Finite-time adaptive optimal control of uncertain strict-feedback nonlinear systems based on fuzzy observer and reinforcement learning

Abstract

This paper proposes an adaptive optimal control strategy of finite-time control for high-order uncertain strict-feedback nonlinear systems. Firstly, a reinforcement learning (RL) based an optimal control scheme is employed to design a optimal controller, to achieve global optimisation. Additionally, considering the unmeasurable states, we construct a fuzzy observer and utilise fuzzy logic systems to approximate the unknown functions. Meanwhile, the inclusion of command filtering and time-based control simplifies the controller design and enhances the system's response rapidity. Finally, the effectiveness and feasibility of the proposed approach are validated through a numerical simulation and a single link-robot system simulation.