Uncertainty-Aware Constraint Learning for Adaptive Safe Motion Planning from Demonstrations

Glen Chou, Dmitry Berenson, Necmiye Ozay
Proceedings of the 2020 Conference on Robot Learning, PMLR 155:1612-1639, 2021.

Abstract

We present a method for learning to satisfy uncertain constraints from demonstrations. Our method uses robust optimization to obtain a belief over the potentially infinite set of possible constraints consistent with the demonstrations, and then uses this belief to plan trajectories that trade off performance with satisfying the possible constraints. We use these trajectories in a closed-loop policy that executes and replans using belief updates, which incorporate data gathered during execution. We derive guarantees on the accuracy of our constraint belief and probabilistic guarantees on plan safety. We present results on a 7-DOF arm and 12D quadrotor, showing our method can learn to satisfy high-dimensional (up to 30D) uncertain constraints and outperforms baselines in safety and efficiency.

Cite this Paper

BibTeX
@InProceedings{pmlr-v155-chou21a, title = {Uncertainty-Aware Constraint Learning for Adaptive Safe Motion Planning from Demonstrations}, author = {Chou, Glen and Berenson, Dmitry and Ozay, Necmiye}, booktitle = {Proceedings of the 2020 Conference on Robot Learning}, pages = {1612--1639}, year = {2021}, editor = {Kober, Jens and Ramos, Fabio and Tomlin, Claire}, volume = {155}, series = {Proceedings of Machine Learning Research}, month = {16--18 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v155/chou21a/chou21a.pdf}, url = {https://proceedings.mlr.press/v155/chou21a.html}, abstract = {We present a method for learning to satisfy uncertain constraints from demonstrations. Our method uses robust optimization to obtain a belief over the potentially infinite set of possible constraints consistent with the demonstrations, and then uses this belief to plan trajectories that trade off performance with satisfying the possible constraints. We use these trajectories in a closed-loop policy that executes and replans using belief updates, which incorporate data gathered during execution. We derive guarantees on the accuracy of our constraint belief and probabilistic guarantees on plan safety. We present results on a 7-DOF arm and 12D quadrotor, showing our method can learn to satisfy high-dimensional (up to 30D) uncertain constraints and outperforms baselines in safety and efficiency.} }
Endnote
%0 Conference Paper %T Uncertainty-Aware Constraint Learning for Adaptive Safe Motion Planning from Demonstrations %A Glen Chou %A Dmitry Berenson %A Necmiye Ozay %B Proceedings of the 2020 Conference on Robot Learning %C Proceedings of Machine Learning Research %D 2021 %E Jens Kober %E Fabio Ramos %E Claire Tomlin %F pmlr-v155-chou21a %I PMLR %P 1612--1639 %U https://proceedings.mlr.press/v155/chou21a.html %V 155 %X We present a method for learning to satisfy uncertain constraints from demonstrations. Our method uses robust optimization to obtain a belief over the potentially infinite set of possible constraints consistent with the demonstrations, and then uses this belief to plan trajectories that trade off performance with satisfying the possible constraints. We use these trajectories in a closed-loop policy that executes and replans using belief updates, which incorporate data gathered during execution. We derive guarantees on the accuracy of our constraint belief and probabilistic guarantees on plan safety. We present results on a 7-DOF arm and 12D quadrotor, showing our method can learn to satisfy high-dimensional (up to 30D) uncertain constraints and outperforms baselines in safety and efficiency.
APA
Chou, G., Berenson, D. & Ozay, N.. (2021). Uncertainty-Aware Constraint Learning for Adaptive Safe Motion Planning from Demonstrations. Proceedings of the 2020 Conference on Robot Learning, in Proceedings of Machine Learning Research 155:1612-1639 Available from https://proceedings.mlr.press/v155/chou21a.html.

Related Material