基于深度神经网络的多指灵巧手抓取手势优化

doi:10.13973/j.cnki.robot.210376

摘要
图/表
参考文献
相关文章 (15)

全文: PDF (775 KB) HTML ( ) 输出: BibTeX \| EndNote (RIS)
引用本文:
何浩源, 尚伟伟, 张飞, 丛爽. 基于深度神经网络的多指灵巧手抓取手势优化[J]. 机器人, 2023, 45(1): 38-47.DOI: 10.13973/j.cnki.robot.210376. HE Haoyuan, SHANG Weiwei, ZHANG Fei, CONG Shuang. Grasping Gesture Optimization of Multi-fingered Dexterous Hands Based on Deep Neural Networks. ROBOT, 2023, 45(1): 38-47. DOI: 10.13973/j.cnki.robot.210376.

摘要基于深度神经网络模型,提出了一种适用于多指灵巧手的抓取手势优化方法。首先,在仿真环境下构建了一个抓取数据集,并在此基础上训练了一个卷积神经网络,依据目标物体单目视觉信息和多指灵巧手抓取位形来预测抓取质量函数,由此可以将多指灵巧手的抓取规划问题转化为使抓取质量最大化的优化问题,进一步,基于深度神经网络中的反向传播和梯度上升算法实现多指灵巧手抓取手势的迭代与优化。在仿真环境中,比较该网络和仿真平台对同一抓取位形的抓取质量评估结果,再利用所提出的优化方法对随机搜索到的初始手势进行优化,比较优化前后手势的力封闭指标。最后,在实际机器人平台上验证本文方法的优化效果,结果表明,本文方法对未知物体的抓取成功率在80%以上,对于失败的抓取,优化后成功的比例达到90%。

	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
	何浩源
	尚伟伟
	张飞
	丛爽

关键词 ：多指灵巧手, 卷积神经网络, 抓取规划, 未知物体

Abstract：Based on deep neural network model, a grasping gesture optimization method is proposed for multi-fingered dexterous hands. Firstly, a grasp dataset is constructed in simulation environment, and then a convolutional neural network is trained on this basis to predict the grasp quality function from the monocular visual information of the target object and the grasp configuration of the multi-fingered dexterous hand. Therefore, the grasp planning problem of the multi-fingered dexterous hands is transformed into an optimization problem about maximizing the grasping quality. Further the backpropagation and gradient ascent algorithm in deep learning is used to iterate and optimize the grasping gestures of the multi-fingered dexterous hands. In simulation, the evaluation results of the grasping quality, separately computed by the proposed network and the simulation platform for the same grasp configuration, are compared. Then the proposed method is implemented to optimize the initial gestures searched randomly, and the force closure metrics of the gestures before and after optimization are compared. Finally, the optimization performance of the proposed method is validated on the actual robot platform. The results show that the grasping success rate of the proposed method for the unknown objects is more than 80%, and for the failed grasps, the success rate after optimization reaches 90%.

Key words： multi-fingered dexterous hand convolutional neural network grasp planning unknown object

收稿日期: 2021-08-24 录用日期: 2021-12-01

ZTFLH:

TP391

基金资助:国家自然科学基金(62173316,51675501);安徽省杰出青年科学基金(2108085J32).

通讯作者: 尚伟伟, wwshang@ustc.edu.cn E-mail: wwshang@ustc.edu.cn

作者简介: 何浩源(1997-),女,硕士生。研究领域:机器人抓取,深度学习等。
尚伟伟(1981-),男,博士,副教授。研究领域:并联机器人,绳索机器人,工业机器人,仿人机器人,机器人学习控制等。

链接本文:

https://robot.sia.cn/CN/10.13973/j.cnki.robot.210376 或 https://robot.sia.cn/CN/Y2023/V45/I1/38

[1] Bohg J, Morales A, Asfour T, et al. Data-driven grasp synthesis-A survey[J]. IEEE Transactions on Robotics, 2014, 30(2):289-309.
[2] Kirkpatrick D G, Mishra B, Yap C K. Quantitative Steinitz's theorems with applications to multifingered grasping[C]//TwentySecond Annual ACM Symposium on Theory of Computing. New York, USA:ACM, 1990:341-351.
[3] Ferrari C, Canny J. Planning optimal grasps[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA:IEEE, 1992:2290-2295.
[4] Ciocarlie M T, Allen P K. Hand posture subspaces for dexterous robotic grasping[J]. International Journal of Robotics Research, 2009, 28(7):851-867.
[5] Sung J, Jin S H, Saxena A. Robobarista:Object part based transfer of manipulation trajectories from crowd-sourcing in 3D pointclouds[C]//International Symposium on Robotics Research. Cham, Switzerland:Springer, 2015:701-720.
[6] Romero J. Human-to-robot mapping of grasps[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems, WS on Grasp and Task Learning by Imitation. 2008:9-15.
[7] Levine S, Pastor P, Krizhevsky A, et al. Learning hand-eye coordination for robotic grasping with deep learning and large-scale data collection[J]. International Journal of Robotics Research, 2018, 37(4-5):421-436.
[8] Goldfeder C, Ciocarlie M, Peretzman J, et al. Data-driven grasping with partial sensor data[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2009:1278-1283.
[9] Lowe D G. Object recognition from local scale-invariant features[C]//Seventh IEEE International Conference on Computer Vision. Piscataway, USA:IEEE, 1999:1150-1157.
[10] Goldfeder C, Ciocarlie M, Dang H, et al. The Columbia grasp database[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA:IEEE, 2009:1710-1716.
[11] Jiang Y, Moseson S, Saxena A. Efficient grasping from RGBD images:Learning using a new rectangle representation[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA:IEEE, 2011:3304-3311.
[12] Lenz I, Lee H, Saxena A. Deep learning for detecting robotic grasps[J]. International Journal of Robotics Research, 2015, 34(4-5):705-724.
[13] Redmon J, Angelova A. Real-time grasp detection using convolutional neural networks[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA:IEEE, 2015:1316-1322.
[14] Guo D, Sun F C, Liu H P, et al. A hybrid deep architecture for robotic grasp detection[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA:IEEE, 2017:1609-1614.
[15] 夏晶,钱堃,马旭东,等.基于级联卷积神经网络的机器人平面抓取位姿快速检测[J].机器人, 2018, 40(6):794-802. Xia J, Qian K, Ma X D, et al. Fast planar grasp pose detection for robot based on cascaded deep convolutional neural networks[J]. Robot, 2018, 40(6):794-802.
[16] 喻群超,尚伟伟,张驰.基于三级卷积神经网络的物体抓取检测[J].机器人, 2018, 40(5):762-768. Yu Q C, Shang W W, Zhang C. Object grab detecting based on three-level convolution neural network[J]. Robot, 2018, 40(5):762-768.
[17] Zhou Y L, Kris H. 6DOF grasp planning by optimizing a deep learning scoring function[C]//Robotics:Science and Systems (RSS) Workshop on Revisiting Contact Turning a Problem into a Solution. 2017.
[18] Varley J, Weisz J, Weiss J, et al. Generating multi-fingered robotic grasps via deep learning[C]//IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway, USA:IEEE, 2015:4415-4420.
[19] Kappler D, Bohg J, Schaal S. Leveraging big data for grasp planning[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA:IEEE, 2015:4304-4311.
[20] Song F J, Zhao Z Z, Ge W, et al. Learning optimal grasping posture of multi-fingered dexterous hands for unknown objects[C]//IEEE International Conference on Robotics and Biomimetics. Piscataway, USA:IEEE, 2018:2310-2315.
[21] Zhao Z Z, Shang W W, He H Y, et al. Grasp prediction and evaluation of multi-fingered dexterous hands using deep learning[J]. Robotics and Autonomous Systems, 2020, 129. DOI:10.1016/j.robot.2020.103550.
[22] Lu Q K, Chenna K, Sundaralingam B, et al. Planning multifingered grasps as probabilistic inference in a learned deep network[M]//Springer Proceedings in Advanced Robotics, Vol.10. Cham, Switzerland:Springer, 2020:455-472.
[23] Miller A T, Allen P K. GraspIt! A versatile simulator for robotic grasping[J]. IEEE Robotics & Automation Magazine, 2004, 11(4):110-122.
[24] Cornell University. Cornell grasping dataset[DB/OL]. (2015-11-31)[2016-09-01]. http://pr.cs.cornell.edu/grasping/rectdata/data.php.
[25] Mahler J, Pokorny F T, Hou B, et al. Dex-Net 1.0:A cloudbased network of 3D objects for robust grasp planning using a multi-armed bandit model with correlated rewards[C]//IEEE International Conference on Robotics and Automation. Piscataway, USA:IEEE, 2016:1957-1964.