Data-efficient, explainable and safe box manipulation: Illustrating the advantages of physical priors in model-predictive control

Achkan Salehi, Stephane Doncieux
Proceedings of the 6th Annual Learning for Dynamics & Control Conference, PMLR 242:13-24, 2024.

Abstract

Model-based RL/control have gained significant traction in robotics. Yet, these approaches often remain data-inefficient and lack the explainability of hand-engineered solutions. This makes them difficult to debug/integrate in safety-critical settings. However, in many systems, prior knowledge of environment kinematics/dynamics is available. Incorporating such priors can help address the aforementioned problems by reducing problem complexity and the need for exploration, while also facilitating the expression of the decisions taken by the agent in terms of physically meaningful entities. Our aim with this paper is to illustrate and support this point of view via a case-study. We model a payload manipulation problem based on a real robotic system, and show that leveraging prior knowledge about the dynamics of the environment in an MPC framework can lead to improvements in explainability, safety and data-efficiency, leading to satisfying generalization properties with less data.

Cite this Paper

BibTeX
@InProceedings{pmlr-v242-salehi24a, title = {Data-efficient, explainable and safe box manipulation: {I}llustrating the advantages of physical priors in model-predictive control}, author = {Salehi, Achkan and Doncieux, Stephane}, booktitle = {Proceedings of the 6th Annual Learning for Dynamics & Control Conference}, pages = {13--24}, year = {2024}, editor = {Abate, Alessandro and Cannon, Mark and Margellos, Kostas and Papachristodoulou, Antonis}, volume = {242}, series = {Proceedings of Machine Learning Research}, month = {15--17 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v242/salehi24a/salehi24a.pdf}, url = {https://proceedings.mlr.press/v242/salehi24a.html}, abstract = {Model-based RL/control have gained significant traction in robotics. Yet, these approaches often remain data-inefficient and lack the explainability of hand-engineered solutions. This makes them difficult to debug/integrate in safety-critical settings. However, in many systems, prior knowledge of environment kinematics/dynamics is available. Incorporating such priors can help address the aforementioned problems by reducing problem complexity and the need for exploration, while also facilitating the expression of the decisions taken by the agent in terms of physically meaningful entities. Our aim with this paper is to illustrate and support this point of view via a case-study. We model a payload manipulation problem based on a real robotic system, and show that leveraging prior knowledge about the dynamics of the environment in an MPC framework can lead to improvements in explainability, safety and data-efficiency, leading to satisfying generalization properties with less data.} }
Endnote
%0 Conference Paper %T Data-efficient, explainable and safe box manipulation: Illustrating the advantages of physical priors in model-predictive control %A Achkan Salehi %A Stephane Doncieux %B Proceedings of the 6th Annual Learning for Dynamics & Control Conference %C Proceedings of Machine Learning Research %D 2024 %E Alessandro Abate %E Mark Cannon %E Kostas Margellos %E Antonis Papachristodoulou %F pmlr-v242-salehi24a %I PMLR %P 13--24 %U https://proceedings.mlr.press/v242/salehi24a.html %V 242 %X Model-based RL/control have gained significant traction in robotics. Yet, these approaches often remain data-inefficient and lack the explainability of hand-engineered solutions. This makes them difficult to debug/integrate in safety-critical settings. However, in many systems, prior knowledge of environment kinematics/dynamics is available. Incorporating such priors can help address the aforementioned problems by reducing problem complexity and the need for exploration, while also facilitating the expression of the decisions taken by the agent in terms of physically meaningful entities. Our aim with this paper is to illustrate and support this point of view via a case-study. We model a payload manipulation problem based on a real robotic system, and show that leveraging prior knowledge about the dynamics of the environment in an MPC framework can lead to improvements in explainability, safety and data-efficiency, leading to satisfying generalization properties with less data.
APA
Salehi, A. & Doncieux, S.. (2024). Data-efficient, explainable and safe box manipulation: Illustrating the advantages of physical priors in model-predictive control. Proceedings of the 6th Annual Learning for Dynamics & Control Conference, in Proceedings of Machine Learning Research 242:13-24 Available from https://proceedings.mlr.press/v242/salehi24a.html.

Related Material