Fault-tolerant Control And Active Vibration Suppression of Free-floating Flexible Space Robot

Ziwang Zhao,Li Chen
2016-01-01
Abstract:A method of decentralized fault tolerant control, based on decentralized back-stepping terminal sliding mode neural network control, is proposed for flexible space robot in the presence of partial loss of actuator effectiveness, where base’s attitude is controlled and base’s position isn’t controlled. Based on Lagrange second dynamics equation, assumed mode method and linear momentum conservation of the system, the dynamic model of flexible space robot is established. Furthermore, the system is decentralized into some subsystems, and then an effective factor is integrated into the dynamic model of each subsystem. To solve the problem of trajectory tracking control and active vibration suppression of the flexible space robot, the singular perturbation approach is adopted to decompose the robot system into a slow subsystem and a fast subsystem so that controller can be designed separately for the subsystems. With regard to the slow subsystem, the back-stepping terminal sliding mode techniques is used to design the control scheme which guarantees that the tracking errors of base’s attitude angle and space manipulator’s joint angles converge to zero in finite time, and with the help of radial basis function (RBF) neural networks, uncertain terms and unknown terms are estimated with adaptive compensation of estimation error. As for the fast subsystem, PD control is used to weaken the chattering of the terminal sliding mode controller and the vibration of the flexible arm effectively. Finally, the simulation results prove the validity of the controller.
What problem does this paper attempt to address?