New Smith Predictor and RBF Neural Network Control for Networked Control Systems

This paper aims to the random and uncertain network delay, as well as controlled plant might be time-variant or nonlinear, a novel approach is proposed that novel Smith predictor combined with the radial basis function neural network (RBFNN) control for the networked control systems (NCS). This novel Smith predictor comes true to hide predictor models of the network delays of the forward and return paths into real network data transmission processes, therefore network delays donpsilat need to be measured, identified or estimated on-line. It is applicable to some occasions that the network delays are larger than one, even tens of sampling periods. Based on CSMA/CD (Ethernet), the results of simulation show validity of this control scheme, and indicate that system has better dynamic performance and anti-jamming ability.