ADAPTIVE COORDINATED CONTROL OF SPACE MANIPULATORS USING 'NEURAL' NETWORKS

摘要

High performance operation of free-flying manipulators requires coordinated control of both the manipulator and the spacecraft on which it is mounted. To ensure that uncertainty in the coupled dynamic model does not adversely impact performance, adaptative control techniques can be used. Previous results [15,18,19] have shown that considered use of neural networks can greatly extend more traditional adaptive control techniques, and can be applied to both fixed-base and free-floating space manipulators. In this paper, we demonstrate that by augmenting these designs with appropriate additional signals describing the motion of the spacecraft, the same control methodology provides an effective, reliable adaptive controller for space manipulators operating in free-flying mode. The stability and convergence properties of the algorithm are discussed, and the performance of the new controller is demonstrated on a simulated space manipulator system.