Safe Adaptive Feedback Control via Barrier States

TL;DR

This paper introduces a novel adaptive control framework that ensures safety and stability for nonlinear systems with uncertainties by integrating barrier states into reinforcement learning-based control.

Contribution

It develops a barrier-state augmented adaptive dynamic programming method that guarantees safety and convergence without persistent excitation.

Findings

Successfully enforces safety via barrier states.

Ensures stability and parameter convergence.

Validated through obstacle-avoidance simulations.

Abstract

This paper presents a safe feedback control framework for nonlinear control-affine systems with parametric uncertainty by leveraging adaptive dynamic programming (ADP) with barrier-state augmentation. The developed ADP-based controller enforces control invariance by optimizing a value function that explicitly penalizes the barrier state, thereby embedding safety directly into the Bellman structure. The near-optimal control policy computed using model-based reinforcement learning is combined with a concurrent learning estimator to identify the unknown parameters and guarantee uniform convergence without requiring persistency of excitation. Using a barrier-state Lyapunov function, we establish boundedness of the barrier dynamics and prove closed-loop stability and safety. Numerical simulations on an optimal obstacle-avoidance problem validate the effectiveness of the developed approach.