Fast calibration method for industrial robot kinematics based on motion capture system

This paper proposes a fast kinematics calibration method for industrial robots (IRs). Unlike existing approaches, the proposed method focuses on rapid calibration in complex and temporary IR-related scenarios. To achieve this purpose, the IR kinematic calibration is divided into four specific application situations, each using partial modules from the comprehensive calibration algorithm to streamline the process. The complete calibration method consists of two steps: first, determining the IR's joint axes through sequential joint movements and obtaining rough kinematic parameters, and second, refining these parameters using stochastic gradient descent within an optimization framework. In addition, the existing calibration methods typically ignore the positions of the intermediate links, which impacts IR obstacle avoidance. This paper proposes obstacle-avoidance auxiliary points to address this issue. Finally, simulation and experiment, conducted using a motion capture system, demonstrate that the proposed method can achieve fast, accurate calibration in complex scenarios.