Networked Learning Control Based on Multi-step Prediction Algorithm of Cubic Spline Rolling Optimization

For the adverse effect caused by network delay on networked-learning-based control,a composite control is proposed based on multi-step prediction model of cubic spline rolling optimization and on-line adjustment of neural network.Tracking control simulation is performed on the complex,time-varying and nonlinear controlled object with external disturbances in Token network.Simulation results show that the adverse effect caused by network delay on the controlled complex object can be overcome,and rapidity and high-precision tracking can be achieved by applying the composite control strategy.