ADAPTIVE-CONTROL OF NONLINEAR MULTIVARIABLE SYSTEMS USING NEURAL NETWORKS

Most practical systems have multiple inputs and multiple outputs, and the applicability of neural networks as practical adaptive controllers will eventually be judged by their success in multivariable problems. The representation, identification, and control of nonlinear multivariable systems are rendered difficult by the coupling as well as the delays that exist between the inputs and outputs. In the first part of the paper, theoretical questions related to system representation and existence of a desired control input are discussed. The second part of the paper develops a design methodology using neural networks. It is shown that under appropriate conditions, it may be possible to design efficient neural controllers for nonlinear multivariable systems for which linear controllers are inadequate.