Abstract: The trajectory tracking of elastic base and double flexible arms for a free-floating space robot system with fully flexible arms and elastic base is discussed. The elastic connection between the base and the arm is considered as a linear spring and dynamics model of the system which is regarded as a two- time scale system is derived. For the slow-varying subsystem, the recursive CMAC neural network (CMACNN) which has good dynamic characteristics is used to approximate uncertainties of dynamical equation and a recursive CMAC neural network control scheme which improves the tracking performance is designed to improve the tracking trajectory of rigid motion. For the fast-varying subsystem, an optimal linear quadratic regulator controller is adopted to damp out the vibration of the two flexible links and base elastic. Simulation results verify the effectiveness of the compound control method.

Key words: Space robot with fully flexible arms, elastic base, singular perturbation method, recurrent CMACNN, active vibration suppression