Robust Safety and Fixed-time stability for Multiagent System
Dr. Kunal Garg, Postdoctoral Associate, REALM lab, Department of Aeronautics and Astronautics, MIT
Friday, September 16, 2022 at 2:30 – 3:30 PM in Engineering: 231
Abstract: Various types of constraints are present in real-world applications due to structural and operational requirements. For example, spatial constraints, i.e., constraints requiring the system trajectories to evolve in a safe set at all times, while visiting goal set(s), are common in safety-critical applications. Furthermore, temporal constraints, i.e., constraints on convergence within a given user-defined time, appear in time-critical applications, for instance when a task must be completed within a given time frame. Thus, it is desirable to synthesize control input that can fulfill such spatiotemporal requirements while satisfying input constraints. We will present a quadratic program (QP)-based formulation to compute the control input that renders a safe set forward-invariant and drives the closed-loop trajectories to a goal set within a user-defined time in the presence of input constraints. Such formulations are effective from a practical point of view since QPs can be solved very efficiently for real-time implementation. As a case study, we will demonstrate how this formulation can be used for fault-tolerant control of a quadrotor, when its motors are faulty or under an adversarial cyber-attack.
Bio: Dr. Garg is a postdoctoral associate in the REALM lab in the Department of Aeronautics and Astronautics at MIT. Prior to joining MIT, he spent a year with the Hybrid Systems Lab at UC Santa Cruz as a postdoctoral scholar. He received his Bachelor of Technology degree in Aerospace Engineering from the Indian Institute of Technology, Mumbai, India in 2016, Master of Science in Engineering degree in Aerospace Engineering from the University of Michigan, Ann Arbor in 2019, and Ph.D. from the University of Michigan, Ann Arbor in Aerospace Engineering in 2021. His research interests include Machine-learning approaches for the safety of multi-agent systems, Control-theoretic methods for the security of CPS, Optimization-based control design, Multi-agent distributed control, Finite-time stability theory for safe distributed control design, Hybrid/Switched-Systems: theory and applications, and Accelerated optimization methods.