Scholarlink

This paper proposes a fast collision detection decision method based on CUDA heterogeneous parallel computing technology, designed to meet the demands of obstacle avoidance and collision prevention during UAV flight. With the basic idea of Oriented Bounding Boxes collision detection, the method first projects obstacles and flight trajectories onto three coordinate planes. And within each projection plane, the method parallelly determines whether the obstacles’ boundary segments and flight trajectory segments intersect by CUDA. And the simulation result shows that the CUDA-based method runs faster than the CPU-based method.

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

A CUDA-Based Fast Oriented Bounding Box Collision Detection Method for UAV Flight

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Liu Jiaqi,
  • Huang Xin,
  • Bu Junjie,
  • Zhao Shaobo,
  • Du Bing,
  • Guo Maoyun

摘要

This paper proposes a fast collision detection decision method based on CUDA heterogeneous parallel computing technology, designed to meet the demands of obstacle avoidance and collision prevention during UAV flight. With the basic idea of Oriented Bounding Boxes collision detection, the method first projects obstacles and flight trajectories onto three coordinate planes. And within each projection plane, the method parallelly determines whether the obstacles’ boundary segments and flight trajectory segments intersect by CUDA. And the simulation result shows that the CUDA-based method runs faster than the CPU-based method.