Heuristic Bidirectional Fast Marching Tree for Optimal Motion Planning
Wenxiang Gao, Qing Tang, Jin Yong Yao, Yaru Yang, Deping Yu · 2018
To accelerate the planning efficiency of the sampling-based motion planning algorithms, we combined the heuristic strategy with the bidirectional fast marching tree (BFMT*) to solve the collision-free motion planning problems. The modified algorithm can ensure that the number of the extended points can be reduced as much as possible, in order to speed up the rate of the motion planning. In our works, firstly a new algorithm named Heuristic Bidirectional Fast Matching Tree (HBFMT*) is proposed. Then the extending procedure which explores the sampled points with goal priority is analyzed, and the relationship between the extending procedure and the weight is given. Thirdly, the weights of the heuristic are compared in variety of typical maps. We gives some suggestions on how to choose the appropriate parameters on solving motion planning problems. Finally, the algorithm is compared with other sampling-based algorithms, and the results show that HBFMT* has a fast search ability than other algorithms, such as BFMT*, FMT*, RRT*.