A Model-Free Resilient Tracking Control Scheme for Non-Affine Nonlinear Continuous Systems with Uncertainties

Abstract

In this paper, a model-free integral reinforcement learning (IRL) resilient tracking control scheme has been proposed for a class of non-affine nonlinear systems with uncertainties. By using the precompensation technique, the actual working control has been transformed into the state of compensator, and an augmented tracking system converts the optimal tracking control problem into an optimal regulation problem. Then applying the reinforcement learning (RL) technique, the designed method can avoid the dynamic information directly during the iteration learning process, and obtain the model-free resilient tracking control. Moreover, the developed algorithm overcomes the limitations and strict assumptions of conventional optimal tracking control in dealing with a non-affine nonlinear system. Besides, with the use of polynomial approximation structure and the least square method, practicability in the implementing process has been guaranteed effectively. Finally, two examples in the simulation are given to demonstrate the validity and feasibility of the proposed model-free IRL resilient tracking control.