Learning Stable Task Sequences from Demonstration with Linear Parameter Varying Systems and Hidden Markov Models

Jose R. Medina, Aude Billard
Proceedings of the 1st Annual Conference on Robot Learning, PMLR 78:175-184, 2017.

Abstract

The problem of acquiring multiple tasks from demonstration is typically divided in two sequential processes: (1) the segmentation or identification of different subgoals/subtasks and (2) a separate learning process that parameterizes a control policy for each subtask. As a result, segmentation criteria typically neglect the characteristics of control policies and rely instead on simplified models. This paper aims for a single model capable of learning sequences of complex time-independent control policies that provide robust and stable behavior. To this end, we first present a novel and efficient approach to learn goal-oriented time-independent motion models by estimating \emphboth attractor and dynamic behavior from data guaranteeing stability using linear parameter varying (LPV) systems. This method enables learning complex task sequences with hidden Markov models (HMMs), where each state/subtask is given by a stable LPV system and where transitions are most likely around the corresponding attractor. We study the dynamics of the HMM-LPV model and propose a motion generation method that guarantees the stability of task sequences. We validate our approach in two sets of demonstrated human motions.

Cite this Paper

BibTeX
@InProceedings{pmlr-v78-medina17a, title = {Learning Stable Task Sequences from Demonstration with Linear Parameter Varying Systems and Hidden Markov Models}, author = {Medina, Jose R. and Billard, Aude}, booktitle = {Proceedings of the 1st Annual Conference on Robot Learning}, pages = {175--184}, year = {2017}, editor = {Levine, Sergey and Vanhoucke, Vincent and Goldberg, Ken}, volume = {78}, series = {Proceedings of Machine Learning Research}, month = {13--15 Nov}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v78/medina17a/medina17a.pdf}, url = {https://proceedings.mlr.press/v78/medina17a.html}, abstract = {The problem of acquiring multiple tasks from demonstration is typically divided in two sequential processes: (1) the segmentation or identification of different subgoals/subtasks and (2) a separate learning process that parameterizes a control policy for each subtask. As a result, segmentation criteria typically neglect the characteristics of control policies and rely instead on simplified models. This paper aims for a single model capable of learning sequences of complex time-independent control policies that provide robust and stable behavior. To this end, we first present a novel and efficient approach to learn goal-oriented time-independent motion models by estimating \emphboth attractor and dynamic behavior from data guaranteeing stability using linear parameter varying (LPV) systems. This method enables learning complex task sequences with hidden Markov models (HMMs), where each state/subtask is given by a stable LPV system and where transitions are most likely around the corresponding attractor. We study the dynamics of the HMM-LPV model and propose a motion generation method that guarantees the stability of task sequences. We validate our approach in two sets of demonstrated human motions.} }
Endnote
%0 Conference Paper %T Learning Stable Task Sequences from Demonstration with Linear Parameter Varying Systems and Hidden Markov Models %A Jose R. Medina %A Aude Billard %B Proceedings of the 1st Annual Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2017 %E Sergey Levine %E Vincent Vanhoucke %E Ken Goldberg %F pmlr-v78-medina17a %I PMLR %P 175--184 %U https://proceedings.mlr.press/v78/medina17a.html %V 78 %X The problem of acquiring multiple tasks from demonstration is typically divided in two sequential processes: (1) the segmentation or identification of different subgoals/subtasks and (2) a separate learning process that parameterizes a control policy for each subtask. As a result, segmentation criteria typically neglect the characteristics of control policies and rely instead on simplified models. This paper aims for a single model capable of learning sequences of complex time-independent control policies that provide robust and stable behavior. To this end, we first present a novel and efficient approach to learn goal-oriented time-independent motion models by estimating \emphboth attractor and dynamic behavior from data guaranteeing stability using linear parameter varying (LPV) systems. This method enables learning complex task sequences with hidden Markov models (HMMs), where each state/subtask is given by a stable LPV system and where transitions are most likely around the corresponding attractor. We study the dynamics of the HMM-LPV model and propose a motion generation method that guarantees the stability of task sequences. We validate our approach in two sets of demonstrated human motions.
APA
Medina, J.R. & Billard, A.. (2017). Learning Stable Task Sequences from Demonstration with Linear Parameter Varying Systems and Hidden Markov Models. Proceedings of the 1st Annual Conference on Robot Learning, in Proceedings of Machine Learning Research 78:175-184 Available from https://proceedings.mlr.press/v78/medina17a.html.

Related Material