Work

QS ‐ RRT *: An Improved Path Planning Algorithm with Goal‐Biased Strategy

Year: 2025

Type: article

Abstract: The rapidly‐exploring Random Tree (RRT) algorithm, based on sampling, is a popular algorithm for path planning problems. The optimal RRT (RRT*) is an extension of the RRT algorithm, which provides asymptotic optimality. However, RRT* still suffers from slow convergence and high initial solution cost. To address these issues, this paper proposes QS‐RRT*, an improved version of the RRT* algorithm. QS‐RRT* can generate a better initial solution and converge to the optimal solution more quickly. Second, QS‐RRT* not only considers a set of vertices around the new node but also takes its ancestor nodes into account up to a given depth, thereby expanding the number of potential parent nodes and generating a path with lower cost. After obtaining a high‐quality initial solution, intelligent sampling is used to generate sampling points near obstacles, further reducing the path cost. The algorithm was compared with Q‐RRT* and RRT*‐Smart through a simulation platform, demonstrating the advantages of QS‐RRT* in terms of initial solution quality and convergence speed. © 2025 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

Source: IEEJ Transactions on Electrical and Electronic Engineering

DOI: 10.1002/tee.70194

Authors: Jie Luo, Xu Zhang, Dawei Tu

Institution: Shanghai University

Language: English

Topic: Robotic Path Planning Algorithms

Subfield: Computer Vision and Pattern Recognition

Field: Computer Science

Domain: Physical Sciences

Open Access status: bronze

Funder: National Natural Science Foundation of China

Awards: G1310850233, G877063096