Safety and Optimality in Learning-Based Control at Low Computational Cost

• 成果类型: Article
• 署名作者: Baumann, Dominik; Kowalczyk, Krzysztof; Rojas, Cristian R.; Tiels, Koen; Wachel, Pawel
• 署名单位: Aalto University; Uppsala University; Wroclaw University of Science & Technology; Royal Institute of Technology; Eindhoven University of Technology
• 刊物名称: IEEE TRANSACTIONS ON AUTOMATIC CONTROL
• ISSN/ISSBN: 0018-9286
• DOI: 10.1109/TAC.2025.3570248
• 发表日期: 2025
• 关键词: bayesian optimization; Gaussian process

摘要

Applying machine learning methods to physical systems that are supposed to act in the real world requires providing safety guarantees. However, methods that include such guarantees often come at a high computational cost, making them inapplicable to large datasets and embedded devices with low computational power. In this article, we propose CoLSafe, a computationally lightweight safe learning algorithm whose computational complexity grows sublinearly with the number of data points. We derive both safety and optimality guarantees and showcase the effectiveness of our algorithm on a seven-degrees-of-freedom robot arm.