Scholarlink

<p>To address challenges faced by heterogeneous aerial-ground unmanned systems in complex and dynamic environments—including discrepancies in perception and mapping, difficulties in multi-source map fusion, and limited robustness in path planning. This paper proposes a real-time mapping and navigation algorithm based on aerial-ground cooperation to improve environmental perception and autonomous decision-making. For mapping, UAV employs the FAST-LIO2 and OctoMap algorithms to efficiently generate high-precision 3D maps. In parallel, a projection-based probabilistic fusion mechanism is used to create 2D grid maps, providing accurate navigation information for UGV. In the context of navigation, a D²ACO algorithm that integrates goal-directed guidance with a dynamic pheromone regulation mechanism is proposed to enable both global path optimization and the construction of local dynamic obstacle avoidance windows. Experimental results demonstrate that the proposed method effectively addresses both map construction and dynamic obstacle avoidance in complex terrain and supports high-precision autonomous navigation even in GPS-denied environments. Compared to the original ACO algorithm, the D²ACO algorithm achieves reductions of 47.3% in search time, 17.1% in path length, and 87.8% in the number of iterations, offering a more efficient and reliable solution for path planning in unmanned systems.</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

Aerial-ground collaborative mapping and path planning algorithm for unmanned systems in dynamic environments

  • 期刊论文

  • 发表日期：

    2025年11月6日

查看全文
  • Shuolin Peng,
  • Zai Luo,
  • Li Yang,
  • Wensong Jiang,
  • Songbai Yu

摘要

To address challenges faced by heterogeneous aerial-ground unmanned systems in complex and dynamic environments—including discrepancies in perception and mapping, difficulties in multi-source map fusion, and limited robustness in path planning. This paper proposes a real-time mapping and navigation algorithm based on aerial-ground cooperation to improve environmental perception and autonomous decision-making. For mapping, UAV employs the FAST-LIO2 and OctoMap algorithms to efficiently generate high-precision 3D maps. In parallel, a projection-based probabilistic fusion mechanism is used to create 2D grid maps, providing accurate navigation information for UGV. In the context of navigation, a D²ACO algorithm that integrates goal-directed guidance with a dynamic pheromone regulation mechanism is proposed to enable both global path optimization and the construction of local dynamic obstacle avoidance windows. Experimental results demonstrate that the proposed method effectively addresses both map construction and dynamic obstacle avoidance in complex terrain and supports high-precision autonomous navigation even in GPS-denied environments. Compared to the original ACO algorithm, the D²ACO algorithm achieves reductions of 47.3% in search time, 17.1% in path length, and 87.8% in the number of iterations, offering a more efficient and reliable solution for path planning in unmanned systems.