Data-efficient, Explainable and Safe Box Manipulation: Illustrating the Advantages of Physical Priors in Model-Predictive Control

TL;DR

This paper demonstrates that incorporating physical priors into model-predictive control for robotic box manipulation enhances explainability, safety, and data efficiency, especially when prior knowledge of system dynamics is available.

Contribution

The paper provides a case-study showing how physical priors improve model-based RL/control in robotics, emphasizing benefits in safety, explainability, and data efficiency.

Findings

Improved safety and explainability in robotic manipulation tasks.

Reduced data requirements for effective control.

Enhanced generalization with less training data.

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.