Resilient inverse optimal control for tracking: Overcoming process noise challenges

This paper studies the Inverse Optimal Control (IOC), aiming to identify the underlying cost functions using observed optimal control paths. An innovative IOC algorithm is developed in this paper by leveraging the closed-loop control law of optimal tracking control, without needing to consider any prior knowledge of the process noise. More explicitly, a convex optimization problem is formulated for the IOC problem by encompassing various linear constraints. The contributions of our work include: (i) Robustly handling process noise, ensuring accuracy without excessive data. (ii) Deriving linear conditions for optimal tracking control law, leading to a closed-form IOC solution that can yield the global optimal solution under sufficient conditions. (iii) No extra LMI constraints are needed when dealing with diverse reference signals. The paper concludes by demonstrating our approach's effectiveness through simulations and comparisons with baseline methods.