Radial Basis Functions Neural Network Based Self-Tuning Regulator

In this paper a new technique is proposed to design an online control algorithm using the Radial Basis Functions Neural Network (RBFNN). The controller is an RBFNN based direct self-tuning regulator (STR) that overcomes several shortcomings of the inverse control design using the neural networks. The control algorithm performs equally good to both minimum phase and non-minimum phase plants. The plant parameters are estimated online and are used to update the weights of the RBFNN. The weight update equations are derived based on the well known least mean squares principle. The RBFNN virtually models the inverse of the plant and thus the output tracks the reference trajectory. The proposed algorithm is successfully verified using simulations for both minimum and non-minimum phase plants.