Local Bayesian Optimization for Controller Tuning with Crash Constraints

Abstract

Controller tuning is crucial for closed-loop performance but often involves manual adjustments. Although Bayesian optimization (BO) has been established as a data-efficient method for automated tuning, applying it to large and high-dimensional search spaces remains challenging. We extend a recently proposed local variant of BO to include crash constraints, where the controller can only be successfully evaluated in an a-priori unknown feasible region. We demonstrate the efficiency of the proposed method through simulations and hardware experiments. Our findings showcase the potential of local BO to enhance controller performance and reduce the time and resources necessary for tuning.

Published in at-Automatisierungstechnik

References in corpus (8)

• Automatic LQR Tuning Based on Gaussian Process Global Optimization
• Robot Learning with Crash Constraints
• Noisy-Input Entropy Search for Efficient Robust Bayesian Optimization
• On Controller Tuning with Time-Varying Bayesian Optimization
• Local policy search with Bayesian optimization
• Benchmark of Bayesian Optimization and Metaheuristics for Control Engineering Tuning Problems with Crash Constraints
• The Behavior and Convergence of Local Bayesian Optimization
• Local Bayesian optimization via maximizing probability of descent