Smith Predictive Control Based on Neural Network

摘要

A new Smith predictive controller based on neural network is proposed in this paper. By using theories of Smith prediction, self-learning and nonlinear mapping of neural network, the control problem of controlled object with time-varying and time-delay parameters can effectively be solved. By adding auxiliary feed-back loop, the learning rate of neural network is accelerated. The research of simulation of process hatching bird eggs indicates that the presented method is one with stronger robustness, rather rapid rate of convergence and favorable behavior of dynamic and static state, which can newly, effectively and practicably settle the problem in the control over object with time-varying and time-delay parameters.