Bio-mimetic Trajectory Tracking Control of Tendon driven Robot Using Delayed Force Feedback

Tendon driven robots can contact with environment softly at any time by grace of the intrinsic elasticity of actuators. Meanwhile it is difficult to manipulate them accurately due to their own flexibilities. In the biologic systems, the muscles are controlled by the spinal cord and cerebellum at the lowest level. Due to those control systems, we can manipulate our limbs as intended. Thus, this paper attempts to develop a new trajectory tracking controller for the tendon driven robot by integrating the computational models of the spinal cord and cerebellum in the same manner as the biological system. The proposed controller consists of the reflex control inputs composed of the delayed force feedback and PD position error feedback, and the adaptive controller which estimates and eliminates the excess reflex component of reflexes caused by the inertial force. The asymptotic stability of the controlled system is proven along the Liapunov-Krasovskii stability theorem. The effectiveness of the proposed controller and the dependence of the control performance on delays are examined through some computer simulations.