基于位形优化的机器人基坐标系标定与跟踪

Calibration and Tracking of the Robot Base Frame Based on Configuration Optimization

为了精确标定测量设备和机器人基坐标系之间的坐标变换关系,建立了机器人运动学模型以及待标定关系模型,基于差分进化算法对机器人位形进行优选,提出了一种离散型差分进化算法结合线性最小二乘法的标定方法。实验表明,与同样数目随机姿态下的标定结果相比,优选姿态可以显著降低机器人基坐标系的标定误差。针对基坐标系跟踪问题,测量机器人移动前后基座处多个光靶坐标值,通过基于单位四元数法点集匹配,实现基坐标系的实时跟踪。与新位置下的标定结果对比,该方法的跟踪误差与标定误差相近,从而避免了烦琐的机器人标定过程。

In to accurately calibrate the relationship between measuring equipment frame and the base frame of the robot,the forward kinematics of the robot and the mathematical model of the relationship to be calibrated were established firstly.The optimization of the robot configuration can be executed based on the differential evolution algorithm.A calibration method of the discrete differential evolution algorithm combined with linear least squared method was proposed.Experiments show that,compared with the same number of configurations randomly selected,the calibration error of the optimal group can be significantly reduced.For the tracking problem,the coordinate values of the optical targets fixed on the base were measured before and after the robot moves,and the real time tracking of the base frame was realized via the point set matching based on the unit quaternion method.Compared with the calibration results in the new position,the tracking error is similar to that of the calibration error,thus the tedious robot calibration process can be avoided.