Modeling and control strategy of flexible joint servo system in humanoid manipulator driven by tendon-sheath

The control effect of rotational speed in joints directly affects the motion accuracy of a humanoid manipulator driven by tendon-sheath. The dynamic parameters of the joint have time-varying characteristics due to the posture change of the manipulator. The joint driven by tendon-sheath has a specific torsional stiffness, so flexibility should be considered in the humanoid manipulator's servo system. The time-varying of the parameters in the servo system and the joint flexibility can cause fluctuation of the output speed. To improve the motion accuracy of the humanoid manipulator, a fuzzy-tuned PI control strategy is used to suppress the instability of the output speed. First, the change law of the inertia on the motor side of the flexible joint is calculated by the dynamics equation of the humanoid manipulator. Next, a mathematical model of the joint is established, and the transfer function from the load speed to the electromagnetic torque is obtained. Furthermore, according to the pole-placement strategy, the fuzzy-tuned PI controller parameters are selected appropriately for the manipulator in different postures. Finally, the effectiveness of the proposed method is verified by numerical simulations and control experiments of the manipulator. The results show that the fuzzy-tuned PI control strategy can significantly reduce the tracking errors and improve the control performance of the manipulator.