Scholarlink

This paper studies the problem of maneuver decision-making for multi-aircraft air combat. First, considering pertinent air combat scenarios, the aircraft model and the air combat attack zone are established. Then, the Attention Mechanism (AM) is introduced to improve the convergence speed of the Proximal Policy Optimization (PPO) algorithm. Moreover, we design a maneuver decision-making model for multi-aircraft air combat based on the proposed AM-PPO algorithm. Finally, through 3v3 air combat simulation experiments, this paper validates the AM-PPO algorithm against the classical PPO algorithm in terms of convergence speed and stability. The simulation results further illustrate that the proposed algorithm can effectively control the aircraft to execute strategic maneuvers against opponents while avoiding friendly aircraft.

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

Maneuver Decision Making for Multi-aircraft Air Combat Based on Reinforcement Learning with Attention Mechanism

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Peida Li,
  • Xiaoduo Li,
  • Liang Han

摘要

This paper studies the problem of maneuver decision-making for multi-aircraft air combat. First, considering pertinent air combat scenarios, the aircraft model and the air combat attack zone are established. Then, the Attention Mechanism (AM) is introduced to improve the convergence speed of the Proximal Policy Optimization (PPO) algorithm. Moreover, we design a maneuver decision-making model for multi-aircraft air combat based on the proposed AM-PPO algorithm. Finally, through 3v3 air combat simulation experiments, this paper validates the AM-PPO algorithm against the classical PPO algorithm in terms of convergence speed and stability. The simulation results further illustrate that the proposed algorithm can effectively control the aircraft to execute strategic maneuvers against opponents while avoiding friendly aircraft.