ADAPTIVE SLIDING MODE CONTROL BASED ON RBF NEURAL NETWORK FOR TWO TANKS INTERACTING SYSTEM

Abstract

In this paper, an adaptive radial basis function neural network (RBFNN) is proposed to deal with chattering reduction problem in sliding mode control for the two tanks interacting system. The RBFNN is used to approximate the function in the sliding mode control. The signum function in the sliding mode control is replaced by tanh function to test the performance of the chattering reduction problem.  The stability of the proposed algorithm is proved by the Lyapunov theory. To show the suitability of the proposed algorithm, the simulation results in MATLAB/Simulink of this method are compared to the fuzzy control, sliding mode control with conditional integrals, fuzzy PID control and the conventional PID control. The comparison results show that the proposed controller is more effective with the rise time is 0.1271 (s), the percent overshoot is 0 (%), the steady state error converges to zero, the settling time is 0.2464 (s) and the chattering is eliminated.