{"title":"定向没有传感器的可变形多边形部件","authors":"Shawn M. Kristek, Dylan A. Shell","doi":"10.1109/IROS.2012.6386165","DOIUrl":null,"url":null,"abstract":"Sensorless part orienting has proven useful in manufacturing and automation, while the manipulation of deformable objects is an area of growing interest. Existing sensorless orienting techniques may produce forces which have the potential to damage deformable parts. We present an algorithm that, when provided a geometric description of the part and a deformation model, generates a plan to orient the part up to symmetry from any initial orientation. The solution exploits deformation of the object under certain configurations to help resolve ambiguity. The approach has several attractive features: (1) the resulting plan is a short sequence of such actions guaranteed to succeed for all initial configurations; (2) the algorithm operates even with a very simple model of deformation, but is extensible when specialized knowledge is available; (3) failure to find a feasible solution has precise semantics (e.g., inadequate manipulator precision). We validate the algorithm experimentally with a pair of low-precision robot manipulators, orienting 6 parts made of 4 types of materials, with the correct orientation being reached on 80% of the 192 trials. Careful analysis of the failures emphasizes the importance of low-friction conditions, that increased manipulator precision would be beneficial but is not necessary, and a simple deformation model can suffice. In addition to illustrating the feasibility of sensorless manipulation of deformable parts, we note that the algorithm has applications to manipulation of non-deformable parts without the pressure switch sensor employed in existing sensorless orienting strategies.","PeriodicalId":6358,"journal":{"name":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","volume":"19 1","pages":"973-979"},"PeriodicalIF":0.0000,"publicationDate":"2012-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"21","resultStr":"{\"title\":\"Orienting deformable polygonal parts without sensors\",\"authors\":\"Shawn M. Kristek, Dylan A. Shell\",\"doi\":\"10.1109/IROS.2012.6386165\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Sensorless part orienting has proven useful in manufacturing and automation, while the manipulation of deformable objects is an area of growing interest. Existing sensorless orienting techniques may produce forces which have the potential to damage deformable parts. We present an algorithm that, when provided a geometric description of the part and a deformation model, generates a plan to orient the part up to symmetry from any initial orientation. The solution exploits deformation of the object under certain configurations to help resolve ambiguity. The approach has several attractive features: (1) the resulting plan is a short sequence of such actions guaranteed to succeed for all initial configurations; (2) the algorithm operates even with a very simple model of deformation, but is extensible when specialized knowledge is available; (3) failure to find a feasible solution has precise semantics (e.g., inadequate manipulator precision). We validate the algorithm experimentally with a pair of low-precision robot manipulators, orienting 6 parts made of 4 types of materials, with the correct orientation being reached on 80% of the 192 trials. Careful analysis of the failures emphasizes the importance of low-friction conditions, that increased manipulator precision would be beneficial but is not necessary, and a simple deformation model can suffice. In addition to illustrating the feasibility of sensorless manipulation of deformable parts, we note that the algorithm has applications to manipulation of non-deformable parts without the pressure switch sensor employed in existing sensorless orienting strategies.\",\"PeriodicalId\":6358,\"journal\":{\"name\":\"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems\",\"volume\":\"19 1\",\"pages\":\"973-979\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2012-12-24\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"21\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/IROS.2012.6386165\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"2012 IEEE/RSJ International Conference on Intelligent Robots and Systems","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IROS.2012.6386165","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Orienting deformable polygonal parts without sensors
Sensorless part orienting has proven useful in manufacturing and automation, while the manipulation of deformable objects is an area of growing interest. Existing sensorless orienting techniques may produce forces which have the potential to damage deformable parts. We present an algorithm that, when provided a geometric description of the part and a deformation model, generates a plan to orient the part up to symmetry from any initial orientation. The solution exploits deformation of the object under certain configurations to help resolve ambiguity. The approach has several attractive features: (1) the resulting plan is a short sequence of such actions guaranteed to succeed for all initial configurations; (2) the algorithm operates even with a very simple model of deformation, but is extensible when specialized knowledge is available; (3) failure to find a feasible solution has precise semantics (e.g., inadequate manipulator precision). We validate the algorithm experimentally with a pair of low-precision robot manipulators, orienting 6 parts made of 4 types of materials, with the correct orientation being reached on 80% of the 192 trials. Careful analysis of the failures emphasizes the importance of low-friction conditions, that increased manipulator precision would be beneficial but is not necessary, and a simple deformation model can suffice. In addition to illustrating the feasibility of sensorless manipulation of deformable parts, we note that the algorithm has applications to manipulation of non-deformable parts without the pressure switch sensor employed in existing sensorless orienting strategies.