Adaptive fuzzy-neural network effectively disturbance compensate in sliding mode control for dual arm robot

Abstract

In this study, an Adaptive Backstepping Sliding Mode Controller (ABSMC) is introduced based on the Radial Basis Function (RBF) neural network and a fuzzy logic modifier. The proposed method is used to control a Dual-Arm Robot (DAR) – a nonlinear structure with unstable parameters and external disturbances. The control aims to track the motion trajectory of both arms in the flat surface coordinate within a short time, maintaining stability, and ensuring that the tracking error converges in finite time, especially when influenced by unforeseen external disturbances. The nonlinear Backstepping Sliding Mode Control (BSMC) is effective in trajectory tracking control; however, undesired phenomena may occur if there are uncertain disturbances affecting the system or model parameters change. It is proposed to use a neural network to estimate a nonlinear function to handle unknown uncertainties of the system. The neural network parameters can be adaptively adjusted to optimal values through adaptation rules derived from Lyapunov's theorem. Additionally, fuzzy logic theory is also employed to adjust the controller parameters to accommodate changes or unexpected impacts. The performance of the Fuzzy Neural Network Backstepping Sliding Mode Control (FNN-BSMC) is evaluated through simulation results using Matlab/Simulink software. Two simulation cases are conducted: the first case assumes stable model parameters without uncertain disturbances affecting the joints, while the second case considers a model with changing parameters and disturbances. Simulation results demonstrate the effective adaptability of the proposed method when the system model is affected by various types of uncertainties from the environment