Improved RRT*-Smart Algorithm for UGV Path Planning in Emergency Rescue Scenarios
Shaoqun Li, Chen Zhu · 2024
In emergency rescue scenarios, path planning for UGVs is a mission-critical task. However, many algorithms based on the Rapidly-exploring Random Trees (RRT) often exhibit instability and inefficiency in such scenarios. Therefore, to address the challenge of UGV path planning in emergency rescue scenarios, this paper proposes an improved version of the RRT*-Smart algorithm named TAD-RRT*-Smart. The improvement of the proposed algorithm mainly involves three aspects: First, a target point attraction module is introduced, which integrates APF to make the random tree oriented during expansion, thus accelerating the convergence rate. Second, a dynamic step size adjustment module is introduced, which enhances the adaptability of UGV in different environments. Finally, the tree node expansion process is refined, which enhances the expansion efficiency of the intelligent sampling and the ability to pass through complex areas. The simulation experiments show that the TAD-RRT*-Smart algorithm exhibits outstanding optimal path quality and stability in various environments, as well as superior initial path generation capability. Thus, it can provide a reliable answer for UGV path planning in emergency rescue scenarios.