对抗环境下多无人机轨迹规划与动态三维重建

Trajectory planning and dynamic 3D reconstruction for multiple UAVs in adversarial environments

摘要:

无人机集群已经在俄乌冲突等现代战争中常态化运用,成为对敌打击和侦察的重要力量。面向未来海上无人集群智能作战需要,亟须研究无人机集群在对抗环境下的密集障碍规避、非合作目标协同绕飞与三维重建的问题。为提升无人集群在复杂海上环境中的自主感知与协同建模能力,提出一种面向动态环境的集群智能规划与三维重建一体化方法。首先,在协同轨迹规划阶段,提出了基于领航者的无人机协同跟踪方法,领航机负责实时锁定并跟踪战舰等高价值非合作目标,僚机通过队形保持机制维持编队稳定。其次,当遭遇无人集群移动障碍时,通过路径重规划与机动动作实现威胁规避,并在安全通过后迅速恢复目标追踪。最后,在无人机集群抵近兴趣目标后进入动态三维重建阶段,设计了多无人机协同绕飞与三维重建方法,实现对非合作机动目标的多视角观测与动态三维重建。通过海上仿真试验表明,文中所提出的方法使得无人机集群在复杂对抗环境下自主完成对战舰的实时跟踪与高效重建,能够兼顾任务完成率与飞行安全性,为无人机集群在未来战场的侦察监视和情报获取任务提供了有效方案。

Abstract:

Unmanned aerial vehicle (UAV) swarms have become a normalized element of modern warfare-as exemplified by the Russia-Ukraine conflict-serving as critical assets for strikes and reconnaissance. To satisfy the requirements of future maritime swarm operations, it is imperative to investigate dense-obstacle avoidance, cooperative bypass of non-cooperative targets, and three-dimensional reconstruction for UAV swarms operating in adversarial environments. To enhance autonomous perception and cooperative mapping in complex maritime scenarios, this work proposes an integrated approach for swarm intelligence planning and 3D reconstruction tailored to dynamic environments. Initially, in the cooperative trajectory planning stage, a leader-based UAV cooperative tracking method is proposed. The leader UAV is responsible for real-time locking and tracking of high-value non-cooperative targets such as warships, while the follower UAVs maintain formation stability through a formation-keeping mechanism. Furthermore, when encountering mobile obstacles from other unmanned swarms, the swarm avoids threats through path re-planning and maneuvering actions, and quickly resumes target tracking after safely passing through. Ultimately, in the dynamic 3D reconstruction phase, after the UAV swarm approaches the target of interest, a multi-UAV cooperative circumnavigation and 3D reconstruction method is designed to achieve multi-perspective observation and dynamic 3D reconstruction of non-cooperative maneuvering targets. Sea-based simulation experiments demonstrate that the methods proposed in this paper enable the UAV swarm to autonomously complete real-time tracking and efficient reconstruction of warships in complex adversarial environments, balancing mission completion rate and flight safety. This provides an effective solution for the reconnaissance, surveillance, and intelligence gathering tasks of UAV swarms in future battlefields.