A BIT* algorithm with dilated vertices-based path stretching strategy

摘要

The path planning of mobile robots is a critic issue in robotics. The existing algorithms face challenges in terms of path quality and convergence speed. To enhance planning efficiency and path optimality, this paper proposes a novel algorithm: dilated vertices-based stretching BIT* (DS-BIT*) algorithm. The proposed method integrates a path-stretching strategy based on dilated vertices into the BIT* framework. After finding a path, the algorithm improves it using the dilated vertices-based path-stretching algorithm. This approach effectively removes redundant waypoints, enhances path smoothness, and reduces the overall path cost. A lower path cost facilitates faster contraction of the informed search region, thereby improving the path convergence speed. In addition, DS-BIT* performs bidirectional improvement of the final path at the end of the algorithm, further reducing the final path cost. Comparative experiments with RRT*, Informed RRT*, and BIT* in both simple and complex environments demonstrate the superior performance of DS-BIT* in terms of planning efficiency and path quality. • A dilated vertices-based path stretching BIT* (DS-BIT*) method is proposed. • Modeled by rectangular, triangular, and circular, the obstacles are dilated and then the dilated vertices are obtained. • A bidirectional implementing mechanism of the DS-BIT* is proposed, which can further reduce the path cost.