Assuring Safe Operation of Robotic Systems under Uncertainty: Control and Learning Methods

Assuring Safe Operation of Robotic Systems under Uncertainty: Control and Learning Methods applies set-theoretic and reinforcement learning approaches to formulate, analyze, and solve the challenge of ensuring safe operation of robotic systems in an uncertain environment. The authors adopt learning-supported, set-theoretic methods—specifically, the barrier Lyapunov function and the control barrier function—to achieve desirable robust safety with guaranteed performance in continuous-time nonlinear control applications. They also combine reinforcement learning with control theory to ensure safe learning and optimization. The reinforcement learning-based optimization framework incorporates safety and robustness guarantees by applying theoretical analysis tools from the field of control. This book will be of interest to researchers, engineers, and students specializing in robot planning and control.

Table of Contents:
1 Introduction to Safety under Uncertainty Section I Set-Theoretic Methods 2 Guaranteed Safety and Performance via Concurrent Learning 3 Provable Robust Safety Through Barrier Lyapunov Function 4 Safe Navigation via Integrated Perception and Control Section II Reinforcement Learning Approaches 5 Constrained Optimal Control Through Risk-Sensitive RL 6 Safe Approximate Optimal Control via Filtered RL 7 Time-Delayed Data Informed RL for Optimal Tracking Control

About the Author :
Cong Li earned a PhD from the Chair of Automatic Control Engineering, Technical University of Munich, Germany in 2022. He was also a research associate at the Chair of Automatic Control Engineering, Technical University of Munich. Yongchao Wang is at the Xi'an Research Institution of Hi-Technology and a professor at the School of Aerospace Science and Technology, Xidian University, Xi'an, China. He was at the Chair of Automatic Control Engineering, Technical University of Munich, Germany. Fangzhou Liu received the Doktor-Ingenieur degree in electrical engineering from the Technical University of Munich, Germany in 2019. He was a lecturer and a research fellow at the Chair of Automatic Control Engineering, Technical University of Munich, Germany. He is now a full professor at the School of Astronautics, Harbin Institute of Technology, Harbin, China. Xinglong Zhang earned a BE in mechanical enineering from Zhejiang University, Hangzhou, China in 2011 and a PhD in system and control from the Politecnico di Milano, Italy, 2018. He is presently an associate professor at the College of Intelligence Science and Technology, National University of Defense Technology, Changsha, China. His research interests include Koopman operators, learning-based model predictive control, reinforcement learning, and approximate dynamic programming, and their applications in automotive systems.