狭窄空间约束下的多无人机协同规划研究

Cooperative Path Planning for Multiple UAVs in Constrained Narrow Spaces

摘要:

多无人机系统因其高度的灵活性和任务互补性，十分适合复杂且狭窄的工作环境，在军事、救援、监测和物流等多种领域展现出了巨大的应用潜力。然而，在受限空间中实现多架无人机安全、高效的路径规划仍然面临诸多挑战。一方面，安全搜索空间的减少限制了符合无人机运动约束飞行路径的生成；另一方面，狭窄的空间增加了机间碰撞风险。因此，提出一种基于机体约束树的多无人机协同路径生成方法，采用分层协作规划框架，其中包含高层的冲突规避模块和底层的路径生成模块。在该框架中，通过改进了增强冲突搜索算法（Enhanced Conflict-Based Search，ECBS），使算法能够有效地处理全局路径搜索中的冲突问题， 避免局部死锁的产生。通过在全局规划阶段优化冲突处理策略，实现了多无人机在复杂环境中稳定、高效地飞行。此外，在 Rviz 仿真环境下对随机、对立、交错的多飞行任务进行了实验验证。实验结果显示，所提出的方法能够在数分钟内为至少 60 架无人机生成平滑的协同路径。

Abstract:

Due to the high flexibility and complementary，Multi-UAV systems are suitable for complex and constrained operational environments， demonstrating significant application potential in military， rescue， monitoring，and logistics fields. However，there are still challenges to achieving safe and efficient path planning for multiple UAVs in constrained spaces. On the one hand，the reduction of the safe search space restricts the generation of flight paths conforming to UAV motion constraints. On the other hand，the constrained space increases the risk of collision between UAVs. Thus，in this paper，a cooperative path generation method based on a body constraint tree is proposed. It adopts a hierarchical collaboration planning framework that includes a high-level conflict avoidance module and a low-level path generation module. In the framework，the Enhanced Conflict-Based Search（ECBS）algorithm is improved to address conflict in global path planning and to avoid the occurrence of local deadlocks. By optimizing the conflict resolution strategy during the global planning phase，a stable and efficient collaboration of multi-UAV systems in complex environments is achieved. In addition，experiments are conducted in the Rviz simulation environment with random，adversarial，and interleaved multi-flight missions. The experimental results demonstrate that the proposed method can generate smooth cooperative paths for at least 60 UAVs within a few minutes.