基于模仿学习的双曲率曲面零件复合材料织物机器人铺放

doi:10.13973/j.cnki.robot.210475

摘要
图/表
参考文献
相关文章 (2)

全文: PDF (786 KB) HTML ( ) 输出: BibTeX \| EndNote (RIS)
引用本文:
段宝阁, 杨尚尚, 谢啸, 肖晓晖. 基于模仿学习的双曲率曲面零件复合材料织物机器人铺放[J]. 机器人, 2022, 44(4): 504-512.DOI: 10.13973/j.cnki.robot.210475. DUAN Baoge, YANG Shangshang, XIE Xiao, XIAO Xiaohui. Robotic Layup of Woven Composite on Double Curvature Surface Parts Based on Imitation Learning. ROBOT, 2022, 44(4): 504-512. DOI: 10.13973/j.cnki.robot.210475.

摘要针对双曲率曲面零件的复合材料织物铺放，手工铺放效率低、质量均一性差，机器人铺放的相关研究未能准确地描述手工铺放技能。为此，本文提出基于模仿学习的铺放技能采集、描述与重现的相关方法。首先，利用拖动示教获取织物铺放的轨迹信息，以压力阈值为分割依据进行有监督的轨迹分割，再采用高斯混合模型（Gaussian mixture model，GMM）与高斯混合回归（Gaussian mixture regression，GMR）学习得到铺放轨迹概率模型。在此基础上，以示教姿态和曲面法向为参考，计算末端执行器的法向姿态，再次进行示教获得法向压力信息，由GMM/GMR学习得到铺放压力概率模型；然后，针对接触状态建立过程，在力／位混合控制框架下，提出预控制量退出机制，实现接触状态建立过程中无冲击的控制器切换；最后，基于UR5e机械臂、ATI六维力／力矩传感器和ROS（robot operating system）搭建试验平台，进行试验。结果表明：所提出的技能采集、描述方法可实现对人工铺放经验的准确提取与表达，采用预控制量退出机制，接触状态建立过程的最大压力误差为1.8 N，小于曲面模具所有16条轨迹的铺放压力均方根误差2.0 N，完成的铺层无肉眼可见的鼓包与褶皱等缺陷，深度相机测量织物点云距模具点云距离小于1.0 mm的点约占97.3%。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	段宝阁
	杨尚尚
	谢啸
	肖晓晖

关键词 ：复合材料织物, 双曲率曲面, 模仿学习, 概率模型, 接触任务, 状态切换

Abstract：In layup of woven composite on double curvature surface, the manual layup is of low efficiency and poor quality uniformity, while research findings on robotic layup are not accurate enough to describe manual layup skills. Therefore, a method of acquisition, representation and recurrence of woven composite layup skills is proposed based on imitation learning. Firstly, the dragging demonstration method is used to obtain trajectory information of woven layup, and supervised trajectory segmentation is carried out based on pressure threshold. After that, probabilistic models of layup trajectory are learned by Gaussian mixture model (GMM) and Gaussian mixture regression (GMR). On this basis, the normal orientation of the endeffector is calculated taking demonstration orientation and surface norm as reference. Then, another demonstration is carried out to obtain normal pressure information, and probabilistic models of layup pressure are learned though GMM/GMR. For the establishment process of contact status, a pre-control quantity exit mechanism is proposed within the framework of hybrid force/motion control to accomplish non-impact switching of the controllers during the establishment of contact status. Finally, an experiment platform is built based on UR5e robot, ATI six-dimensional force/torque sensor and ROS (robot operating system). The result shows that the proposed skills acquisition and description method can realize accurate extraction and expression of manual layup experience, and the maximum pressure error during contact status establishment is 1.8 N when using the proposed pre-control quantity exit mechanism, which is less than 2.0 N, the root mean square error of layup pressure of all 16 trajectories on the curvature surface mould. The finished layer has no visible defects such as bulge and folding. At about 97.3% of points in the woven point cloud measured by depth camera, their distances from the mould point cloud are less than 1.0 mm.

Key words： woven composite double curvature surface imitation learning probabilistic model contact task status switching

收稿日期: 2021-10-29 录用日期: 2022-02-14

TP242.6

基金资助:国家自然科学基金（51705371）.

通讯作者: 肖晓晖，xhxiao@whu.edu.cn E-mail: xhxiao@whu.edu.cn

作者简介: 段宝阁（1997-），男，硕士生。研究领域：复合材料自动铺放，机械臂运动规划与控制。
杨尚尚（1990-），男，博士生。研究领域：复合材料自动铺放，机械臂运动规划与控制。
谢啸（1998-），男，硕士生。研究领域：复合材料自动铺放，机械臂运动规划与控制。

链接本文:

https://robot.sia.cn/CN/10.13973/j.cnki.robot.210475 或 https://robot.sia.cn/CN/Y2022/V44/I4/504

[1] 官威,李文晓,戴瑛,等.纺织复合材料预制体变形研究综述[J].航空制造技术, 2021, 64(Z1):22-37.Guan W, Li W X, Dai Y, et al.A review of study on deformation of textile composite preforms[J].Aeronautical Manufacturing Technology, 2021, 64(Z1):22-37.
[2] Elkington M, Bloom L D, Ward C, et al.On prepreg properties and manufacturability[C]//19th International Conference on Composite Materials.Montreal, Canada:Concordia Centre for Composites, 2013:4397-4409.
[3] Ward C, Hazra K, Potter K.Development of the manufacture of complex composite panels[J].International Journal of Materials and Product Technology, 2011, 42(3-4):131-155.
[4] Such M, Ward C, Hutabarat W, et al.Intelligent composite layup by the application of low cost tracking and projection technologies[J].Procedia CIRP, 2014, 25:122-131.
[5] Elkington M P, Bloom D, Ward C, et at.Understanding the lamination process[C]//19th International Conference on Composite Materials.Montreal, Canada:Concordia Centre for Composites, 2013:4385-4396.
[6] Prabhu V A, Elkington M, Crowley D, et al.Digitisation of manual composite layup task knowledge using gaming technology[J].Composites, Part B:Engineering, 2017, 112:314-326.
[7] Crowley D, Elkington M, Ward C, et al.Hand lay-up of complex geometries-Prediction, capture and feedback[C]//International SAMPE Technical Conference.Diamond Bar, USA:Society for the Advancement of Material and Process Engineering, 2016.
[8] Malhan R K, Kabir A M, Shembekar A V, et al.Hybrid cells for multi-layer prepreg composite sheet layup[C]//IEEE 14th International Conference on Automation Science and Engineering.Piscataway, USA:IEEE, 2018:1466-1472.
[9] Malhan R K, Shembekar A V, Kabir A M, et al.Automated planning for robotic layup of composite prepreg[J].Robotics and Computer-Integrated Manufacturing, 2021, 67.DOI:10.1016/j.rcim.2020.102020.
[10] 黄艳龙,徐德,谭民.机器人运动轨迹的模仿学习综述[J].自动化学报, 2022, 48(2):315-334.Huang Y L, Xu D, Tan M.On imitation learning of robot movement trajectories:A survey[J].Acta Automatica Sinica, 2022, 48(2):315-334.
[11] Gao X, Ling J, Xiao X H, et al.Learning force-relevant skills from human demonstration[J].Complexity, 2019.DOI:10.1155/2019/5262859.
[12] Wang Y, Beltran-Hernandez C C, Wan W W, et al.Hybrid trajectory and force learning of complex assembly tasks:A combined learning framework[J].IEEE Access, 2021, 9:60175-60186.
[13] Wang N, Chen C Z, di Nuovo A.A framework of hybrid force/motion skills learning for robots[J].IEEE Transactions on Cognitive and Developmental Systems, 2021, 13(1):162-170.
[14] 周欣宇.一种基于轨迹模仿学习和任务规划的机械臂演示学习方法研究[D].哈尔滨:哈尔滨工业大学, 2019.Zhou X Y.A research on robotic-arm learning method by demonstration based on trajectory imitation learning and task planning[D].Harbin:Harbin Institute of Technology, 2019.
[15] 李琳,肖佳栋,张铁,等.基于自适应迭代的机器人曲面恒力跟踪[J].北京航空航天大学学报, 2019, 45(4):641-649.Li L, Xiao J D, Zhang T, et al.Constant-force curved-surfacetracking with robotic manipulator based on adaptive iterative algorithm[J].Journal of Beijing University of Aeronautics and Astronautics, 2019, 45(4):641-649.
[16] 高培阳.基于力传感器的工业机器人恒力磨抛系统研究[D].武汉:华中科技大学, 2019.Gao P Y.Research on force control grinding system of industrial robot based on force sensor[D].Wuhan:Huazhong University of Science and Technology, 2019.
[17] 连学军.面向大型风电叶片的机器人阻抗控制顺应打磨研究[D].武汉:华中科技大学, 2017.Lian X J.The research of robot adaptable grinding large wind blade by impedance control[D].Wuhan:Huazhong University of Science and Technology, 2017.
[18] Yamane K.Admittance control with unknown location of interaction[J].IEEE Robotics and Automation Letters, 2021, 6(2):4079-4086.
[19] 迟明善,姚玉峰,刘亚欣.模仿学习示教轨迹自动分割方法的研究进展[J].控制与决策, 2019, 34(7):1345-1354.Chi M S, Yao Y F, Liu Y X.Recent advances on automatic segmentation method of teaching trajectory for imitation learning[J].Control and Decision, 2019, 34(7):1345-1354.
[20] 李航.统计学习方法[M].北京:清华大学出版社, 2012:162-166.Li H.Statistical learning method[M].Beijing:Tsinghua University Press, 2012:162-166.
[21] Calinon S, Guenter F, Billard A.On learning, representing, and generalizing a task in a humanoid robot[J].IEEE Transactions on Systems, Man, and Cybernetics, Part B:Cybernetics, 2007, 32(2):286-298.
[22] 熊有伦,李文龙,陈文斌,等.机器人学:建模、控制与视觉[M].武汉:华中科技大学出版社, 2018:267-269.Xiong Y L, Chen W L, Chen W B, et al.Robotics:Modeling, control and vision[M].Wuhan:Huazhong University of Science and Technology Press, 2018:267-269.