Backstepping position tracking controller design with neural network deadzone compensation

This paper presents a deadzone compensation scheme using neural networks together with a controller designed by backstepping method. The control system consists of a deadzone precompensator with two neural networks following a backstepping controller. The neural precompensator is employed to improve the performance of the backstepping controller. The proposed control scheme shows superior transient and steady-state performance compared to conventional PD and PID controllers. Moreover, the scheme is robust to uncertainty in deadzone nonlinearities. In the end, computer simulation examples are given to illustrate the effectiveness of the scheme