cs.RO提交于 2026-10-26已译

UAV-SwarmAgent：通过代码生成与视觉语言推理实现可扩展的多无人机侦察与任务分配

UAV-SwarmAgent: Scalable Multi-UAV Reconnaissance and Task Allocation via Code Generation and Vision-Language Reasoning

Oleg Sautenkov · Yasheerah Yaqoot · Muhammad Ahsan Mustafa · Grigorii Kalinin · Chih-Yung Wen · Dzmitry Tsetserukou

中文摘要

多无人机侦察任务通常需要规划航点并持续监控每架无人机。UAV SwarmAgent 是一种分层多智能体系统：操作员用一句自然语言描述任务，地面站智能体编写并执行代码，将其转换为各无人机的独立路线。该智能体借助视觉语言模型(vision-language model)在空中地图中定位检查目标，通过平衡路线长度将目标分配给 N 架无人机，并调度工作智能体使用紧凑的机载决策模型核验事件。仿真涵盖 4 种事件类型和 3 种基础模型，集群在 255 次任务中的 244 次成功定位事件，成功率为 95.7%。针对 1 至 5 架无人机的配对扩展研究显示，飞行时间符合 Amdahl 型定律；第二架无人机贡献了可实现节省量的一半，加速比上限为 1.52x。论文讨论了这种可预测性及观察到的误报对集群操作员的意义，并概述了检验该系统的用户研究。

关键要点

01分层多智能体系统可将自然语言任务转换为各无人机的路线与目标分配
02结合视觉语言模型定位目标，并用紧凑机载决策模型核验事件
03仿真在 255 次任务中成功定位 244 次事件，成功率为 95.7%
041 至 5 架无人机的扩展实验表明飞行时间符合 Amdahl 型定律
05第二架无人机贡献一半可实现节省量，整体加速比上限为 1.52x

解读

尚无解读。

原始英文摘要

arXiv:2610.32559v1 Announce Type: new Abstract: Supervising a multi-UAV reconnaissance mission usually means plotting waypoints and watching every drone. We present UAV SwarmAgent, a hierarchical multi-agent system in which an operator states the mission in one natural-language sentence and a ground-station agent writes and executes code that turns it into per-UAV routes. The agent grounds inspection targets in an aerial map with a vision-language model, splits them among N UAVs by balancing route length, and dispatches worker agents that verify incidents with a compact onboard decision model. In simulation over 4 incident types and 3 foundation models, the swarm located the incident in 244 of 255 missions (95.7%). A paired scaling study with one to 5 UAVs shows that flight time follows an Amdahl-type law. The second UAV delivers half of the attainable saving and the speedup is bounded at 1.52x. We discuss what this predictability, and the false alarms we observed, imply for an operator who supervises the swarm, and outline the user study needed to test it.