Wang Zhengtuo, Xu Yuetong, Guanhua Xu, Fu Jianzhong, Yu Jiongyan, Gu Tianyi
{"title":"机器人抓取点云仿真与深度学习","authors":"Wang Zhengtuo, Xu Yuetong, Guanhua Xu, Fu Jianzhong, Yu Jiongyan, Gu Tianyi","doi":"10.1108/AA-07-2020-0096","DOIUrl":null,"url":null,"abstract":"\nPurpose\nIn this work, the authors aim to provide a set of convenient methods for generating training data, and then develop a deep learning method based on point clouds to estimate the pose of target for robot grasping.\n\n\nDesign/methodology/approach\nThis work presents a deep learning method PointSimGrasp on point clouds for robot grasping. In PointSimGrasp, a point cloud emulator is introduced to generate training data and a pose estimation algorithm, which, based on deep learning, is designed. After trained with the emulation data set, the pose estimation algorithm could estimate the pose of target.\n\n\nFindings\nIn experiment part, an experimental platform is built, which contains a six-axis industrial robot, a binocular structured-light sensor and a base platform with adjustable inclination. A data set that contains three subsets is set up on the experimental platform. After trained with the emulation data set, the PointSimGrasp is tested on the experimental data set, and an average translation error of about 2–3 mm and an average rotation error of about 2–5 degrees are obtained.\n\n\nOriginality/value\nThe contributions are as follows: first, a deep learning method on point clouds is proposed to estimate 6D pose of target; second, a convenient training method for pose estimation algorithm is presented and a point cloud emulator is introduced to generate training data; finally, an experimental platform is built, and the PointSimGrasp is tested on the platform.\n","PeriodicalId":55448,"journal":{"name":"Assembly Automation","volume":null,"pages":null},"PeriodicalIF":1.9000,"publicationDate":"2021-05-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"7","resultStr":"{\"title\":\"Simulation and deep learning on point clouds for robot grasping\",\"authors\":\"Wang Zhengtuo, Xu Yuetong, Guanhua Xu, Fu Jianzhong, Yu Jiongyan, Gu Tianyi\",\"doi\":\"10.1108/AA-07-2020-0096\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"\\nPurpose\\nIn this work, the authors aim to provide a set of convenient methods for generating training data, and then develop a deep learning method based on point clouds to estimate the pose of target for robot grasping.\\n\\n\\nDesign/methodology/approach\\nThis work presents a deep learning method PointSimGrasp on point clouds for robot grasping. In PointSimGrasp, a point cloud emulator is introduced to generate training data and a pose estimation algorithm, which, based on deep learning, is designed. After trained with the emulation data set, the pose estimation algorithm could estimate the pose of target.\\n\\n\\nFindings\\nIn experiment part, an experimental platform is built, which contains a six-axis industrial robot, a binocular structured-light sensor and a base platform with adjustable inclination. A data set that contains three subsets is set up on the experimental platform. After trained with the emulation data set, the PointSimGrasp is tested on the experimental data set, and an average translation error of about 2–3 mm and an average rotation error of about 2–5 degrees are obtained.\\n\\n\\nOriginality/value\\nThe contributions are as follows: first, a deep learning method on point clouds is proposed to estimate 6D pose of target; second, a convenient training method for pose estimation algorithm is presented and a point cloud emulator is introduced to generate training data; finally, an experimental platform is built, and the PointSimGrasp is tested on the platform.\\n\",\"PeriodicalId\":55448,\"journal\":{\"name\":\"Assembly Automation\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":1.9000,\"publicationDate\":\"2021-05-28\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"7\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Assembly Automation\",\"FirstCategoryId\":\"94\",\"ListUrlMain\":\"https://doi.org/10.1108/AA-07-2020-0096\",\"RegionNum\":4,\"RegionCategory\":\"计算机科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q3\",\"JCRName\":\"AUTOMATION & CONTROL SYSTEMS\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Assembly Automation","FirstCategoryId":"94","ListUrlMain":"https://doi.org/10.1108/AA-07-2020-0096","RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"AUTOMATION & CONTROL SYSTEMS","Score":null,"Total":0}
Simulation and deep learning on point clouds for robot grasping
Purpose
In this work, the authors aim to provide a set of convenient methods for generating training data, and then develop a deep learning method based on point clouds to estimate the pose of target for robot grasping.
Design/methodology/approach
This work presents a deep learning method PointSimGrasp on point clouds for robot grasping. In PointSimGrasp, a point cloud emulator is introduced to generate training data and a pose estimation algorithm, which, based on deep learning, is designed. After trained with the emulation data set, the pose estimation algorithm could estimate the pose of target.
Findings
In experiment part, an experimental platform is built, which contains a six-axis industrial robot, a binocular structured-light sensor and a base platform with adjustable inclination. A data set that contains three subsets is set up on the experimental platform. After trained with the emulation data set, the PointSimGrasp is tested on the experimental data set, and an average translation error of about 2–3 mm and an average rotation error of about 2–5 degrees are obtained.
Originality/value
The contributions are as follows: first, a deep learning method on point clouds is proposed to estimate 6D pose of target; second, a convenient training method for pose estimation algorithm is presented and a point cloud emulator is introduced to generate training data; finally, an experimental platform is built, and the PointSimGrasp is tested on the platform.
期刊介绍:
Assembly Automation publishes peer reviewed research articles, technology reviews and specially commissioned case studies. Each issue includes high quality content covering all aspects of assembly technology and automation, and reflecting the most interesting and strategically important research and development activities from around the world. Because of this, readers can stay at the very forefront of industry developments.
All research articles undergo rigorous double-blind peer review, and the journal’s policy of not publishing work that has only been tested in simulation means that only the very best and most practical research articles are included. This ensures that the material that is published has real relevance and value for commercial manufacturing and research organizations.