DocumentCode :
1086552
Title :
Improving Motion-Planning Algorithms by Efficient Nearest-Neighbor Searching
Author :
Yershova, Anna ; LaValle, Steven M.
Author_Institution :
Dept. of Comput. Sci., Illinois Univ., Urbana, IL
Volume :
23
Issue :
1
fYear :
2007
Firstpage :
151
Lastpage :
157
Abstract :
The cost of nearest-neighbor (NN) calls is one of the bottlenecks in the performance of sampling-based motion-planning algorithms. Therefore, it is crucial to develop efficient techniques for NN searching in configuration spaces arising in motion planning. In this paper, we present and implement an algorithm for performing NN queries in Cartesian products of R, S1, and RP3, the most common topological spaces in the context of motion planning. Our approach extends the algorithm based on kd-trees, called ANN, developed by Arya and Mount for Euclidean spaces. We prove the correctness of the algorithm and illustrate substantial performance improvement over the brute-force approach and several existing NN packages developed for general metric spaces. Our experimental results demonstrate a clear advantage of using the proposed method for both probabilistic roadmaps and rapidly exploring random trees
Keywords :
path planning; search problems; trees (mathematics); Euclidean spaces; kd-trees; motion-planning algorithms; nearest-neighbor searching; probabilistic roadmaps; Costs; Data structures; Extraterrestrial measurements; Motion-planning; Neural networks; Packaging; Path planning; Pattern recognition; Topology; Tree graphs; Configuration space; kd-trees; nearest-neighbor (NN) searching; probabilistic roadmaps (PRMs); rapidly exploring random trees (RRTs); sampling-based motion planning;
fLanguage :
English
Journal_Title :
Robotics, IEEE Transactions on
Publisher :
ieee
ISSN :
1552-3098
Type :
jour
DOI :
10.1109/TRO.2006.886840
Filename :
4084578
Link To Document :
بازگشت