Decentralized control of robot joints based on neural network observer

In actual work, the system parameters of robot joints will change in real time or cannot be measured, and coupling relationship between various subsystems existence modeling errors make model difficult to determine. Based on such problems, a neural network observation is proposed Decentralized control method for robot. manipulator, firstly, each joint subsystem established by decentralized theory, nonlinear function approximation ability used approach uncertain part manipulator online through input output data. The design observer can estimate state system, use estimated sliding mode controller dynamically compensate unknown dynamics independent joint, realize self-control when speed information are unknown. Adaptive greatly enhances robustness adaptability robotic arm system. Finally, stability criterion given Lyapunov method, simulation results prove effectiveness method.