Fuzzy Adaptive Compensation Control for Space Manipulator with Joint Flexibility and Dead Zone Based on Neural Network

Considering the problems of dead zone and flexibility in the joint transmission mechanism of space manipulator, a fuzzy compensation control method based on neural network is proposed. Dynamic equation of the system is established by the system’s linear and angular momentum conservation and Lagrange equation. Based on singular perturbation theory, it is decomposed into two subsystem models of fast variable and slow variable for control, respectively. A moment difference feedback controller is designed to suppress the elastic vibration for the fast-changing flexible subsystem model. Aiming at the unknown uncertainties in the slowly varying stiffness subsystem model, a radial basis function neural network controller is designed to approximate the unknown model and its approximation error is eliminated by a robust controller. Aiming at the dead zone link in the joint transmission mechanism, the mathematical relationship among dead zone estimation, dead zone compensator and controller is deduced. The dead zone estimator and dead zone compensator based on adaptive fuzzy system are designed to realize the online real-time estimation and compensation of dead zone, solving the tracking error caused by joint dead zone and improving the control accuracy. The parameter adaptive learning rate of the designed fuzzy system can realize online real-time adjustment without off-line learning stage. Based on Lyapunov theory, the uniform final boundedness of the signals of the whole closed-loop system is proved. Simulation results verify the effectiveness of the proposed control algorithm.