Search

Robots require the ability to autonomously and continuously react to unexpected online changes in the task definition and in the environment, especially those cohabited with humans. To react to these changes, the task, from the current state up to the finish, must instantly be reconsidered. This implies a prohibitive re-computation cost. This paper proposes a modular control architecture based on Model Predictive Control, that offers a good compromise between optimally achieving the task and the required computation time, by only reconsidering the near future. This framework offers a generic way to formulate task-related objectives and constraints that dissociates the planning from the execution, which depends mainly on the robot dynamics. The proposal exploits a linear formalization of the MPC in SE(3) to implement this architecture in a high-frequency closed-loop controller, achieving task re-planning at the control rate. The pertinence of the proposed control architecture is demonstrated using experiments with the Franka Emika robots in scenarios where the task to be achieved is modified on the fly.

International audience; The concept of signaling motions of a robot interacting with a human was presented in a conference paper. These motions consist in using the redundant degrees of freedom of a robot performing a task as new means of meaningful communication between the robot and the human. They are generated by a quasi-static torque control, consistent with the main task of the robot. A double within-subjects study (N=16) was conducted to evaluate the effects of two signaling motions (one diffuse - breathing - and one explicit - postural encoding) on the participants' performance of a simulated wood-turning task and on their behavior towards the robot. Our results show a positive effect on both task performance and participants' behavior. Furthermore, both signaling movements appear to improve participants' situation awareness by feeding their mental model throughout the interaction. Based on a systematic evaluation of potential situational awareness demons and this work, our long-term goal is to explore a general way to overlay these signaling motions onto robot movements in an automatic, transparent, and appropriate manner. One of the challenges in achieving this goal is related to understanding the relationship between the context as well as the specific characteristics of the user and the perception of the signaling motions.; Le concept des mouvements de signalisation d'un robot interagissant avec un humain a été présenté dans un article de conférence. Ces mouvements consistent à utiliser les degrés de liberté redondants d'un robot exécutant une tâche comme nouveaux moyens de communication significative entre le robot et l'homme. Ils sont générés par un contrôle de couple quasi-statique, en cohérence avec la tâche principale du robot. Une double étude intra-sujet (N=16) a été menée pour évaluer les effets de deux mouvements de signalisation (un diffus - la respiration - et un explicite - l'encodage postural) sur l'exécution d'une tâche de simulation de toupillage de bois par les participants et sur leur comportement envers le robot. Nos résultats montrent un effet positif à la fois sur l'exécution de la tâche et sur le comportement des participants. De plus, les deux mouvements de signalisation semblent améliorer la conscience de la situation des participants en alimentant leur modèle mental tout au long de l'interaction. Sur la base d'une évaluation systématique des démons potentiels de la conscience de situation et de ce travail, notre objectif à long terme est d'explorer une manière générale de superposer ces mouvements de signalisation aux mouvements du robot d'une manière automatique, transparente et appropriée. Un des défis pour atteindre cet objectif est lié à la compréhension de la relation entre le contexte ainsi que les caractéristiques spécifiques de l'utilisateur et la perception des mouvements de signalisation.

Planning a robot's task-space movement according to its actual capacity is a difficult problem because this capacity depends on its state and thus can evolve significantly during the execution of the movement. This paper proposes a method for real-time trajectory planning based on time-optimal Trapezoidal acceleration profile (TAP) trajectories, that adapts to the real-time evolution of the robot's capacity. The method is based on an efficient approach for projecting the robot's kinematic limits in the trajectory direction, based on the convex polytope algebra. The method is experimentally validated on a Franka Emika Panda collaborative robot and compared with the classical approach considering fixed robot's Cartesian space motion capacity. The results show that the proposed method is able to better exploit true robot's motion capacity by generating faster trajectories for the same level of tracking accuracy.

Efficient workspace sharing of collaborative robots and human operators remains an unsolved problem in the industry. This problem goes beyond the use of a priori or a posteriori safety measures and has to be tackled at the control level. To address the need of adaptation to human presence as well as to endow the robot with the ability to adapt interactively to new Cartesian targets, a linear Model Predictive Controller is proposed in this paper. This controller computes accelerationbounded optimal Cartesian trajectories in SE(3) over a receding horizon. The pertinence of the proposed control architecture is demonstrated using experiments with the Franka Emika robots in different scenarios implying both adaptation of the maximum allowed velocity to comply with human presence and on-the-fly update of a Cartesian goal pose.

Inverse Optimal Control (IOC) is a problem that exists in many fields. In the context of human motion analysis, many methods for solving this problem have been proposed. This paper presents the Projected Inverse Optimal Control (PIOC) approach which puts forward a simple and complete vision of the problem. PIOC is composed of several independent elements allowing to cover problems of dimensioning, measurement noise and reliability of solution. The first element is the parameterization of the data, allowing to reduce the dimensions of the problem. The second element corresponds to the decoupling of the conditions including the conditions of identifiability of the parameters, singularity of the problem and feasibility of the solution. The last element refers to the choice of solution. We show with this approach that the classical solution methods are in fact projections in the space of trajectories. PIOC also allows us to propose a simple algorithm for the choice of basis, a recurrent problem in IOC

International audience; In this paper, the concept of signaling motions of a robot interacting with a human is presented. These motions consist in using the redundant degrees of freedom of a robot performing a task as new means of meaningful robot-human communication. They are generated through quasi-static torque control, in consistency with the main robot task. A double withinsubject (N=16) study is conducted to evaluate the effects of two signaling motions on the performance of a task by participants and on their behavior towards the robot. Our results show a positive effect on both the task execution and the participants behavior. Additionally, both signaling motions seem to improve the situation awareness of the participants by fueling their mental model throughout the interaction.

International audience

International audience; Inverse Optimal Control (IOC) is a popular method for human motion analysis. In the context of these methods it is necessary to pay attention to the reliability of the results. This paper proposes an approach based on the evaluation of Karush-Kuhn-Tucker conditions relying on a complete analysis with Singular Value Decomposition and provides a detailed analysis of reliability. With respect to a ground truth, our simulations illustrate how the proposed method analyzes the reliability of the resolution. After introducing a clear methodology, the properties of the matrices are studied with different noise levels and different experimental models and conditions. We show how to implement the method, step by step, by explaining the numerical difficulties encountered during the resolution and thus how to make the results of the IOC problem reliable.

This article presents a quantification approach of manual expertise. The long term goal is to better understand the notion of manual expertise in order to improve the design of collaborative robots, both from a mechanical and control point of view. Based on the existing literature and through exchanges with highly skilled manual operators, we first propose a definition of manual expertise. Based on this definition, we propose quantitative evaluation criteria for three important dimensions of a manual task completion: safety, discomfort and performance. These criteria are evaluated in experiments relying on a physically realistic mock-up of a wood milling task, a highly demanding task in the carpentry domain. This mock-up includes a 7-DOF collaborative robot used both to reproduce the cutting tool wrenches and provide measurement of the wooden part motion. This experimental setup is used in a training protocol including two groups of 5 novice subjects. This protocol confirms that the proposed approach allows to observe and analyse the handling strategy developed by an operator through training as well as to correlate the type of training to the nature of the developed expertise.