Scholarlink

<p>To address the challenges of frequent sudden obstacles and delayed dynamic response in autonomous vehicle path planning within urban weakly regulated zones (WR-Zones), this paper proposes GAPF-PRRT—a real-time path planning algorithm integrating Gaussian artificial potential fields (GAPF) with a pruned rapidly exploring random tree (PRRT). Firstly, Gaussian decay factors are proposed to reconstruct the repulsive field function, enabling dynamic adjustment of obstacle influence ranges while effectively resolving artificial potential fields (APF)'s inherent local optima limitation. Secondly, a comprehensive multi-modal field model is created, incorporating both road boundary repulsion and dynamic obstacle-inflated repulsion fields to significantly enhance real-time obstacle avoidance in complex WR-Zones environments. Subsequently, an efficient goal-biased KD-tree pruning mechanism is implemented to substantially accelerate rapidly exploring random tree (RRT)'s global search capability. Finally, a Bézier curve optimizer with integrated collision cone detection is innovatively developed to achieve a Pareto-optimizer balance between path smoothness and length. Simulations under critical WR-zones scenarios—such as sudden lane occupancy, variable-width roads, and narrow-lane encounters—demonstrate the superior performance of the proposed method compared to improved artificial potential field (improved-APF), rapidly exploring random tree star algorithm (RRT*), and bi-directional RRT-APF fusion algorithm (Bi-RRT-APF), achieving an average reduction in path length of 17.05%, an 81.08% improvement in planning speed, and obstacle response times consistently below 11.1&#xa0;ms.</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

Synergistically integrated Gaussian artificial potential field and pruned rapidly exploring random tree for real-time path planning of autonomous vehicles in weakly regulated zones

  • 期刊论文

  • 发表日期：

    2025年10月3日

查看全文
  • Xiaolan Wu,
  • Xinyang Wang,
  • Zhifeng Bai,
  • Guifang Guo

摘要

To address the challenges of frequent sudden obstacles and delayed dynamic response in autonomous vehicle path planning within urban weakly regulated zones (WR-Zones), this paper proposes GAPF-PRRT—a real-time path planning algorithm integrating Gaussian artificial potential fields (GAPF) with a pruned rapidly exploring random tree (PRRT). Firstly, Gaussian decay factors are proposed to reconstruct the repulsive field function, enabling dynamic adjustment of obstacle influence ranges while effectively resolving artificial potential fields (APF)'s inherent local optima limitation. Secondly, a comprehensive multi-modal field model is created, incorporating both road boundary repulsion and dynamic obstacle-inflated repulsion fields to significantly enhance real-time obstacle avoidance in complex WR-Zones environments. Subsequently, an efficient goal-biased KD-tree pruning mechanism is implemented to substantially accelerate rapidly exploring random tree (RRT)'s global search capability. Finally, a Bézier curve optimizer with integrated collision cone detection is innovatively developed to achieve a Pareto-optimizer balance between path smoothness and length. Simulations under critical WR-zones scenarios—such as sudden lane occupancy, variable-width roads, and narrow-lane encounters—demonstrate the superior performance of the proposed method compared to improved artificial potential field (improved-APF), rapidly exploring random tree star algorithm (RRT*), and bi-directional RRT-APF fusion algorithm (Bi-RRT-APF), achieving an average reduction in path length of 17.05%, an 81.08% improvement in planning speed, and obstacle response times consistently below 11.1 ms.