GAUSSIAN NEURAL NETWORKS FOR CONTROL FUNCTION IMPLEMENTATION

The systematic storage in neural networks of prior information to be u sed in the design of various control subsystems is investigated. Assum ing that the prior information is available in a certain form (namely, input/output data points and specifications between the data points), a particular neural network using Gaussian nonlinearities and a corre sponding parameter design method are introduced. The proposed neural n etwork addresses the issue of effectively using prior information, and its use is illustrated in an implementation of a control law schedule r.