为了提高机器人的绝对定位精度，建立了机器人绝对定位误差模型并进行了补偿方法研究． 将定位误差分为几何参数误差与柔度误差，分别建立相应的误差模型． 几何
参数误差研究以MD-H( 修正型D-H) 运动模型为基础，对柔度误差的影响进行了解耦，并考虑了机器人基坐标系与测量坐标系的转换误差，提出了基于相对位置的几何参数误差模型． 柔度误差研究针对机器人的构造特点，建立了针对关节2 和3 的误差模型，简化了计算模型． 最后基于所建立的两种误差模型，提出了误差补偿方法，并采用该方法对机器人进行了实际补偿实验． 结果表明，平均绝对定位精度由补偿前的1． 173 mm 降至补偿后的0． 158mm，说明文中方法可有效提高机器人的绝对定位精度，扩展机器人的应用范围．

In order to improve the absolute positioning accuracy of robots,an absolute positioning error model is constructed and an error compensation method is investigated.First,positioning errors are divided into geometrical parameter errors and compliance errors,and two models are respectively constructed for the two kinds of errors.The geometrical parameter errors are investigated on the basis of modified Denavit-Hartenberg kinematics,and the impact of the compliance errors is decoupled.Then,by taking into consideration the conversion error between the base coordinate system and the measurement one of robots,a geometrical parameter error model on the basis of relative positions is constructed.For the compliance errors,a model for joint 2 and joint 3 is constructed according to the construction features of robots,thus simplifying the calculation model of the compliance errors greatly.Finally,on the basis of the two kinds of error models,the error compensation method is proposed.The results of error compensation experiments show that average absolute positioning accuracy decreases from 1.173 to 0.158mm after calibration,which means that the proposed method can improve the absolute positioning accuracy of industrial robots effectively and extend its range of applications greatly.