No description available.

No description available.

Humans learn to interact with many complex dynamic systems such as helicopters, bicycles, and automobiles. This dissertation develops a subsystem identification method to model the control strategies that human subjects use in experiments where they interact with dynamic systems. This work provides new results on the control strategies that humans learn. We present a novel subsystem identification algorithm, which can identify unknown linear time-invariant feedback and feedforward subsystems interconnected with a known linear time-invariant subsystem. These subsystem identification algorithms are analyzed in the cases of noiseless and noisy data. We present results from human-in-the-loop experiments, where human subjects in- teract with a dynamic system multiple times over several days. Each subject’s control behavior is assumed to have feedforward (or anticipatory) and feedback (or reactive) components, and is modeled using experimental data and the new subsystem identifi- cation algorithms. The best-fit models of the subjects’ behavior suggest that humans learn to control dynamic systems by approximating the inverse of the dynamic system in feedforward. This observation supports the internal model hypothesis in neuro- science. We also examine the impact of system zeros on a human’s ability to control a dynamic system, and on the control strategies that humans employ.

This dissertation demonstrates a new method for inferring a representation of the motor command, generated by the central nervous system for interactive point-to-point movements. This new tool, the input inference neural network or IINN, allows estimation of the complexity of the motor command. The IINN was applied to experimental data gathered from 7 volunteer subjects who performed point-to-point tasks while interacting with a specially constructed haptic robot. The motor plan inference demonstrates that, for the point-to-point movement tasks executed during experiments, the motor command can be projected onto a low-dimensional manifold. This dimension is estimated to be 4 or 5 and far less than the degrees of freedom available in the arm. It is hypothesized that subjects simplify the problem of adapting to changing environments by projecting the motor control problem onto a motor manifold of low dimension. Reducing the dimension of the movement optimization problem through the development of a motor manifold can explain rapid adaptation to new motor tasks.

This dissertation demonstrates a new method for inferring a representation of the motor command, generated by the central nervous system for interactive point-to-point movements. This new tool, the input inference neural network or IINN, allows estimation of the complexity of the motor command. The IINN was applied to experimental data gathered from 7 volunteer subjects who performed point-to-point tasks while interacting with a specially constructed haptic robot. The motor plan inference demonstrates that, for the point-to-point movement tasks executed during experiments, the motor command can be projected onto a low-dimensional manifold. This dimension is estimated to be 4 or 5 and far less than the degrees of freedom available in the arm. It is hypothesized that subjects simplify the problem of adapting to changing environments by projecting the motor control problem onto a motor manifold of low dimension. Reducing the dimension of the movement optimization problem through the development of a motor manifold can explain rapid adaptation to new motor tasks.

L'émergence Troubles Musculo-Squelettiques (TMS) en industrie constitue un véritable fléau ayant de lourdes conséquences socio-économique en France. Afin de réduire la pénibilité au travail et les risques TMS, les industriels s'engagent dans une politique de réaménagement des postes de travail par la mise en œuvre de moyens robotisés d'assistance aux opérateurs. Dans cette politique de prévention, le groupe PSA Peugeot Citroën aspire à utiliser des cobots et des exosquelettes comme dispositifs d'assistance pour améliorer les conditions de travail des opérateurs. Mais pour mettre en œuvre ces types de robot en usine, il est nécessaire de quantifier leurs apports ergonomiques. C'est dans ce contexte que s'inscrit cette thèse, dont l'objectif est de proposer une méthode d'évaluation de robots collaboratifs visant à être mis en œuvre dans les usines PSA Peugeot Citroën. Dans le cadre de ces travaux, nous avons utilisé l'exosquelette mono-bras droit ABLE, conçu par le CEA-LIST. A partir d'une analyse biomécanique d'une tâche de manipulation humaine, nous avons pu évaluer l'apport de l'exosquelette en termes de réduction de charge physique de l'utilisateur. Aussi avons-nous proposé dans ces travaux d'analyser les mécanismes neuromusculaires résultants du mouvement effectué en interaction avec l'exosquelette. Sur la base de la théorie du contrôle moteur humain et en utilisant une méthode d'optimisation inverse, les fonctions objectifs telles que jerk, le couple articulaire, ou l'énergie, caractérisant la tâche de manipulation humaine en termes d'efforts, de cinématique et de temps d'exécution, ont été identifiées. Cette meilleure compréhension du mouvement du membre supérieur humain a permis ensuite de revenir sur la conception de l'exosquelette afin de proposer une stratégie de commande optimisée à l'exécution de tâches de travail en environnement industriel. / The emergence of Musculo-Squelettal Disorders (MSD) in the industry is a real blight, having major socioeconomic consequences in France. In order to reduce work painfulness and MSD risks, some industries are committing to modifying workstations by assisting operators with robotic devices. Following this MSD prevention policy, PSA Peugeot Citroen aims to use cobots or exoskeletons as assistive devices to improve workers conditions. However, implementing this type of robot in factories requires quantifying their ergonomic benefit. In this context, the objective of this thesis is to develop a method to assess collaborative robot that are intended to be used in PSA Peugeot Citroen factories. In this framework, the right mono-arm ABLE exoskeleton, designed by the CEA-LIST has been used. With a biomechanical analysis of an industrial manipulation task, we have been able to assess the benefit of the exoskeleton in terms of physical load reduction. We also proposed in this work to assess neuromuscular mechanisms underlying the industrial task performed in interaction with the exoskeleton. On the basis of the human motor control theory and using an inverse optimisation method, objectives functions such as jerk, joint torque or energy that characterize the human manipulation task in terms of efforts, kinematics and execution time, have been identified. This improved understanding of human upper limb movements then allowed reviewing the exoskeleton design in order to propose an optimal command strategy adapted to the execution of industrial tasks.

No description available.

A novel controller termed as Experience Mapping based Prediction Controller (EMPC) is developed in this work. EMPC is developed utilizing the broad control concepts of human motor control (HMC). The concepts of HMC are utilized to develop the core concepts of EMPC for the control of ideal Type-1 LTI systems. The control accuracy of the developed concepts is studied and the mathematical stability criterion for the controller is developed. The applicability of EMPC for the control of real world problems is tested on a Permanent Magnet DC motor based position control system. 1. Novel learning methods are presented to form experience mapped knowledge-base (EMK) which is used for the creation of the forward and inverse models. 2. Control and Adaptation Techniques which overcome the presence of non idealities are developed using the inverse model. 3. Two separate techniques which utilize the forward model for improving the adaptation capabilities of EMPC are developed. 4. Two novel techniques are developed for the improvement of the tracking performance in terms of the accuracy and smoothness of tracking. These techniques are tested under various system conditions including large dynamic parameter changes on a simulation model and a practical setup. The performance of EMPC is compared against that of PID, MRAC and LQG controllers for all the proposed techniques and EMPC is found to perform significantly better under the various system conditions in terms of transient and steady state characteristics. Finally, the effectiveness of EMPC in stabilizing unstable systems using the concepts developed is tested on a practical Inverted Pendulum system. The problem of the simultaneous development of experiences and control of the system is addressed with the stabilizing problem. The proposed controller, EMPC provides an alternative approach for the existing control of systems without the requirement of an accurate system mathematical model. Its capability to learn by directly interacting with the system and adapt using experiences makes it an attractive alternative to other control techniques present in literature. Keywords: EMPC, Position Control, PMDC motors

abstract: Daily collaborative tasks like pushing a table or a couch require haptic communication between the people doing the task. To design collaborative motion planning algorithms for such applications, it is important to understand human behavior. Collaborative tasks involve continuous adaptations and intent recognition between the people involved in the task. This thesis explores the coordination between the human-partners through a virtual setup involving continuous visual feedback. The interaction and coordination are modeled as a two-step process: 1) Collecting data for a collaborative couch-pushing task, where both the people doing the task have complete information about the goal but are unaware of each other's cost functions or intentions and 2) processing the emergent behavior from complete information and fitting a model for this behavior to validate a mathematical model of agent-behavior in multi-agent collaborative tasks. The baseline model is updated using different approaches to resemble the trajectories generated by these models to human trajectories. All these models are compared to each other. The action profiles of both the agents and the position and velocity of the manipulated object during a goal-oriented task is recorded and used as expert-demonstrations to fit models resembling human behaviors. Analysis through hypothesis teasing is also performed to identify the difference in behaviors when there are complete information and information asymmetry among agents regarding the goal position. / Dissertation/Thesis / Masters Thesis Electrical Engineering 2020

La recherche portant sur l'identification des principes neurobiologiques qui sous-tendent le contrôle moteur humain est actuellement très active. Les mouvements humains ont en effet un niveau de robustesse et de dextérité encore inégalé dans la réalisation robotique de tâches complexes. L'objectif est donc de mieux comprendre l'origine de cette performance et de la reproduire en robotique bio-inspirée. Il a déjà été démontré que des réseaux spinaux rythmiques sont présents dans la moelle épinière des vertébrés. Ils constituent des systèmes dynamiques non-linéaires composés de neurones en inhibition réciproque et seraient à l’origine de la génération des mouvements rythmiques comme la locomotion et la respiration. Les attracteurs de ces systèmes dynamiques seraient modulés de manière continue ou intermittente par des signaux sensoriels et des signaux descendant du cortex moteur, de manière à adapter le comportement de l’agent à la dynamique de l’environnement.La présente étude émet l'hypothèse que des informations visuelles sont également couplées aux réseaux spinaux rythmiques et que ces couplages sont responsables des synchronisations temporelles et spatiales observées lors de la réalisation de tâches visuomotrices rythmiques. Cette proposition est confrontée à des résultats expérimentaux de frappe cyclique de balle, un benchmark bien connu des neuroscientifiques et des dynamiciens en raison de ses propriétés dynamiques intrinsèques. Il rend possible à la fois l’étude de la génération de mouvements rythmiques par des réseaux spinaux, la synchronisation temporelle avec l’environnement, la correction en-ligne des erreurs spatiales et l’interception de projectiles balistiques.Cette thèse propose ainsi un modèle comportemental mathématique innovant reposant sur un modèle d’oscillateur neuronal dont l’attracteur, qui définit les trajectoires de la raquette, est modulé en ligne par les perceptions visuelles de la trajectoire de la balle. La pertinence du modèle est validée par comparaison aux données expérimentales et aux modèles précédemment proposés dans la littérature. La robustesse de cette stratégie de contrôle est également quantifiée par une analyse de stabilité asymptotique du système hybride défini par le couplage entre le système neuro-musculo-squelettique et la balle. Le correcteur bio-inspiré proposé dans cette thèse réunit de manière harmonieuse un contrôle prospectif de la synchronisation balle-raquette, un contrôle paramétrique intermittent dimensionnant le mouvement et un contrôle émergeant du cycle-limite du système couplé. Il reproduit efficacement les modulations des actions motrices et les performances des humains durant la tâche de frappe cyclique de balle, y compris en présence de perturbations, et ce sans avoir recours à une planification du mouvement ou à des représentations internes explicites de l’environnement. Les résultats de cette étude conduisent à l’affirmation réaliste que les mouvements humains sont directement structurés par l’information sensorielle disponible et par des stratégies correctives en-ligne, en accord avec la théorie des dynamiques comportementales. Cette architecture de contrôle pourrait offrir de nombreux avantages aux robots humanoïdes qui en seraient munis, en assurant stabilité et économie d’énergie, par l’intermédiaire de lois de commande de faible complexité et peu gourmandes en ressources computationnelles. / The identification of the neurbiological principles underlying human motor control is a very active reseach topic. Indeed, human movement has a level of robustness and dexterity still unmatched by robots. The objective is therefore to better understand the origin of this efficiency to replicate these performances in robotics. It has been shown that spinal rhythm generators, known as Central Pattern Generators (CPG), are responsible for the generation of rhythmic movements such as locomotion and respiration in vertebrates. These CPG constitute dynamic nonlinear systems modulated by sensory signals and descending signals from the cortex to adapt the behavior to the changing environment.The present study hypothesizes that visual information is also coupled to the CPG and that these couplings are responsible for the temporal and spatial synchronization observed during rhythmic visuomotor tasks. This assumption is confronted with experimental results from human participants performing ball bouncing, a well-known benchmark in neuroscience and robotics for its intrinsic dynamic properties. This task allows for the investigation of rhythmic movement generation by spinal networks, the temporal synchronization with the environment, the on-line correction of spatial errors and the interception of ballistic projectiles.This thesis proposes an innovative mathematical behavioral model based on a neuronal oscillator whose attractor, which defines the paddle trajectories, is modulated on-line by the visual perception of the ball trajectory. The relevance of the model is validated by comparison with experimental data and models previously proposed in the literature. The robustness of this control strategy is quantified by an asymptotic stability analysis. The bio-inspired controller presented in this thesis harmoniously combines a prospective control of the ball-paddle synchronization, an intermittent parametric control that scales the movement and a control emerging from the coupled system limit cycle. It efficiently reproduces the human modulation in motor action and performance during ball bouncing, without relying on movement planning or explicit internal representation of the environment. The results of this study lead to the realistic assumption that much part of the human behavior during ball bouncing is directly structured by sensory information and on-line error correction processes, in agreement with the behavioral dynamics theory. This control architecture holds promise for the control of humanoid robots as it is able to ensure stability and energy saving through control laws of reduced complexity and computational cost.