Local Bayesian optimization for controller tuning with crash constraints

at - Automatisierungstechnik, vol. 72, pp. 281–292

Abstract

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.

Authors 3

1. Alexander von Rohr corresponding Aachen

   RWTH Aachen University

   Affiliation as printed

   9165 Institute for Data Science in Mechanical Engineering, RWTH Aachen University , Aachen , Germany

2. Dominik Scheurenberg Aachen

   RWTH Aachen University

   Affiliation as printed

   9165 Institute of Automatic Control, RWTH Aachen University , Aachen , Germany

3. Sebastian Trimpe Aachen

   RWTH Aachen University

   Affiliation as printed

   9165 Institute for Data Science in Mechanical Engineering, RWTH Aachen University , Aachen , Germany