**Êtes-vous un étudiant de l'EPFL à la recherche d'un projet de semestre?**

Travaillez avec nous sur des projets en science des données et en visualisation, et déployez votre projet sous forme d'application sur GraphSearch.

Publication# Learning from demonstration using products of experts: Applications to manipulation and task prioritization

Résumé

Probability distributions are key components of many learning from demonstration (LfD) approaches, with the spaces chosen to represent tasks playing a central role. Although the robot configuration is defined by its joint angles, end-effector poses are often best explained within several task spaces. In many approaches, distributions within relevant task spaces are learned independently and only combined at the control level. This simplification implies several problems that are addressed in this work. We show that the fusion of models in different task spaces can be expressed as products of experts (PoE), where the probabilities of the models are multiplied and renormalized so that it becomes a proper distribution of joint angles. Multiple experiments are presented to show that learning the different models jointly in the PoE framework significantly improves the quality of the final model. The proposed approach particularly stands out when the robot has to learn hierarchical objectives that arise when a task requires the prioritization of several sub-tasks (e.g. in a humanoid robot, keeping balance has a higher priority than reaching for an object). Since training the model jointly usually relies on contrastive divergence, which requires costly approximations that can affect performance, we propose an alternative strategy using variational inference and mixture model approximations. In particular, we show that the proposed approach can be extended to PoE with a nullspace structure (PoENS), where the model is able to recover secondary tasks that are masked by the resolution of tasks of higher-importance.

Official source

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Concepts associés

Chargement

Publications associées

Chargement

Publications associées (60)

Chargement

Chargement

Chargement

Concepts associés (18)

Robot

vignette|Atlas (2013), robot androïde de Boston Dynamics
vignette|Bras manipulateurs dans un laboratoire (2009)
vignette|NAO (2006), robot humanoïde éducatif d'Aldebaran Robotics
vignette|DER1 (2005),

Loi de probabilité

thumb|400px
3 répartitions.png
En théorie des probabilités et en statistique, une loi de probabilité décrit le comportement aléatoire d'un phénomène dépendant du hasard. L'étude des phénomènes aléa

Robotique

thumb|upright=1.5|Nao, un robot humanoïde.
thumb|upright=1.5|Des robots industriels au travail dans une usine.
La robotique est l'ensemble des techniques permettant la conception et la réalisation de

We consider the problem of learning robust models of robot motion through demonstration. An approach based on Hidden Markov Model (HMM) and Gaussian Mixture Regression (GMR) is proposed to extract redundancies across multiple demonstrations, and build a time- independent model of a set of movements demonstrated by a human user. Two experiments are presented to validate the method, that consist of learning to hit a ball with a robotic arm, and of teaching a humanoid robot to manipulate a spoon to feed another humanoid. The experiments demonstrate that the proposed model can efficiently handle several aspects of learning by imitation. We first show that it can be utilized in an unsupervised learning manner, where the robot is autonomously organizing and encoding variants of motion from the multiple demonstrations. We then show that the approach allows to robustly generalize the observed skill by taking into account multiple constraints in task space during reproduction.

2009We present a Programming by Demonstration (PbD) 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 some simple manipulatory tasks. A probability based estimation of the relevance is suggested, by first projecting the joint angles, hand paths, and object-hand trajectories onto a generic latent space using Principal Component Analysis (PCA). The resulting signals were then encoded using a mixture of Gaussian/Bernoulli distributions (GMM/BMM). 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 (GMR). Finally, we analytically compute the trajectory which optimizes the imitation metric and use this to generalize the skill to different contexts and to the robot's specific bodily constraints.

2007Robot 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 focused on systems that learn a specific subset of tasks, our work explores the problem of recognition, generalization, and reproduction of 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.

2004