Safe Adaptive Learning-based Control for Constrained Linear Quadratic Regulators with Regret Guarantees

Abstract

We study the adaptive control of an unknown linear system with a quadratic cost function subject to safety constraints on both the states and actions. The challenges of this problem arise from the tension among safety, exploration, performance, and computation. To address these challenges, we propose a polynomial-time algorithm that guarantees feasibility and constraint satisfaction with high probability under proper conditions. Our algorithm is implemented on a single trajectory and does not require system restarts. Further, we analyze the regret of our learning algorithm compared to the optimal safe linear controller with known model information. The proposed algorithm can achieve a $\tilde{O}(T^{2/3})$ regret, where $T$ is the number of stages and $\tilde{O}(\cdot )$ absorbs some logarithmic terms of $T$.

References in corpus (9)

• Global Convergence of Policy Gradient Methods for the Linear Quadratic Regulator
• LMI Properties and Applications in Systems, Stability, and Control Theory
• Online Control with Adversarial Disturbances
• Logarithmic Regret for Online Control
• Learning nonlinear dynamical systems from a single trajectory
• Adaptive MPC under Time Varying Uncertainty: Robust and Stochastic
• Non-asymptotic and Accurate Learning of Nonlinear Dynamical Systems
• Online Optimal Control with Affine Constraints
• Geometric Exploration for Online Control