Optimal control design based on reinforcement learning for a class of nonlinear distributed systems

Abstract

This paper proposes an optimal control scheme for a class of non-affine nonlinear distributed systems. The research is conducted for a tethered parafoil system. The reference inputs are optimized by reinforcement learning method for two optimization goals respectively. A dynamic model approximation method is introduced to approximate the non-affine nonlinear terms. The tracking controller is designed and the stability analysis is given. The methodology is demonstrated by simulations.