基于迁移学习的铣削机器人定位误差补偿方法

doi:10.3901/JME.2022.14.170

机械工程学报 ›› 2022, Vol. 58 ›› Issue (14): 170-180.doi: 10.3901/JME.2022.14.170

• 特邀专栏：大型构件视觉测量与机器人加工 • 上一篇下一篇

扫码分享

基于迁移学习的铣削机器人定位误差补偿方法

邓柯楠, 高栋, 马守东, 路勇

哈尔滨工业大学机电工程学院哈尔滨 150000

收稿日期:2021-05-31 修回日期:2021-12-10 出版日期:2022-07-20 发布日期:2022-09-07
通讯作者: 路勇(通信作者)，男，1971年出生，博士，教授，博士研究生导师。主要研究方向为智能刀柄、加工过程监控与大型构件加工。E-mail：luyong@hit.edu.cn
作者简介:邓柯楠，男，1992年出生，博士研究生。主要研究方向为大型构件移动机器人加工理论与方法。E-mail：1110801922@hit.edu.cn;高栋，男，1970年出生，博士，教授，博士研究生导师。主要研究方向超大构件加工方法，重型机床加工误差补偿。
基金资助:
国家重点研发计划资助项目(2018YFB1306800)。

An Efficient Error Compensation Method for Milling Robot Based on Transfer Learning

DENG Kenan, GAO Dong, MA Shoudong, LU Yong

School of Mechatronics Engineering, Harbin Institute of Technology, Harbin 150000

Received:2021-05-31 Revised:2021-12-10 Online:2022-07-20 Published:2022-09-07

摘要/Abstract

摘要： 空间网格补偿法是提高机器人定位误差的有效方法之一，然而由于所需采样位姿多导致误差测量环节非常耗时，为提高机器人定位误差补偿效率，提出了一种机理分析与数据驱动的铣削机器人定位误差补偿方法，基于迁移学习来预测机器人工作空间内不同区域的定位误差。首先建立机器人刚柔耦合误差模型，研究立方体与柱体工作空间内不同区域的误差分布特性；之后，考虑误差区域相似性将机器人工作空间分为源域空间与目标域空间，在源域空间基于分级采样策略将完备的机器人采样位姿及误差测量数据作为源域数据，对于目标域空间只需要将少量的采样位姿及误差数据作为目标域数据，源域数据与目标域数据均用于训练高斯过程回归模型,通过基于加权拟合误差的子空间对齐和自适应权重迭代方法提升迁移学习模型预测精度，根据指定机器人位姿参数预测并补偿机器人定位误差；最后，使用KR160铣削机器人系统进行了误差补偿试验以验证该方法的可行性和有效性，试验结果表明，经过补偿后机器人定位误差1.499 mm降低到0.182 mm，所需机器人采样位姿数目减少了70%，使用铣削机器人加工法兰孔，其轮廓误差和位置误差达到0.269 mm和0.331 mm，该方法可以提高补偿效率和机器人定位精度。

关键词: 定位误差, 迁移学习, 铣削机器人, 补偿效率, 耦合误差模型

Abstract: The grid compensation is one of the effective methods to improve the positioning accuracy of industrial robot. However, error measurement is time-consuming due to the many sampling robot configurations required. To improve positioning error compensation efficiency, proposes a mechanism analysis and data-driven method for milling robot, which predict the positioning error of different regions in the robot workspace based on transfer learning. Firstly, the rigid-flexible coupling error model of the robot is established, and the similarity of the error distribution in different regions of the cube and cylindrical workspace is analyzed. Then, the robot workspace is divided into source domain and target domain by considering error similarity of different areas. Complete robot sampling configurations and errors measurement data are obtained by hierarchical sampling method as source data for the source workspace, and only a few robot configurations are required to measure the positioning errors as the target data for targe regions. The source domain data and target domain data are used to train the Gaussian process regression model (GPR). The prediction accuracy of the transfer learning model is improved by subspace alignment and adaptive weight method based on weighted fitting error. The robot positioning error is predicted and compensated according to the robot positions. Finally, a KR160 milling robot is used for error compensation experiments to verify the feasibility and effectiveness of the proposed method. The experimental results prove that the robot positioning error is reduced from 1.499 mm to 0.182 mm after compensation, and the number of robot sampling poses is reduced by 70%, the contour error and position error of the flange hole are 0.269 mm and 0.331 mm, which proves that the method can improve the compensation efficiency and machining accuracy.

Key words: positioning errors, transfer learning, milling robot, compensation efficiency, coupling error model

中图分类号:

TP242

邓柯楠, 高栋, 马守东, 路勇. 基于迁移学习的铣削机器人定位误差补偿方法[J]. 机械工程学报, 2022, 58(14): 170-180.

DENG Kenan, GAO Dong, MA Shoudong, LU Yong. An Efficient Error Compensation Method for Milling Robot Based on Transfer Learning[J]. Journal of Mechanical Engineering, 2022, 58(14): 170-180.

参考文献

[1] TAO Bo, ZHAO Xingwei, DING Han. Mobile-robotic machining for large complex components:A review study[J]. Science China, 2019, 62(8):1388-1400.
[2] 谢福贵,梅斌,刘辛军,等.一种大型复杂构件加工新模式及新装备探讨[J].机械工程学报, 2020, 56(19):70-78. XIE Fugui, MEI Bin, LIU Xinjun, et al. Novel mode and equipment for machining large complex components[J]. Journal of Mechanical Engineering, 2020, 56(19):70-78.
[3] VERL A, VALENTE A, MELKOTE S, et al. Robots in machining[J]. CIRP Annals, 2019, 68(2):799-822.
[4] 曾远帆,田威,廖文和.面向飞机自动钻铆系统的工业机器人精度补偿技术[J].航空制造技术, 2016, 59(18):46-52. ZENG Yuanfan, TIAN Wei, LIAO Wenhe. Industrial robot error compensation methods for aircraft automatic drilling and riveting system[J]. Aeronautical Manufacturing Technology, 2016, 59(18):46-52.
[5] 史晓佳,张福民,曲兴华,等. KUKA工业机器人位姿测量与在线误差补偿[J].机械工程学报, 2017, 53(8):1-7. SHI Xiaojia, ZHANG Fumin, QU Xinghua, et al. Position and attitude measurement and online errors compensation for KUKA industrial robots[J]. Journal of Mechanical Engineering, 2017, 53(8):1-7.
[6] MEHLENHOFF T, BLOEDORN C. Solution for automated drilling in composite structures with a standard articulating robot system[R]. SAE Technical Paper, 2010.
[7] WANG S M, LIU Y L, KANG Y. An efficient error compensation system for CNC multi-axis machines[J]. International Journal of Machine Tools and Manufacture, 2002, 42(11):1235-1245.
[8] 文科,张加波,乐毅,等.数控驱动的移动铣削机器人精度提升方法[J].机械工程学报, 2021, 57(5):72-80. WEN Ke, ZHANG Jiabo, YUE Yi, et al. Method for improving accuracy of NC-driven mobile milling robot[J]. Journal of Mechanical Engineering, 2021, 57(5):72-80.
[9] ZENG Y, TIAN W, LIAO W. Positional error similarity analysis for error compensation of industrial robots[J]. Robotics and Computer-Integrated Manufacturing, 2016, 42:113-120.
[10] 周炜,廖文和,田威.基于空间插值的工业机器人精度补偿方法理论与试验[J].机械工程学报, 2013, 49(3):42-48. ZHOU Wei, LIAO Wenhe, TIAN Wei. Theory and experiment of industrial robot accuracy compensation method based on spatial interpolation[J]. Journal of Mechanical Engineering, 2013, 49(3):42-48.
[11] NGUYEN H N, ZHOU J, KANG H J. A calibration method for enhancing robot accuracy through integration of an extended Kalman filter algorithm and an artificial neural network[J]. Neurocomputing, 2015, 151:996-1005.
[12] LI B, TIAN W, ZHANG C, et al. Positioning error compensation of an industrial robot using neural networks and experimental study[J]. Chinese Journal of Aeronautics, 2022, 35(2):346-360.
[13] ZHAO G, ZHANG P, MA G, et al. System identification of the nonlinear residual errors of an industrial robot using massive measurements[J]. Robotics and ComputerIntegrated Manufacturing, 2019, 59:104-114.
[14] ZHAO D, BI Y, KE Y. An efficient error compensation method for coordinated CNC five-axis machine tools[J]. International Journal of Machine Tools and Manufacture, 2017, 123:105-115.
[15] ZHU W, LI G, DONG H, et al. Positioning error compensation on two-dimensional manifold for robotic machining[J]. Robotics and Computer-Integrated Manufacturing, 2019, 59:394-405.
[16] MEI B, XIE F, LIU X J, et al. Elasto-geometrical error modeling and compensation of a five-axis parallel machining robot[J]. Precision Engineering, 2021, 69:48-61.
[17] LIU X, LI Y, CHEN G. Multimode tool tip dynamics prediction based on transfer learning[J]. Robotics and Computer-Integrated Manufacturing, 2019, 57:146-154.
[18] CHEN G, LI Y, LIU X. Pose-dependent tool tip dynamics prediction using transfer learning[J]. International Journal of Machine Tools and Manufacture, 2019, 137:30-41.
[19] PARDOE D, STONE P. Boosting for regression transfer[C/CD]//Proceedings of the 27th International Conference on Machine Learning (ICML-10), June 21-24, 2010, Haifa, Israel.
[20] 乔贵方,孙大林,宋光明,等.串联机器人标定系统的坐标系快速转换方法[J].机械工程学报, 2020, 56(14):1-8. QIAO Guifang, SUN Dalin, SONG Guangming, et al. A rapid coordinate transformation method for serial robot calibration system[J]. Journal of Mechanical Engineering 2020, 56(14):1-8.
[21] KENAN D, DONG G, SHOUDONG M, et al. Contouring errors and feedrate fluctuation of serial industrial robot in complex toolpath with different controller[C]//International Conference on Intelligent Robotics and Applications. Springer, Cham, 2021:100-108.

基于迁移学习的铣削机器人定位误差补偿方法

An Efficient Error Compensation Method for Milling Robot Based on Transfer Learning

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	邓聪颖, 邓子豪, 林丽君, 陈翔, 马莹, 禄盛. 基于迁移学习的变刀具-刀柄夹持状态下数控铣削稳定性预测研究[J]. 机械工程学报, 2024, 60(3): 296-304.
[2]	朱永超, 朱才朝, 谭建军, 冉峯, 宋朝省. 考虑数据特征差异的风电齿轮箱群组健康状态预测研究[J]. 机械工程学报, 2024, 60(18): 64-75.
[3]	李响, 付春霖, 雷亚国, 李乃鹏, 杨彬. 保证数据隐私的装备协同智能故障诊断联邦迁移学习方法[J]. 机械工程学报, 2023, 59(6): 1-9.
[4]	刘伟, 刘顺, 邓朝晖, 葛吉民. 工业机器人定位误差补偿技术研究进展[J]. 机械工程学报, 2023, 59(17): 1-16.
[5]	刘佳宁, 曹宏瑞, 闫鹏辉, 姬汶辰. 变工况下磨削振纹的深度迁移智能辨识方法[J]. 机械工程学报, 2023, 59(16): 128-136.
[6]	吴锦辉, 张德权, 韩旭. 工业机器人参数容差稳健性设计[J]. 机械工程学报, 2023, 59(11): 147-158.
[7]	王文东, 李杰, 张俊博, 秦雷, 袁小庆. 基于迁移学习的手部离散动作识别方法研究[J]. 机械工程学报, 2022, 58(7): 12-19.
[8]	宰红斌, 刘建国, 张文岗, 封士永, 祖国强. 基于深度迁移学习的配电线路绝缘子状态监测方法^*[J]. 电气工程学报, 2022, 17(4): 275-281.
[9]	杨淞元, 田勇, 田劲东. 基于iCEEMDAN和迁移学习的锂离子电池SOH估计^*[J]. 电气工程学报, 2022, 17(4): 2-10.
[10]	鲍劲松, 张荣, 李婕, 陆玉前, 彭涛. 面向人-机-环境共融的数字孪生协同技术[J]. 机械工程学报, 2022, 58(18): 103-115.
[11]	雷亚国, 杨彬, 李乃鹏, 李响, 武通海. 跨设备的机械故障靶向迁移诊断方法[J]. 机械工程学报, 2022, 58(12): 1-9.
[12]	文科, 张加波, 乐毅, 周莹皓, 杨继之, 陈钦韬. 数控驱动的移动铣削机器人精度提升方法[J]. 机械工程学报, 2021, 57(5): 72-80.
[13]	陈祝云, 钟琪, 黄如意, 廖奕校, 李霁蒲, 李巍华. 基于增强迁移卷积神经网络的机械智能故障诊断[J]. 机械工程学报, 2021, 57(21): 96-105.
[14]	赵树煊, 张洁, 汪俊亮, 徐楚桥. 基于两阶段深度迁移学习的面料疵点检测算法[J]. 机械工程学报, 2021, 57(17): 86-97.
[15]	雷亚国, 杨彬, 杜兆钧, 吕娜. 大数据下机械装备故障的深度迁移诊断方法[J]. 机械工程学报, 2019, 55(7): 1-8.