CODE GENERATION OF A DRIVER FOR CMAC NN AND ITS IMPLEMENTATION FOR MIMO NONLINEAR SYSTEM

Abstract

Almost all difficult control problems are nonlinear, but few universally applicable nonlinear control design approaches are broadly used in practice. Neural Networks have important capability of learning the nonlinear mappings to any extent of accuracy. It is this property, which immediately opens the field: of control theory for the neural network applications. The field of neurocontrol has rapidly expanded over the last several years by virtue of its incontestable success in solving practical control problems. These Neurai controllerscan be simulated using CMAC. This Project is aimed in this direction to provide code of CMAC driver which trains waig. hted factor on the basis of training data & * odify these factor sdynamical. This driver can be used for MIMO nonlinear dynamical system. The driver is based on CMAC proposed by ALBUS [8-10] with some modification in it.