ARobotPathPlanningFramework.PDFVIP

A Robot Path Planning Framework that Learns from Experience Dmitry Berenson Pieter Abbeel Ken Goldberg University of California, Berkeley, Berkeley, CA, USA {berenson, pabbeel}@eecs.berkeley.edu, goldberg@berkeley.edu Abstract— We propose a framework, called Lightning, for planning paths in high-dimensional spaces that is able to learn from experience, with the aim of reducing computation time. This framework is intended for manipulation tasks that arise in applications ranging from domestic assistance to robot-assisted surgery. Our framework consists of two main modules, which run in parallel: a planning-from-scratch module, and a module that retrieves and repairs paths stored in a path library. After a path is generated for a new query, a library manager decides whether to store the path based on computation time and the generated path’s similarity to the retrieved path. To retrieve an appropriate path from the library we use two heuristics that exploit two key aspects of the problem: (i) A correlation between the amount a path violates constraints and the amount Fig. 1. Diagram of the Lightning framework. of time needed to repair that path, and (ii) the implicit division of constraints into those that vary across environments in which approaches to create a framework that is both general and the robot operates and those that do not. We evaluated an implementation of the framework on several ef?cient. It is also straightforward to enhance the framework tasks for the PR2 mobile manipulator and a minimally-invasive with parallelization and cloud computing by parallelizing the surgery robot in simulation. We found that the retrieve-and- retrieval and repair algorithms and by storing the experience repair module produced paths faster than planning-from- of multiple robots in the cloud. scratch in over 90% of test cases for the PR2 and in 58%