SOF-RRT: An improved path planning algorithm using spatial offset sampling

Recently, sampling-based algorithms have been demonstrated advantages in complex and high-dimensional environments. The RRT* algorithm, as the best variant of RRT, provides progressive optimality. It is shown that the RRT* algorithm has a slow convergence speed and a high initial path cost that resu...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Engineering applications of artificial intelligence 2023-11, Vol.126, p.106875, Article 106875
Hauptverfasser: Yu, Shanen, Chen, Jianke, Liu, Guangyu, Tong, Xiaolong, Sun, Yingyi
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Recently, sampling-based algorithms have been demonstrated advantages in complex and high-dimensional environments. The RRT* algorithm, as the best variant of RRT, provides progressive optimality. It is shown that the RRT* algorithm has a slow convergence speed and a high initial path cost that result in the low efficiency of the algorithm, due to the low probability of adequate sampling caused by the invalid expansion of the tree and redundant sampling. To overcome these limitations, the SOF-RRT* algorithm is proposed to generate better initial solutions with higher stability and faster convergence. The SOF-RRT* algorithm is an improvement on the F-RRT* algorithm. The SOF-RRT* algorithm introduces a spatial probability weight sampling strategy into sampling, which makes the sampling probability higher in the area with a large feasible area. It reduces redundant sampling and increases the effective sampling rate. In the aspect of tree expansion, the strategy of target bias and obstacle tangential bias is introduced to improve the tree expansion efficiency, which makes tree expansion away from obstacles and bias to the target point. The proposed algorithm improves the expansion efficiency of sampling and tree. Finally, the initial path generation and convergence speed are simulated and compared, which proves that the algorithm has been greatly optimized in terms of the number of iterations, path quality, and convergence speed. •A path planning algorithm using spatial bias sampling and extended bias strategy.•Suitable for different environments at the same time.•Use target bias and obstacle tangential bias to improve expansion efficiency.•The proposed algorithm has the best results and the most stable operation.
ISSN:0952-1976
1873-6769
DOI:10.1016/j.engappai.2023.106875