Scholarlink

<p>This study presents an enhanced variant of the Rapidly Exploring Random Tree (RRT) algorithm tailored for efficient local path planning in autonomous vehicles. Building upon foundational RRT principles, the proposed algorithm integrates novel modifications to improve search efficiency during the sampling phase. These enhancements include the use of advanced techniques for obstacle detection and selective sampling. Additionally, factors such as the actual vehicle size and surrounding obstacles are considered during the expansion phase. The enhanced algorithm was rigorously evaluated through extensive simulations conducted in three distinct autonomous driving scenarios. The results demonstrate significant improvements over RRT, RRT*, and A-RRT* across several indicators: computational cost is reduced by up to 78%, planning time by 64% to 78%, and trajectory cost is minimized by more than 97%. In addition, path smoothness is significantly improved, with sharp turns and unplanned detours reduced. This research contributes to advancing autonomous vehicle technology by providing a versatile and reliable solution to local path-planning challenges.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

An Improved RRT Algorithm for Efficient Local Path Planning in Autonomous Vehicles

  • 期刊论文

  • 发表日期：

    2025年6月18日

查看全文
  • Jawad Abdouni,
  • Tarik Jarou,
  • Abderrahim Waga,
  • Karima Bensassi,
  • Younes El koudia

摘要

This study presents an enhanced variant of the Rapidly Exploring Random Tree (RRT) algorithm tailored for efficient local path planning in autonomous vehicles. Building upon foundational RRT principles, the proposed algorithm integrates novel modifications to improve search efficiency during the sampling phase. These enhancements include the use of advanced techniques for obstacle detection and selective sampling. Additionally, factors such as the actual vehicle size and surrounding obstacles are considered during the expansion phase. The enhanced algorithm was rigorously evaluated through extensive simulations conducted in three distinct autonomous driving scenarios. The results demonstrate significant improvements over RRT, RRT*, and A-RRT* across several indicators: computational cost is reduced by up to 78%, planning time by 64% to 78%, and trajectory cost is minimized by more than 97%. In addition, path smoothness is significantly improved, with sharp turns and unplanned detours reduced. This research contributes to advancing autonomous vehicle technology by providing a versatile and reliable solution to local path-planning challenges.