Experiment Design with Gaussian Process Regression with Applications to Chance-Constrained Control

TL;DR

This paper presents an experimental design method using Gaussian process regression to improve data collection for data-driven control, especially in chance-constrained problems, demonstrating superior performance over benchmarks.

Contribution

The paper introduces a novel experimental design approach leveraging Gaussian processes to optimize data collection for control tasks with uncertain dynamics.

Findings

Outperforms benchmark methods in numerical tests

Effective in chance-constrained control scenarios

Utilizes Gaussian process structure for better experiment design

Abstract

Learning for control in repeated tasks allows for well-designed experiments to gather the most useful data. We consider the setting in which we use a data-driven controller that does not have access to the true system dynamics. Rather, the controller uses inferred dynamics based on the available information. In order to acquire data that is beneficial for this controller, we present an experimental design approach that leverages the current data to improve expected control performance. We focus on the setting in which inference on the unknown dynamics is performed using Gaussian processes. Gaussian processes not only provide uncertainty quantification but also allow us to leverage structures inherent to Gaussian random variables. Through this structure, we design experiments via gradient descent on the expected control performance with respect to the experiment input. In particular, we focus on a chance-constrained minimum expected time control problem. Numerical demonstrations of our approach indicate our experimental design outperforms relevant benchmarks.