A Neural Network-Based Fast Terminal Sliding Mode Controller for Dual-Arm Robots

The paper studies a robust adaptive fast terminal sliding mode controller (FTSMC) based on a radial basis function neural network (RBF NN) for a dual-arm robot manipulator system that coordinates the motion of the general object. First, kinematics, a general dynamic model of the system consisting of manipulators and the object, is inferred about the position and direction of the object as the states of the derived model. Second, an FTSMC method is designed, followed by the construction of two RBF systems: one approximates uncertainties and external disturbances, and the other estimates the force applied to the object according to the object’s uncertainty model. Next, the Lyapunov theory is employed to demonstrate the stability of the closed-loop system and derive the adaptation laws for RBF NN. Finally, simulation results of the dual-arm robot system with three degrees of freedom (3-DOF) manipulators are presented to illustrate the effectiveness of the proposed method.