|
|
Foundations of Robotics
Seminar, April 22, 2009
Time
and Place | Seminar Abstract
Manipulation Planning with Task-space Regions
Dmitry Berenson
Carnegie Mellon University - Robotics Institute
NSH 1507
Talk 4:30 pm
This talk will present an approach to manipulation planning that uses Task-space Regions (TSRs) to specify goal end-effector poses. Instead of specifying a discrete set of goals in the manipulator's configuration space, we specify goals more intuitively as volumes in the task space (the space of poses of the robot's end-effector). We show that TSRs provide a common framework for describing goal regions that are useful for grasping and manipulation. We also describe two randomized planning algorithms capable of planning with TSRs. The first is an extension of RRT-JT that interleaves exploration using a Rapidly-exploring Random Tree (RRT) with exploitation using Jacobian-based gradient descent toward TSR samples. The second is the IKBiRRT algorithm, which uses a forward-searching tree rooted at the start and a backward-searching tree that is seeded by TSR samples.
We demonstrate both simulation and experimental results for a 7DOF WAM arm with a mobile base performing reaching and pick-and-place tasks. Our results show that planning with TSRs provides an intuitive and powerful method of specifying goals for a variety of tasks without sacrificing efficiency or desirable completeness properties.
The Robotics Institute is part of the School of Computer Science, Carnegie Mellon University.