Tasneem, Sameeha (University of California, Riverside), Wu, Guoyuan (University of California-Riverside), Stas, Mike (University of California Riverside), Barth, Matthew (University of California-Riverside)

ARRT*: A Hybrid Heuristic RRT* Approach for Efficient Path Planning

Scheduled for presentation during the Regular Session "S41c-Motion Planning, Trajectory Optimization, and Control for Autonomous Vehicles" (FR-LA-T41), Friday, November 21, 2025, 16:00−16:20, Broadbeach 1&2

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 18, 2025

Abstract

Path planning algorithms must balance computational efficiency and path optimality, particularly in real-time applications. Rapidly-exploring Random Tree (RRT) and its variants provide fast execution but often generate suboptimal paths, while grid-based algorithms like A* and hybrid A* yield optimal paths at a high computational cost. This paper presents a novel path-planning approach that integrates heuristic functions from A* and hybrid A* into the RRT* framework. By refining node cost evaluation, our method enhances both path quality and computational efficiency. Experimental evaluations and ablation studies demonstrate its generalizability across multiple RRT variants. Furthermore, we explore the impact of different sampling strategies in conjunction with our modified cost function to optimize key metrics such as path length, execution time, and smoothness. As a practical application, we implement this enhanced algorithm in an automated parking system, a key feature in Connected and Automated Vehicles (CAVs). Quantitative results show that ARRT* reduces path length by up to 17% compared to standard RRT*, while also improving computational efficiency and minimizing steering variations, making it more suitable for real-world autonomous navigation in constrained environments.