Fuzzy RBF Neural Network Control for Networked Control Systems Based on Modified Smith Predictor

Network delay highly degrades the control performance of networked control systems (NCS). Aiming to random and uncertain network delay, time-variant or nonlinear controlled plant and imprecise Smith predictor models, a novel approach is proposed that modified Smith predictor combined with fuzzy radial basis function neural network (FRBFNN). This approach can identify the controlled plant on-fine, timely adjusts the weights of the adaptive controller. Because modified Smith predictor does not include network delays, therefore, it is no need for measuring, identifying or estimating network delays on-line. It is applicable to some occasions that network delays are larger than one, even tens of sampling periods. Based on CSMA/CD (Ethernet) and CSMA/AMP (CAN Bus), the simulation results show that modified Smith dynamic predictor combined with FRBFNN is effective.