Control of systems with deadzones using neural-network based learning controller

Conventional controllers, such as PD or PID controllers, are widely used in industrial applications, since it is simple, cheap and robust. Such controllers exhibit poor performance when applied to systems containing non-smooth nonlinearity. In this paper, the authors present a neural-network based learning controller for systems having a non-smooth nonlinearity with unknown parameters, specifically, a deadzone. The control scheme consists of a conventional PD controller and CMAC network. The authors illustrate the effectiveness of their scheme using computer simulation examples