Sufficiently Accurate Model Learning

TL;DR

This paper introduces a regularization method for data-driven model learning in robotics, producing models with predictable error characteristics that improve control and planning performance.

Contribution

It proposes a primal-dual approach to learn Sufficiently Accurate models with explicit error constraints, enhancing model reliability for control tasks.

Findings

Models have more predictable error profiles.

Improved control and planning performance in simulations.

Explicit error constraints lead to better model utility.

Abstract

Modeling how a robot interacts with the environment around it is an important prerequisite for designing control and planning algorithms. In fact, the performance of controllers and planners is highly dependent on the quality of the model. One popular approach is to learn data driven models in order to compensate for inaccurate physical measurements and to adapt to systems that evolve over time. In this paper, we investigate a method to regularize model learning techniques to provide better error characteristics for traditional control and planning algorithms. This work proposes learning "Sufficiently Accurate" models of dynamics using a primal-dual method that can explicitly enforce constraints on the error in pre-defined parts of the state-space. The result of this method is that the error characteristics of the learned model is more predictable and can be better utilized by planning and control algorithms. The characteristics of Sufficiently Accurate models are analyzed through experiments on a simulated ball paddle system.