Title :
An approach to speed up RRT
Author :
Yun-xiao Shan ; Bi-jun Li ; Jian-Zhou ; Yue-Zhang
Author_Institution :
State Key Lab. of Inf. Eng. in Surveying Mapping & Remote Sensing, Wuhan Univ., Wuhan, China
Abstract :
We present an improved algorithm to RRT* in this paper. Our algorithm tries to increase the efficiency by replacing the Initialtree() of RRT*with a RRT tree only contains one solution to speed up finding a feasible solution, and then applying a RRT*_S algorithm to optimize the current solution. This RRT*_S is similar with RRT* in principle, but it does not insert extra nodes into the current tree and optimizes just in fixed nodes instead of all the nearby nodes. We demonstrate the performance of this improved algorithm on two benchmark scenarios to show the effectiveness compared with RRT* and RRT.
Keywords :
mobile robots; path planning; trees (mathematics); Initialtree; RRT tree; RRT*_S algorithm; motion planning; rapidly-exploring randomized tree; robotics; Benchmark testing; Convergence; Measurement; Optimization; Planning; Robots; Standards;
Conference_Titel :
Intelligent Vehicles Symposium Proceedings, 2014 IEEE
Conference_Location :
Dearborn, MI
DOI :
10.1109/IVS.2014.6856473