Stiffness Optimization of Cable-Driven Humanoid Manipulators

Cable-driven humanoid manipulators offer significant advantages in terms of lightweight design, flexibility, and safety, thus they have promising prospects for applications with frequent human–robot interactions. However, cable flexibility and joint mechanism lead to low stiffness in the cable-driven joint, resulting in a reduction in the Cartesian stiffness of the humanoid manipulators. To improve the stiffness, we optimize the stiffness performance of the manipulators from two aspects: joint stiffness and Cartesian stiffness. To enhance the joint stiffness, we design a cable-driven joint based on the tension amplification principle of movable pulley groups. The amplification ratio of the joint stiffness is proportional to the square of the amplification factor of the movable pulley groups, and the stiffness does not vary with the joint angle. On the other hand, to improve the Cartesian stiffness, we propose a stiffness optimization algorithm for fixed point and continuous path tasks by combining the kinematic and stiffness models of humanoid manipulators. Finally, two cable-driven humanoid manipulator prototypes are constructed, and experimental results demonstrate that the cable-driven joint and proposed algorithm can effectively improve both the joint stiffness and Cartesian stiffness during path planning tasks.