The Rapidly Exploring Random Tree (RRT*) algorithm is widely recognized for its probabilistic completeness and asymptotic path optimization, but it faces challenges with suboptimal path quality and high computational time costs. This paper proposes a fast RRT* algorithm based on dynamic probability adjustment and bi-directional insertion point optimization (DPI-RRT*), aiming to improve path quality and reduce discovery time.
The RRT* algorithm is widely recognized for its probabilistic completeness and asymptotic path optimization, but it faces challenges with suboptimal path quality and high computational time costs. This paper proposes a fast RRT* algorithm based on dynamic probability adjustment and bi-directional insertion point optimization (DPI-RRT*), aiming to improve path quality and reduce discovery time.
Simulation experiments show that DPI-RRT* outperforms traditional RRT*, Quick-RRT* and Informed-Quick RRT* algorithms, with at least 23.6% improvement in suboptimal solution discovery and at least 21.1% reduction in discovery time.
The DPI-RRT* algorithm significantly enhances path planning efficiency and quality by dynamically adjusting the generation probability of random points and optimizing turning points via bi-directional insertion, demonstrating superior performance compared to traditional RRT* algorithms.
