The International Journal of Robotics Research Study, Ahead of Publish.
Comprehending the geometry of collision-free arrangement room (C-free) in the existence of Cartesian-space barriers is an important component for collision-free movement preparation. While it is feasible to look for crashes at a factor making use of typical formulas, to day no sensible technique exists for calculating C-free areas with extensive certifications as a result of the intricacy of mapping Cartesian-space barriers via the kinematics. In this job, we provide the initial to our understanding extensive technique for about breaking down a logical parametrization of C-free right into licensed polyhedral areas. Our technique, called C-Iris (C-space Iterative Regional Rising cost of living by Semidefinite programs), produces huge, raised polytopes in a logical parameterization of the arrangement room which are carefully licensed to be collision-free. Such areas have actually been revealed to be valuable for both optimization-based and randomized movement preparation. Based upon convex optimization, our technique operates in approximate measurements, just makes presumptions concerning the convexity of the barriers in the 3D Cartesian room, and is quickly adequate to range to reasonable troubles in control. We show our formula’s capacity to load a non-trivial quantity of collision-free C-space in a number of 2-DOF instances where the C-space can be pictured, along with the scalability of our formula on a 7-DOF KUKA iiwa, a 6-DOF UR3e, and 12-DOF bimanual manipulators. An application of our formula is open-sourced in Drake. We in addition offer instances of our formula in interactive Python note pads.
发布者：Hongkai Dai，转转请注明出处：https://robotalks.cn/certified-polyhedral-decompositions-of-collision-free-configuration-space/
