In aircraft assembly, robot drilling system has been widely used to enhance the efficiency and quality of assembly holes' manufacturing. However, the industrial robot used by the drilling robot has low absolute positioning accuracy, which cannot meet the hole positioning accuracy requirement of aircraft components, so the drilling robot must be calibrated. The kinematic calibration effect is sensitive to the selection of measurement configurations. Although different observability indexes have been proposed to evaluate the measurement configurations, it is difficult to obtain precise kinematic parameters and minimize the uncertainty of end-effector positions simultaneously with a single observability index. At the same time, current measurement configurations optimization algorithms are still prone to fall into local optimization trap and boundary optimization trap. In order to improve the calibration accuracy of the drilling robot, an improved multi-objective particle swarm optimization algorithm was proposed to search the measurement configurations in the limited workspace, and the “rebound” particle was proposed to avoid local convergence. The effectiveness of the proposed algorithm was verified by simulations and calibration experiments of a drilling robot with KUKA KR500L340-2. Results show that the proposed algorithm can effectively obtain the measurement configurations with the maximum comprehensive observability index, and the positions of the measurement configurations could avoid the search boundary. Meanwhile, more accurate kinematic parameters of the drilling robot can be calculated by using the optimal measurement configurations searched by the proposed algorithm, and the end-effector position accuracy is 26.94% higher than that calibrated with randomly selected measurement configurations.

在飞机装配中，机器人钻孔系统已被广泛应用于提高装配孔制造的效率和质量。然而，钻孔机器人所使用的工业机器人绝对定位精度较低，无法满足飞机部件的孔定位精度要求，因此必须对钻孔机器人进行标定。运动学校准效果对测量配置的选择很敏感。尽管已经提出了不同的可观测性指标来评估测量配置，但很难通过单一的可观测性指标同时获得精确的运动学参数和最小化末端执行器位置的不确定性。同时，目前的测量配置优化算法仍然容易陷入局部优化陷阱和边界优化陷阱。为了提高钻井机器人的标定精度，提出了一种改进的多目标粒子群优化算法，在有限的工作空间内搜索测量配置，并提出了“反弹”粒子避免局部收敛。通过使用KUKA KR500L340-2钻井机器人的仿真和标定实验验证了所提出算法的有效性。结果表明，该算法能够有效地获得综合可观测性指数最大的测量构型，并且测量构型的位置可以避开搜索边界。同时，