ADAPTIVE CONTROL OF A CABLE-DRIVEN SERPENTINE MANIPULATOR BASED ON NEURAL NETWORK OBSERVER

Liang Han, Zhentao Li, and Yunzhi Huang

Keywords

Cable-driven manipulator, state observer, neural network, output feedback control, trajectory tracking

Abstract

For the cable-driven serpentine manipulator (CDSM), the imprecise dynamics and the unmeasurable joint states pose challenges to its high-precision trajectory tracking. In this paper, output feedback control with neural network (NN) and state observer for CDSM is proposed to achieve high-precision trajectory tracking. For the unmeasurable joint state, a state observer with an adaptive NN is constructed. Another adaptive NN is used for output feedback control to compensate for the dynamics of the CDSM and unknown disturbances. It can be proved by Lyapunov stability theory that all signals are ultimately bounded. Based on the above analysis process, we carried out numerical simulations. The simulation results proved that the states of the joint can be evaluated effectively by the proposed NN-based observer, and the constructed NN-based output feedback control scheme has stronger anti-disturbance ability and better trajectory tracking performance than traditional PD control law.

Important Links:

Go Back