Position Control of Servomotors Using Neural Dynamic Sliding Mode

Abstract

In this paper, position control of servomotors is addressed. A radial basis function neural network is employed to identify the unknown nonlinear function of the plant model, and then a robust adaptive law is developed to train the parameters of the neural network, which does not require any preliminary off-line weight learning. Moreover, base on the identified model, we propose a new dynamic sliding mode control (DSMC) for a general class of nonaffine nonlinear systems by defining a new adaptive proportional-integral sliding surface and employing a linear state feedback. The main property of proposed controller is that it does not need an upper bound for the uncertainty and identified model; moreover, the switching gain increases and decreases according to the system circumstance by employing an adaptive procedure. Then, chattering is removed completely by using the DSMC with a small switching gain.