INVERSION OF RBF NETWORKS AND APPLICATIONS TO ADAPTIVE-CONTROL OF NONLINEAR-SYSTEMS

The paper investigates the application of inversion of a radial basis function network (RBFN) to nonlinear control problems for which the st ructure of the nonlinearity is unknown. Initially, the RBF network is trained to learn the forward dynamics of the plant. Two different cont roller structures are then proposed based on this identified RBFN mode l. In one scheme, a feedback control law is derived based on the input prediction by inversion of the RBFN model so that the system is Lyapu nov stable. The second kind of controller structure predicts the feedf orward control action, while the fixed controller actuates the feedbac k stabilising signal. An extended Kalman filtering based algorithm is employed to carry out the network inversion during each sampling inter val. Two examples are presented to verify the proposed scheme. Simulat ion results show that the performance of the controller based on the p roposed network inversion scheme is efficient.