Scholarlink

This paper introduces a two-stage framework for quadrotor unmanned aerial vehicle (UAV) swarm cooperation. First, we model coalition formation as a potential game, balancing coverage and distance costs with saturation effects for uniform distribution. Second, we apply Minimax Deep Deterministic Policy Gradient (Minimax-DDPG) for cooperative flight control based on the coalitions. Our approach combines coalition theory with reinforcement learning for effective target allocation and coordinated engagement in adversarial environments.

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

Two-Stage Decision-Making for UAV Swarms via Potential Games and Reinforcement Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Zekun Duan,
  • Genjiu Xu,
  • Zesheng Li

摘要

This paper introduces a two-stage framework for quadrotor unmanned aerial vehicle (UAV) swarm cooperation. First, we model coalition formation as a potential game, balancing coverage and distance costs with saturation effects for uniform distribution. Second, we apply Minimax Deep Deterministic Policy Gradient (Minimax-DDPG) for cooperative flight control based on the coalitions. Our approach combines coalition theory with reinforcement learning for effective target allocation and coordinated engagement in adversarial environments.