An Integrated Two-Level Self-Calibration Method for a Cable-Driven Humanoid Arm

摘要

This paper addresses the kinematic calibration issues for a 7-DOF cable-driven humanoid arm in order to improve its motion control accuracy. The proposed 7-DOF humanoid arm has a hybrid parallel-serial kinematic structure, which consists of three serially connected parallel cable-driven modules, i.e., a 3-DOF shoulder module, a 1-DOF elbow module, and a 3-DOF wrist module. Due to the unique arm design features such as hybrid parallel-serial structure, modular configuration, and redundant sensors, an integrated two-level self-calibration method is proposed in this work. The first level of self-calibration, termed as the central linkage mechanism calibration, is to identify the kinematics errors existed in the 7-DOF central linkage mechanism based on its self-motion capability. The second level of calibration, termed as the cable-driven module calibration, is to identify the kinematics errors existed in each of the parallel cable-driven modules based on its sensing redundancy. To simplify the formulation of the calibration algorithms, the error model of the serial central linkage mechanism is derived from its forward kinematics, in which the Products-Of-Exponential (POE) formula is employed, while the error models of the parallel cable-driven modules are derived from their inverse kinematics. The simulation and experimental results have shown that the proposed self-calibration algorithms can effectively improve the accuracy of the 7-DOF cable-driven humanoid arm.