A unified approach of control and identification for optimal control of non-linear systems

Abstract

This paper proposes a novel unified approach of control and identification based on a new control performance criterion to maximize the performance of optimal control of unknown non-linear systems. Hamilton-Jacobi-Bellman equation is used for developing the proposed criterion. Because the proposed criterion still includes unknown components due to the non-linearity of the system, two important ideas are proposed to represent the criterion as a function of identified system parameters. After constructing the proposed criterion which maximizes the control performance, the system identification and the controller design are performed simultaneously based on the criterion. A numerical simulation shows that the proposed method improves the optimal control performance compared with the other methods.