Model reference adaptive control based on MLP network for dynamic system

Abstract

This paper presents the Model Reference Adaptive Control based on Multi Layered Perceptron (MLP) Network. The error between the output of the plant and the reference model is used to adapt the controller parameters. The Reference Model Adaptive Control (MRAC) will be used to control a time varying system where the performance specifications are given in terms of a reference model. The MLP is used for function approximation and its weights can be adjusted to model any function up to certain accuracy from input to output using the Recursive Least Square (RLS) algorithm. Simulation results show that the controller is adequate to control the system with unpredictable conditions and disturbances.