Scholarlink

<p>Manned-Unmanned Teaming (MUM-T) improves pilot survivability, increases mission effectiveness, and reduces operational costs. For cost effectiveness, Unmanned Aerial Vehicles (UAVs) are likely to have lower maneuverability and performance than those of the manned aircraft. In a close-range combat, the reduced agility is likely to impact the mission success rate and proper tactics should be devised to overcome the weakness of individual aircraft. This study proposes a cooperative air combat tactic for heterogeneous MUM-T operations. We apply QMIX-based Multi-Agent Reinforcement Learning (MARL) in an environment combining high-performance manned fighters and low-performance UAVs. A hierarchical framework is used: the Low Level employs a five-degree-of-freedom (5-DOF) aircraft dynamics model and an intelligent pilot maneuver model, while the High Level focuses on cooperative target selection. Simulation results show that the outnumbered MUM-T is very effective compared to the manned-aircraft-only team. The addition of less agile UAVs significantly improved win rates in all simulated cases.</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

Cooperative Air Combat between High-Performance Manned Fighters and Low-Performance UAVs

  • 期刊论文

  • 发表日期：

    2026年2月14日

查看全文
  • Hyeri Ha,
  • Keeyoung Choi

摘要

Manned-Unmanned Teaming (MUM-T) improves pilot survivability, increases mission effectiveness, and reduces operational costs. For cost effectiveness, Unmanned Aerial Vehicles (UAVs) are likely to have lower maneuverability and performance than those of the manned aircraft. In a close-range combat, the reduced agility is likely to impact the mission success rate and proper tactics should be devised to overcome the weakness of individual aircraft. This study proposes a cooperative air combat tactic for heterogeneous MUM-T operations. We apply QMIX-based Multi-Agent Reinforcement Learning (MARL) in an environment combining high-performance manned fighters and low-performance UAVs. A hierarchical framework is used: the Low Level employs a five-degree-of-freedom (5-DOF) aircraft dynamics model and an intelligent pilot maneuver model, while the High Level focuses on cooperative target selection. Simulation results show that the outnumbered MUM-T is very effective compared to the manned-aircraft-only team. The addition of less agile UAVs significantly improved win rates in all simulated cases.