving a Franka Emika robot arm and unseen objects.

Soft object manipulation tasks in domestic scenes pose a significant challenge for existing robotic skill learning techniques due to their complex dynamics and variable shape characteristics. Since learning new manipulation skills from human demonstration is an effective way for robot applications, developing prior knowledge of the representation and dynamics of soft objects is necessary. In this regard, we propose a pretrained soft object manipulation skill learning model, namely SoftGPT, that is trained using large amounts of exploration data, consisting of a three-dimensional heterogeneous graph representation and a GPT-based dynamics model. For each downstream task, a goal-oriented policy agent is trained to predict the subsequent actions, and SoftGPT generates the consequences of these actions. Integrating these two approaches establishes a thinking process in the robot's mind that provides rollout for facilitating policy learning. Our results demonstrate that leveraging prior knowledge through this thinking process can efficiently learn various soft object manipulation skills, with the potential for direct learning from human demonstrations.

This paper presents a study on the use of the Talos humanoid robot for performing assistive sit-to-stand or stand-to-sit tasks. In such tasks, the human exerts a large amount of force (100–200 N) within a very short time (2–8 s), posing significant challenges in terms of human unpredictability and robot stability control. To address these challenges, we propose an approach for finding a spatial reference for the robot, which allows the robot to move according to the force exerted by the human and control its stability during the task. Specifically, we focus on the problem of finding a 1D manifold for the robot, while assuming a simple controller to guide its movement on this manifold. To achieve this, we use a functional representation to parameterize the manifold and solve an optimization problem that takes into account the robot's stability and the unpredictability of human behavior. We demonstrate the effectiveness of our approach through simulations and experiments with the Talos robot, showing robustness and adaptability.

Many problems in robotics are fundamentally problems of geometry, which have led to an increased research effort in geometric methods for robotics in recent years. The results were algorithms using the various frameworks of screw theory, Lie algebra, and dual quaternions. A unification and generalization of these popular formalisms can be found in geometric algebra. The aim of this article is to showcase the capabilities of geometric algebra when applied to robot manipulation tasks. In particular, the modeling of cost functions for optimal control can be done uniformly across different geometric primitives leading to a low symbolic complexity of the resulting expressions and a geometric intuitiveness. We demonstrate the usefulness, simplicity, and computational efficiency of geometric algebra in several experiments using a Franka Emika robot. The presented algorithms were implemented in c++20 and resulted in the publicly available librarygafro. The benchmark shows faster computation of the kinematics than state-of-the-art robotics libraries.

Daily manipulation tasks are characterized by geometric primitives related to actions and object shapes. Such geometric descriptors are poorly represented by only using Cartesian coordinate systems. In this paper, we propose a learning approach to extract the optimal representation from a dictionary of coordinate systems to encode an observed movement/behavior. This is achieved by using an extension of Gaussian distributions on Riemannian manifolds, which is used to analyse a set of user demonstrations statistically, by considering multiple geometries as candidate representations of the task. We formulate the reproduction problem as a general optimal control problem based on an iterative linear quadratic regulator (iLQR), where the Gaussian distribution in the extracted coordinate systems are used to define the cost function. We apply our approach to object grasping and box opening tasks in simulation and on a 7-axis Franka Emika robot. The results show that the robot can exploit several geometries to execute the manipulation task and generalize it to new situations, by maintaining the invariant characteristics of the task in the coordinate system(s) of interest.

In Receding Horizon Planning (RHP), it is critical that the motion being executed facilitates the completion of the task, e.g. building momentum to overcome large obstacles. This requires a value function to inform the desirability of robot states. However, given the complex dynamics, value functions are often approximated by expensive computation of trajectories in an extended planning horizon. In this work, to achieve online multi-contact Receding Horizon Planning (RHP), we propose to learn an oracle that can predict local objectives (intermediate goals) for a given task based on the current robot state and the environment. Then, we use these local objectives to construct local value functions to guide a short-horizon RHP. To obtain the oracle, we take a supervised learning approach, and we present an incremental training scheme that can improve the prediction accuracy by adding demonstrations on how to recover from failures. We compare our approach against the baseline (long-horizon RHP) for planning centroidal trajectories of humanoid walking on moderate slopes as well as large slopes where static stability cannot be achieved. We validate these trajectories by tracking them via a whole-body inverse dynamics controller in simulation. We show that our approach can achieve online RHP for 95%-98.6% cycles, outperforming the baseline (8%-51.2%).

In robotics, ergodic control extends the tracking principle by specifying a probability distribution over an area to cover instead of a trajectory to track. The original problem is formulated as a spectral multiscale coverage problem, typically requiring the spatial distribution to be decomposed as Fourier series. This approach does not scale well to control problems requiring exploration in search space of more than two dimensions. To address this issue, we propose the use of tensor trains, a recent low-rank tensor decomposition technique from the field of multilinear algebra. The proposed solution is efficient, both computationally and storagewise, hence making it suitable for its online implementation in robotic systems. The approach is applied to a peg-in-hole insertion task requiring full 6-D end-effector poses, implemented with a seven-axis Franka Emika Panda robot. In this experiment, ergodic exploration allows the task to be achieved without requiring the use of force/torque sensors.

This paper presents the first successful experiment implementing whole-body model predictive control with state feedback on a torque-control humanoid robot. We demonstrate that our control scheme is able to do whole-body target tracking, control the balance in front of strong external perturbations and avoid collision with an external object. The key elements for this success are threefold. First, optimal control over a receding horizon is implemented with Crocoddyl, an optimal control library based on differential dynamics programming, providing state-feedback control in less than 10 ms. Second, a warm start strategy based on memory of motion has been implemented to overcome the sensitivity of the optimal control solver to initial conditions. Finally, the optimal trajectories are executed by a low-level torque controller, feedbacking on direct torque measurement at high frequency. This paper provides the details of the method, along with analytical benchmarks with the real humanoid robot Talos.A video of the experiment is available at https://peertube.laas.fr/videos/watch/cbc25927-337c-4635-a1bc-153b9aeb4135

Collaborative robots offer increased interaction capabilities at relatively low cost but in contrast to their industrial counterparts they inevitably lack precision. Moreover, in addition to the robots' own imperfect models, day-to-day operations entail various sources of errors that despite being small rapidly accumulate. This happens as tasks change and robots are re-programmed, often requiring time-consuming calibrations. These aspects strongly limit the application of collaborative robots in tasks demanding high precision (e.g. watch-making). We address this problem by relying on a dual-arm system with laser-based sensing to measure relative poses between objects of interest and compensate for pose errors coming from robot proprioception. Our approach leverages previous knowledge of object 3D models in combination with point cloud registration to efficiently extract relevant poses and compute corrective trajectories. This results in high-precision assembly behaviors. The approach is validated in a needle threading experiment, with a 150μm thread and a 300μm hole, and a USB insertion task using two 7-axis Panda robots.

Optimal control is often used in robotics for planning a trajectory to achieve some desired behavior, as expressed by the cost function. Most works in optimal control focus on finding a single optimal trajectory, which is then typically tracked by another controller. In this work, we instead consider trajectory distribution as the solution of an optimal control problem, resulting in better tracking performance and a more stable controller. A Gaussian distribution is first obtained from an iterative Linear Quadratic Regulator (iLQR) solver. A short horizon Model Predictive Control (MPC) is then used to track this distribution. We show that tracking the distribution is more cost-efficient and robust as compared to tracking the mean or using iLQR feedback control. The proposed method is validated with kinematic control of 7-DoF Panda manipulator and dynamic control of 6-DoF quadcopter in simulation.

In this paper, we propose a framework to build a memory of motion for warm-starting an optimal control solver for the locomotion task of a humanoid robot. We use HPP Loco3D, a versatile locomotion planner, to generate offline a set of dynamically consistent whole-body trajectory to be stored as the memory of motion. The learning problem is formulated as a regression problem to predict a single-step motion given the desired contact locations, which is used as a building block for producing multi-step motions. The predicted motion is then used as a warm-start for the fast optimal control solver Crocoddyl. We have shown that the approach manages to reduce the required number of iterations to reach the convergence from ~9.5 to only ~3.0 iterations for the single-step motion and from ~6.2 to ~4.5 iterations for the multi-step motion, while maintaining the solution's quality.

This paper addresses the problem of efficiently achieving visual predictive control tasks. To this end, a memory of motion, containing a set of trajectories built off-line, is used for leveraging precomputation and dealing with difficult visual tasks. Standard regression techniques, such as k-nearest neighbors and Gaussian process regression, are used to query the memory and provide on-line a warm-start and a way point to the control optimization process. The proposed technique allows the control scheme to achieve high performance and, at the same time, keep the computational time limited. Simulation and experimental results, carried out with a 7-axis manipulator, show the effectiveness of the approach.

We propose to formulate the problem of representing a distribution of robot configurations (e.g. joint angles) as that of approximating a product of experts. Our approach uses variational inference, a popular method in Bayesian computation, which has several practical advantages over sampling-based techniques. To be able to represent complex and multimodal distributions of configurations, mixture models are used as approximate distribution. We show that the problem of approximating a distribution of robot configurations while satisfying multiple objectives arises in a wide range of problems in robotics, for which the properties of the proposed approach have relevant consequences. Several applications are discussed, including learning objectives from demonstration, planning, and warm-starting inverse kinematics problems. Simulated experiments are presented with a 7-DoF Panda arm and a 28-DoF Talos humanoid.

Learning from demonstration (LfD) is an intuitive framework allowing non-expert users to easily (re-)program robots. However, the quality and quantity of demonstrations have a great influence on the generalization performances of LfD approaches. In this paper, we introduce a novel active learning framework in order to improve the generalization capabilities of control policies. The proposed approach is based on the epistemic uncertainties of Bayesian Gaussian mixture models (BGMMs). We determine the new query point location by optimizing a closed-form information-density cost based on the quadratic Rényi entropy. Furthermore, to better represent uncertain regions and to avoid local optima problem, we propose to approximate the active learning cost with a Gaussian mixture model (GMM). We demonstrate our active learning framework in the context of a reaching task in a cluttered environment with an illustrative toy example and a real experiment with a Panda robot.

Humans exhibit outstanding learning, planning and adaptation capabilities while performing different types of industrial tasks. Given some knowledge about the task requirements, humans are able to plan their limbs motion in anticipation of the execution of specific skills. For example, when an operator needs to drill a hole on a surface, the posture of her limbs varies to guarantee a stable configuration that is compatible with the drilling task specifications, e.g. exerting a force orthogonal to the surface. Therefore, we are interested in analyzing the human arms motion patterns in industrial activities. To do so, we build our analysis on the so-called manipulability ellipsoid, which captures a posture-dependent ability to perform motion and exert forces along different task directions. Through thorough analysis of the human movement manipulability, we found that the ellipsoid shape is task dependent and often provides more information about the human motion than classical manipulability indices. Moreover, we show how manipulability patterns can be transferred to robots by learning a probabilistic model and employing a manipulability tracking controller that acts on the task planning and execution according to predefined control hierarchies.

Most policy search (PS) algorithms require thousands of training episodes to find an effective policy, which is often infeasible with a physical robot. This survey article focuses on the extreme other end of the spectrum: how can a robot adapt with only a handful of trials (a dozen) and a few minutes? By analogy with the word “big-data,” we refer to this challenge as “micro-data reinforcement learning.” In this article, we show that a first strategy is to leverage prior knowledge on the policy structure (e.g., dynamic movement primitives), on the policy parameters (e.g., demonstrations), or on the dynamics (e.g., simulators). A second strategy is to create data-driven surrogate models of the expected reward (e.g., Bayesian optimization) or the dynamical model (e.g., model-based PS), so that the policy optimizer queries the model instead of the real system. Overall, all successful micro-data algorithms combine these two strategies by varying the kind of model and prior knowledge. The current scientific challenges essentially revolve around scaling up to complex robots, designing generic priors, and optimizing the computing time.

In this paper we show how different choices regarding compliance affect a dual-arm assembly task. In addition, we present how the compliance parameters can be learned from a human demonstration. Compliant motions can be used in assembly tasks to mitigate pose errors originating from, for example, inaccurate grasping. We present analytical background and accompanying experimental results on how to choose the center of compliance to enhance the convergence region of an alignment task. Then we present the possible ways of choosing the compliant axes for accomplishing alignment in a scenario where orientation error is present. We show that an earlier presented Learning from Demonstration method can be used to learn motion and compliance parameters of an impedance controller for both manipulators. The learning requires a human demonstration with a single teleoperated manipulator only, easing the execution of demonstration and enabling usage of manipulators at difficult locations as well. Finally, we experimentally verify our claim that having both manipulators compliant in both rotation and translation can accomplish the alignment task with less total joint motions and in shorter time than moving one manipulator only. In addition, we show that the learning method produces the parameters that achieve the best results in our experiments.

Bimanual operations in humanoids offer the possibility to carry out more than one manipulation task at the same time, which in turn introduces the problem of task prioritization. We address this problem from a learning from demonstration perspective, by extending the task-parameterized Gaussian mixture model to Jacobian and null space structures. The proposed approach is tested on bimanual skills but can be applied in any scenario where the prioritization between potentially conflicting tasks needs to be learned. We evaluate the proposed framework in: two different tasks with humanoids requiring the learning of priorities and a loco-manipulation scenario, showing that the approach can be exploited to learn the prioritization of multiple tasks in parallel.

For a safe and successful daily living assistance, far from the highly controlled environment of a factory, robots should be able to adapt to ever-changing situations. Programming such a robot is a tedious process that requires expert knowledge. An alternative is to rely on a high-level planner, but the generic symbolic representations used are not well suited to particular robot executions. Contrarily, motion primitives encode robot motions in a way that can be easily adapted to different situations. This paper presents a combined framework that exploits the advantages of both approaches. The number of required symbolic states is reduced, as motion primitives provide “smart actions” that take the current state and cope online with variations. Symbolic actions can include interactions (e.g., ask and inform) that are difficult to demonstrate. We show that the proposed framework can adapt to the user preferences (in terms of robot speed and robot verbosity), can readjust the trajectories based on the user movements, and can handle unforeseen situations. Experiments are performed in a shoe-dressing scenario. This scenario is particularly interesting because it involves a sufficient number of actions, and the human-robot interaction requires the handling of user preferences and unexpected reactions.

When learning skills from demonstrations, one is often required to think in advance about the appropriate task representation (usually in either operational or configuration space). We here propose a probabilistic approach for simultaneously learning and synthesizing torque control commands which take into account task space, joint space and force constraints. We treat the problem by considering different torque controllers acting on the robot, whose relevance is learned probabilistically from demonstrations. This information is used to combine the controllers by exploiting the properties of Gaussian distributions, generating new torque commands that satisfy the important features of the task. We validate the approach in two experimental scenarios using 7- DoF torque-controlled manipulators, with tasks that require the consideration of different controllers to be properly executed.

Generalizing manipulation skills to new situations requires extracting invariant patterns from demonstrations. For example, the robot needs to understand the demonstrations at a higher level while being invariant to the appearance of the objects, geometric aspects of objects such as its position, size, orientation and viewpoint of the observer in the demonstrations. In this paper, we propose an algorithm that learns a joint probability density function of the demonstrations with invariant formulations of hidden semi-Markov models to extract invariant segments (also called sub-goals or options), and smoothly follow the generated sequence of states with a linear quadratic tracking controller. The algorithm takes as input the demonstrations observed with respect to different coordinate systems describing virtual landmarks or objects of interest, and adapts the segments according to the environmental changes in a systematic manner. We present variants of this algorithm in latent space with low-rank covariance decompositions, semi-tied covariances, and non-parametric online estimation of model parameters under small variance asymptotics; yielding considerably low sample and model complexity for acquiring new manipulation skills. The algorithm allows a Baxter robot to learn a pick-and-place task while avoiding a movable obstacle based on only 4 kinesthetic demonstrations.

We present a general approach for online learning and optimal control of manipulation tasks in a supervisory teleoperation context, targeted to underwater remotely operated vehicles (ROVs). We use an online Bayesian nonparametric learning algorithm to build models of manipulation motions as task-parametrized hidden semi-Markov models (TP-HSMM) that capture the spatiotemporal characteristics of demonstrated motions in a probabilistic representation. Motions are then executed autonomously using an optimal controller, namely a model predictive control (MPC) approach in a receding horizon fashion. This way the remote system locally closes a high-frequency control loop that robustly handles noise and dynamically changing environments. Our system automates common and recurring tasks, allowing the operator to focus only on the tasks that genuinely require human intervention. We demonstrate how our solution can be used for a hot-stabbing motion in an underwater teleoperation scenario. We evaluate the performance of the system over multiple trials and compare with a state-of-the-art approach. We report that our approach generalizes well with only a few demonstrations, accurately performs the learned task and adapts online to dynamically changing task conditions.

We present a trajectory optimization framework for legged locomotion on rough terrain. We jointly optimize the center of mass motion and the foothold locations, while considering terrain conditions. We use a terrain costmap to quantify the desirability of a foothold location. We increase the gait's adaptability to the terrain by optimizing the step phase duration and modulating the trunk attitude, resulting in motions with guaranteed stability. We show that the combination of parametric models, stochastic-based exploration and receding horizon planning allows us to handle the many local minima associated with different terrain conditions and walking patterns. This combination delivers robust motion plans without the need for warm-starting. Moreover, we use soft-constraints to allow for increased flexibility when searching in the cost landscape of our problem. We showcase the performance of our trajectory optimization framework on multiple terrain conditions and validate our method in realistic simulation scenarios and experimental trials on a hydraulic, torque controlled quadruped robot.

In many sensing and control applications, data are represented in the form of symmetric positive definite (SPD) matrices. Considering the underlying geometry of this data space can be beneficial in many robotics applications. In this paper, we present an extension of Gaussian mixture regression (GMR) with input and/or output data on SPD manifolds. As the covariance of SPD datapoints is a 4th-order tensor, we develop a method for parallel transport of high order covariances on SPD manifolds. The proposed approach is experimented in the context of prosthetic hands, with the estimation of wrist movements based on spatial covariance features computed from surface electromyography (sEMG) signals.

Posture body variation is one of the ways in which humans skillfully and naturally augment their motion and strength capabilities along specific task-space directions in order to successfully perform complex manipulation tasks. Posture variation also has a significant role in robot manipulation, where manipulability arises as a useful criterion to analyze and control the robot dexterity as a function of its joint configuration. In this context, this paper introduces the promising idea of manipulability transfer, a method that allows robots to learn and reproduce desired manipulability ellipsoids from expert demonstrations. The proposed framework is built on a tensor-based formulation of Gaussian mixture model that takes into account that manipulability ellipsoids lie on the manifold of symmetric positive definite matrices. This geometry-aware method is used to design a manipulability-based redundancy resolution that allows the robot to modify its posture so that its manipulability ellipsoid coincides with the desired one. Experiments in simulation validate the functionality of the proposed approach, which extends the robot learning capability beyond trajectory, force and impedance learning approaches.

Performing remote manipulation tasks by tele-operation with limited bandwidth, communication delays and environmental differences is a challenging problem. In this paper, we learn a task-parameterized generative model from the teleoperator demonstrations using a hidden semi-Markov model that provides assistance in performing remote manipulation tasks. We present a probabilistic formulation to capture the intention of the teleoperator, and subsequently assist the teleoperator by time-independent shared control and/or time-dependent autonomous control formulations of the model. In the shared control mode, the model corrects the remote arm movement based on the current state of the teleoperator; whereas in the autonomous control mode, the model generates the movement of the remote arm for autonomous task execution. We show the formulation of the model with virtual fixtures and provide comparisons to benchmark our approach. Teleoperation experiments with the Baxter robot for reaching a movable target and opening a valve reveal that the proposed methodology improves the performance of the teleoperator and caters for environmental differences in performing remote manipulation tasks.

In the context of robotic control, synergies can form elementary units of behavior. By specifying task-dependent coordination behaviors at a low control level, one can achieve task-specific disturbance rejection. In this work we present an approach to learn the parameters of such low-level controllers by demonstration. We identify a synergy by extracting covariance information from demonstration data. The extracted synergy is used to derive a time-invariant state feedback controller through optimal control. To cope with the non-Euclidean nature of robot poses, we utilize Riemannian geometry, where both estimation of the covariance and the associated controller take into account the geometry of the pose manifold. We demonstrate the efficacy of the approach experimentally in a bimanual manipulation task.

Programming by Demonstration (PbD) offers a user-friendly way to transfer skills from human to robot. Typically, demonstration data do not contain the control inputs required to reproduce the demonstrated skill. These can be obtained from a low-level controller that tracks the modeled movement. We present a PbD approach for minimal intervention control - a control strategy that only corrects perturbations that interfere with task performance. The novelty of our approach is the probabilistic encoding of the movement duration, providing a performance measure that enables minimal intervention control in a temporal sense. This is achieved by combining a probabilistic movement encoding based on Hidden Semi-Markov Model (HSMM) with Model Predictive Control (MPC). The probabilistic model is used to construct an objective function, hereby assuming that variance is a measure for task performance. The proposed method is demonstrated in a robot experiment and compared with our earlier work.

We present an approach to generate rapid and fluid drawing movements on a compliant Baxter robot, by taking advantage of the kinematic redundancy and torque control capabilities of the robot. We concentrate on the task of reproducing graffiti-stylised letter-forms with a marker. For this purpose, we exploit a compact lognormal-stroke based representation of movement to generate natural drawing trajectories. An Expectation-Maximisation (EM) algorithm is used to iteratively improve tracking performance with low gain feedback control. The resulting system captures the aesthetic and dynamic features of the style under investigation and permits its reproduction with a compliant controller that is safe for users surrounding the robot.

We present a framework for online coordinated obstacle avoidance with formal safety guarantees. Such a formally verified trajectory planner can be used in shared human-robot workspaces to guarantee safety. The obstacle avoidance is based on estimation of the human occupancy on two different time scales. A long-term plan is created based on a probabilistic task representation, learned by demonstration, and an estimate of the human occupancy to be avoided. Using an additional overapproximative, short-term prediction of human motion we guarantee that the robot can always account for sudden or reflex movements. We demonstrate our two-level obstacle avoidance in simulation. The results show that our method reduces the number of safety stops one would encounter when using only the formal safety verification, and synthesizes alternative movement plans that preserves the coordination observed in the original demonstrations.

Robots are becoming safe and smart enough to work alongside people not only on manufacturing production lines, but also in spaces such as houses, museums, or hospitals. This can be significantly exploited in situations in which a human needs the help of another person to perform a task, because a robot may take the role of the helper. In this sense, a human and the robotic assistant may cooperatively carry out a variety of tasks, therefore requiring the robot to communicate with the person, understand his/her needs, and behave accordingly. To achieve this, we propose a framework for a user to teach a robot collaborative skills from demonstrations. We mainly focus on tasks involving physical contact with the user, in which not only position, but also force sensing and compliance become highly relevant. Specifically, we present an approach that combines probabilistic learning, dynamical systems, and stiffness estimation to encode the robot behavior along the task. Our method allows a robot to learn not only trajectory following skills, but also impedance behaviors. To show the functionality and flexibility of our approach, two different testbeds are used: a transportation task and a collaborative table assembly.

Human-robot collaboration seeks to have humans and robots closely interacting in everyday situations. For some tasks, physical contact between the user and the robot may occur, originating significant challenges at safety, cognition, perception and control levels, among others. This paper focuses on robot motion adaptation to parameters of a collaborative task, extraction of the desired robot behavior, and variable impedance control for human-safe interaction. We propose to teach a robot cooperative behaviors from demonstrations, which are probabilistically encoded by a task-parametrized formulation of a Gaussian mixture model. Such encoding is later used for specifying both the desired state of the robot, and an optimal feedback control law that exploits the variability in position, velocity and force spaces observed during the demonstrations. The whole framework allows the robot to modify its movements as a function of parameters of the task, while showing different impedance behaviors. Tests were successfully carried out in a scenario where a 7 DOF backdrivable manipulator learns to cooperate with a human to transport an object.

Very often, when addressing the problem of human-robot skill transfer in task space, only the Cartesian position of the end-effector is encoded by the learning algorithms, instead of the full pose. However, orientation is just as important as position, if not more, when it comes to successfully performing a manipulation task. In this paper, we present a framework that allows robots to learn the full poses of their end-effectors in a task-parameterized manner. Our approach permits the encoding of complex skills, such as those found in bimanual manipulation scenarios, where the generalized coordination patterns between end-effectors (i.e. position and orientation patterns) need to be considered. The proposed framework combines a dynamical systems formulation of the demonstrated trajectories, both in ℝ3and SO(3), and task-parameterized probabilistic models that build local task representations in both spaces, based on which it is possible to extract the relevant features of the demonstrated skill. We validate our approach with an experiment in which two 7-DoF WAM robots learn to perform a bimanual sweeping task.

Robot learning from demonstrations requires the robot to learn and adapt movements to new situations, often characterized by position and orientation of objects or landmarks in the robot's environment. In the task-parameterized Gaussian mixture model framework, the movements are considered to be modulated with respect to a set of candidate frames of reference (coordinate systems) attached to a set of objects in the robot workspace. Following a similar approach, this paper addresses the problem of having missing candidate frames during the demonstrations and reproductions, which can happen in various situations such as visual occlusion, sensor unavailability, or tasks with a variable number of descriptive features. We study this problem with a dust sweeping task in which the robot requires to consider a variable amount of dust areas to clean for each reproduction trial.

A new challenge for surgical robotics is placed in the use of flexible manipulators, to perform procedures that are impossible for currently available rigid robots. Since the surgeon only controls the end-effector of the manipulator, new control strategies need to be developed to correctly move its flexible body without damaging the surrounding environment. This paper shows how a positional controller for a new surgical robot (STIFF-FLOP) can be learnt from the demonstrations given by an expert user. The proposed algorithm exploits the variability of the task to comply with the constraints only when needed, by implementing a minimal intervention principle control strategy. The results are applied to scenarios involving movements inside a constrained environment and disturbance rejection.

We present a task-parameterized probabilistic model encoding movements in the form of virtual spring-damper systems acting in multiple frames of reference. Each candidate coordinate system observes a set of demonstrations from its own perspective, by extracting an attractor path whose variations depend on the relevance of the frame at each step of the task. This information is exploited to generate new attractor paths in new situations (new position and orientation of the frames), with the predicted covariances used to estimate the varying stiffness and damping of the spring-damper systems, resulting in a minimal intervention control strategy. The approach is tested with a 7-DOFs Barrett WAM manipulator whose movement and impedance behavior need to be modulated in regard to the position and orientation of two external objects varying during demonstration and reproduction.

Physical interaction between humans and robots arises a large set of challenging problems involving hardware, safety, control and cognitive aspects, among others. In this context, the cooperative (two or more people/robots) transportation of bulky loads in manufacturing plants is a practical example where these challenges are evident. In this paper, we address the problem of teaching a robot collaborative behaviors from human demonstrations. Specifically, we present an approach that combines: probabilistic learning and dynamical systems, to encode the robot's motion along the task. Our method allows us to learn not only a desired path to take the object through, but also, the force the robot needs to apply to the load during the interaction. Moreover, the robot is able to learn and reproduce the task with varying initial and final locations of the object. The proposed approach can be used in scenarios where not only the path to be followed by the transported object matters, but also the force applied to it. Tests were successfully carried out in a scenario where a 7 DOFs backdrivable manipulator learns to cooperate, with a human, to transport an object while satisfying the position and force constraints of the task.

Skills can often be performed in many different ways. In order to provide robots with human-like adaptation capabilities, it is of great interest to learn several ways of achieving the same skills in parallel, since eventual changes in the environment or in the robot can make some solutions unfeasible. In this case, the knowledge of multiple solutions can avoid relearning the task. This problem is addressed in this paper within the framework of Reinforcement Learning, as the automatic determination of multiple optimal parameterized policies. For this purpose, a model handling a variable number of policies is built using a Bayesian non-parametric approach. The algorithm is first compared to single policy algorithms on known benchmarks. It is then applied to a typical robotic problem presenting multiple solutions.

Research in learning from demonstration has focused on transferring movements from humans to robots. However, a need is arising for robots that do not just replicate the task on their own, but that also interact with humans in a safe and natural way to accomplish tasks cooperatively. Robots with variable impedance capabilities opens the door to new challenging applications, where the learning algorithms must be extended by encapsulating force and vision information. In this paper we propose a framework to transfer impedance-based behaviors to a torque-controlled robot by kinesthetic teaching. The proposed model encodes the examples as a task-parameterized statistical dynamical system, where the robot impedance is shaped by estimating virtual stiffness matrices from the set of demonstrations. A collaborative assembly task is used as testbed. The results show that the model can be used to modify the robot impedance along task execution to facilitate the collaboration, by triggering stiff and compliant behaviors in an on-line manner to adapt to the user's actions.

Gestures are characterized by intermediary or final landmarks (real or virtual) in task space or joint space that can change during the course of the motion, and that are described by varying accuracy and correlation constraints. Generalizing these trajectories in robot learning by imitation is challenging, because of the small number of demonstrations provided by the user. We present an approach to statistically encode movements in a task-parameterized mixture model, and derive an expectation-maximization (EM) algorithm to train it. The model automatically extracts the relevance of candidate coordinate systems during the task, and exploits this information during reproduction to adapt the movement in real-time to changing position and orientation of landmarks or objects. The approach is tested with a robotic arm learning to roll out a pizza dough. It is compared to three categories of task-parameterized models: 1) Gaussian process regression (GPR) with a trajectory models database; 2) Multi-streams approach with models trained in several frames of reference; and 3) Parametric Gaussian mixture model (PGMM) modulating the Gaussian centers with the task parameters. We show that the extrapolation capability of the proposed approach outperforms existing methods, by extracting the local structures of the task instead of relying on interpolation principles.

Robot programming by demonstration encompasses a wide range of learning strategies, from simple mimicking of the demonstrator's actions to the higher level extraction of the underlying intent. By focusing on this last form, we study the problem of extracting the reward function explaining the demonstrations from a set of candidate reward functions, and using this information for self-refinement of the skill. This definition of the problem has links with inverse reinforcement learning problems in which the robot autonomously extracts an optimal reward function that defines the goal of the task. By relying on Gaussian mixture models, the proposed approach learns how the different candidate reward functions are combined, and in which contexts or phases of the task they are relevant for explaining the user's demonstrations. The extracted reward profile is then exploited to improve the skill with a self-refinement approach based on expectation-maximization, allowing the imitator to reach a skill level that goes beyond the demonstrations. The approach can be used to reproduce a skill in different ways or to transfer tasks across robots of different structures. The proposed approach is tested in simulation with a new type of continuum robot (STIFF-FLOP), using kinesthetic demonstrations from a Barrett WAM manipulator.

A summary of the state-of-the-art reinforcement learning in robotics is given, in terms of both algorithms and policy representations. Numerous challenges faced by the policy representation in robotics are identified. Two recent examples for application of reinforcement learning to robots are described: pancake flipping task and bipedal walking energy minimization task. In both examples, a state-of-the-art Expectation-Maximization-based reinforcement learning algorithm is used, but different policy representations are proposed and evaluated for each task. The two proposed policy representations offer viable solutions to four rarely-addressed challenges in policy representations: correlations, adaptability, multi-resolution, and globality. Both the successes and the practical difficulties encountered in these examples are discussed.

We present an integrated approach allowing a free-standing humanoid robot to acquire new motor skills by kinesthetic teaching. The proposed method controls simultaneously the upper and lower body of the robot with different control strategies. Imitation learning is used for training the upper body of the humanoid robot via kinesthetic teaching, while at the same time Reaction Null Space method is used for keeping the balance of the robot. During demonstration, a force/torque sensor is used to record the exerted forces, and during reproduction, we use a hybrid position/force controller to apply the learned trajectories in terms of positions and forces to the end effector. The proposed method is tested on a 25-DOF Fujitsu HOAP-2 humanoid robot with a surface cleaning task.

We study the use of different weighting mechanisms in robot learning to represent a movement as a combination of linear systems. Kinesthetic teaching is used to acquire a skill from demonstrations which is then reproduced by the robot. The behaviors of the systems are analyzed when the robot faces perturbation introduced by the user physically interacting with the robot to momentarily stop the task. We propose the use of a Hidden Semi-Markov Model (HSMM) representation to encapsulate duration and position information in a robust manner with parameterization on the involvement of time and space constraints. The approach is tested in simulation and in two robot experiments, where a 7 DOFs manipulator is taught to play a melody by pressing three big keys and to pull a model train on its track.

We present a learning-based approach for minimizing the electric energy consumption during walking of a passively-compliant bipedal robot. The energy consumption is reduced by learning a varying-height center-of-mass trajectory which uses efficiently the robot's passive compliance. To do this, we propose a reinforcement learning method which evolves the policy parameterization dynamically during the learning process and thus manages to find better policies faster than by using fixed parameterization. The method is first tested on a function approximation task, and then applied to the humanoid robot COMAN where it achieves significant energy reduction.

When collaboration between human users and robots involves physical interaction, the importance of the safety issue arises. We propose a method to transfer to robots several tasks demonstrated by the user through kinesthetic teaching and subsequently learned using a weighted combination of dynamical systems (DS). The approach used to encode the desired skills ensures a safe robot behavior during the task reproduction, allowing physical interaction with the user who can employ the manipulator as a tangible interface. By using a force sensor-less impedance controller with a back-drivable robot, this concept is exploited in two physical human-robot interaction (pHRI) scenarios. The first considers an emergency situation in which the user can stop or pause a task execution by grasping and moving the robot away from the region of space associated to the skill. The second studies the possibility to select one among several learned tasks and switch to its execution by physically guiding the robot towards the task region.

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 propose a control strategy for a robotic manipulator operating in an unstructured environment while interacting with a human operator. The proposed system takes into account the important characteristics of the task and the redundancy of the robot to determine a controller that is safe for the user. The constraints of the task are first extracted using several examples of the skill demonstrated to the robot through kinesthetic teaching. An active control strategy based on task-space control with variable stiffness is proposed, and combined with a safety strategy for tasks requiring humans to move in the vicinity of robots. A risk indicator for human-robot collision is defined, which modulates a repulsive force distorting the spatial and temporal characteristics of the movement according to the task constraints. We illustrate the approach with two human-robot interaction experiments, where the user teaches the robot first how to move a tray, and then shows it how to iron a napkin.

We present an approach allowing a robot to acquire new motor skills by learning the couplings across motor control variables. The demonstrated skill is first encoded in a compact form through a modified version of Dynamic Movement Primitives (DMP) which encapsulates correlation information. Expectation-Maximization based Reinforcement Learning is then used to modulate the mixture of dynamical systems initialized from the user's demonstration. The approach is evaluated on a torque-controlled 7 DOFs Barrett WAM robotic arm. Two skill learning experiments are conducted: a reaching task where the robot needs to adapt the learned movement to avoid an obstacle, and a dynamic pancake-flipping task.

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 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.

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.

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).

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 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

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.

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.