Upcoming Event: Center for Autonomy Seminar

Safe Control for Learning-Enabled Autonomous Systems

Xiangru Xu, Assistant Professor, Dept. of Mechanical Engineering, University of Wisconsin-Madison

Abstract

This talk presents recent advances in provably safe control for learning-enabled autonomous systems. In the first part, I will discuss reachability analysis and controlled invariance of neural network control systems. I will introduce methods for safety verification and safe control synthesis based on forward and backward reachable set computations using constrained and hybrid zonotopes, together with an interval-based invariance operator for computing the maximum controlled invariant set. In the second part, I will present hierarchical safe control architectures that integrate a high-level optimization-based motion planner with a low-level safety filter, providing formal guarantees of continuous-time safety constraint satisfaction. I will demonstrate their effectiveness through two case studies: safe trajectory planning and tracking for quadrotors, and occlusion-free visual servoing of robotic manipulators.

Biography

Xiangru Xu is an Assistant Professor in the Department of Mechanical Engineering at the University of Wisconsin-Madison. Prior to joining UW-Madison, he held postdoctoral positions in the Department of Electrical Engineering and Computer Science at the University of Michigan and in the Department of Aeronautics and Astronautics at the University of Washington. He received his Ph.D. in Systems and Control from the Chinese Academy of Sciences. He is a recipient of the NSF CAREER Award and the Best New Application Paper Award from IEEE Transactions on Automation Science and Engineering. He currently serves as an Associate Editor for IEEE Transactions on Automatic Control, Control Theory and Technology, and Autonomous Intelligent Systems, and as a member of the IEEE Control Systems Society Conference Editorial Board.