基于GA-SA算法的机器人几何参数误差辨识  

Error Identification of Robot Geometric Parameters Based on GA-SA Algorithm

作　　者：朱振权 殷宝麟[1] 潘瑞冬 郑春雷 ZHU Zhenquan;YIN Baolin;PAN Ruidong;ZHEN Chunlei(School of Mechanical Engineering,Jiamusi University,Jiamusi Heilongjiang 154000,China)

机构地区：[1]佳木斯大学机械工程学院,黑龙江佳木斯154007

出　　处：《佳木斯大学学报（自然科学版）》2024年第1期59-62,共4页Journal of Jiamusi University：Natural Science Edition

摘　　要：几何参数误差对机器人末端绝对定位精度影响最大,而几何误差参数辨识是一个高维非线性问题,求解困难,所以建立一种简单高效的辨识算法是有必要的,本文提出了遗传模拟退火算法(GA-SA)对机器人几何参数误差辨识。以机器人末端位姿误差最小为目标,采用遗传模拟退火算法辨识机器人几何参数误差,以ABB IRB120为算例迭代1100次,遗传算法在200代陷入局部最优,模拟退火参与后最终适应度为0.0914。误差补偿结果表明:机器人末端位置误差沿X,Y,Z轴方向分别降低了88.05%,81.73%,83.72%,姿态误差分别降低了93.92%,83.64%,83.44%,证明遗传模拟退火算法可以有效辨识机器人几何参数误差,提高误差补偿后的机器人末端位姿精度。The geometric parameter error has the greatest influence on the absolute positioning accuracy of the robot end,and the geometric error parameter identification is a high-dimensional nonlinear problem,which is difficult to solve,so it is necessary to establish a simple and efficient identification algorithm,and this paper proposes the Genetic Simulated Annealing Algorithm(GA-SA)for the identification of the robot geometric parameter error.With the goal of minimizing the robot end position error,the genetic simulated annealing algorithm is used to identify the robot geometric parameter error.1100 iterations of ABB IRB120 are used as an example,and the genetic algorithm falls into the local optimum in 200 generations,and the final fitness after simulated annealing is 0.0914.The results of the error compensation show that the robot end position error along the axial direction is reduced by 88.05%,81.73%,83.72%,and the attitude error is reduced by 93.92%,83.64%,83.44%,respectively,which proves that the genetic simulated annealing algorithm can effectively identify the robot geometric parameter errors and improve the robot end position accuracy after error compensation.

关 键 词：机器人误差模型 参数辨识 遗传模拟退火算法 误差补偿 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]