APPLICATION OF MINIMUM PARAMETER LEARNING METHOD AND ARTIFICIAL NEURAL NETWORK FOR MANIPULATOR CONTROL

Abstract

This research designs and implements a robust adaptive sliding mode control (RASMC) based on radial basis function neural network (RBFNN) with minimum parameter learning (MPL) method for a manipulator system. This is a system that has been deployed in the construction materials manufacturing, metallurgy, mechanical engineering and shipbuilding industries. The robust adaptive SMC based on RBFNN is used to ensure the manipulator’s actual position following the desired in a finite time. The weight value parameters of the RBFNN are adjusted online by Quasi Newton algorithm according to adaptive laws for the purpose of controlling the output of the manipulator system to track a given trajectory. The minimum parameter learning (MPL) method is used in this study so that the system has only one online adaptive parameter, reduces the computational burden. The stability of the system is proven by Lyapunov theory. Simulation results in MATLAB/Simulink show the effectiveness of the proposed controller with the rising time, the settling time, the percent overshoot, the steady state error of link 1 are 0.0747(s), 0.1376(s), 0.002 (%),0(rad), and link 2 are 0.0844(s), 0.152(s), 0(%), 0(rad), respectively.