TIME-DELAY CONTROL OF NONLINEAR-SYSTEMS WITH NEURAL-NETWORK MODELING

Time Delay Control (TDC) that uses a multilayer neural network as a no nlinear plant modeler is proposed in this paper. In the proposed contr oller structure, TDC is used to compensate for the changes of the plan t and/or uncertainties including neural network modeling errors. The n eural network adapts to learn the uncertainties and the changes in the system, which are eventually embedded into the neural network. In thi s way, the proposed method exhibits short-term adaptability through TD C and long-term adaptability through neural network adaptation. Becaus e the method uses a neural network as a modeler, it can be effective f or the control of nonlinear systems which are hard to model in an anal ytic way; it also has the ability to cancel out unmodeled dynamics. It is proved that neural network learning error and control error is uni formly bounded. Computer experiments reveal that the proposed algorith m is effective in controlling nonlinear systems. (C) 1997 Elsevier Sci ence Ltd.