AE-RRT*: Adaptive Ellipse-Based Optimal Path Planning
Xu Cao, Hao Yang Shen, Jing Wang, Yangxin Zhao, Kang Wang, Xihong Fei · 2024
Rapidly-exploring Random Tree Star (RRT*) is a well-known sampling-based planning approach, an improvement over the traditional Rapidly Exploring Random Tree (RRT) algorithm. It is suitable for high-dimensional spaces, dynamic environments, and various complex geometric structures. However, the primary challenge of RRT* is its significant computational complexity, especially in high-dimensional spaces. To address these limitations, we propose an improved optimal path planning algorithm, namely the Adaptive Ellipse RRT* (AE-RRT*). The AE-RRT* builds on the RRT* algorithm, integrating heuristic search principles, dynamic ellipse optimization strategies, and phased exploration strategies to improve the performance of the path planning algorithm. In the random sampling phase, heuristic principles generate specific random points, significantly reducing the search space. The dynamic ellipse optimization strategy adjusts the shape of the ellipse dynamically, aligning the sampling region more closely with potential paths. The phased exploration strategy thoroughly explores the search space in the initial stages and focuses energy on path optimization in the later stages. Simulation results indicate that, compared to commonly used traditional algorithms and their enhancements, AE-RRT* can achieve superior convergence performance.