PITD3-based Autonomous Evasion Decision-making Method for Aircraft
作者单位：School of Electrical Engineering and Automation Beijing Aerospace Automatic Control Institute Tianjin Aerospace Institute of Computing Technology
会议名称：《第43届中国控制会议》
会议日期：1000年
学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 081105[工学-导航、制导与控制] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）]
关 键 词：PITD3 Autonomous Evasion Decision-Making Deep Reinforcement Learning Decision Network Aircraft
摘 要：Outstanding autonomous evasion decision-making capability has an important significance for ensuring flight safety and enhancing the autonomy of the aircraft. In this paper, a prediction information-based TD3(PITD3) evasion decision-making algorithm for aircraft is proposed, which can autonomously evade the interceptors attack in an antagonistic environment. Firstly,an antagonistic environment is described according to environmental constraints and the dynamics of the aircraft and interceptor. Secondly, the autonomous decision-making process of the aircraft facing the threat target is modelled as a Markov decision process to describe the state information of the aircraft and interceptor. Finally, an autonomous evasion decision-making method based on the twin delayed deep deterministic policy gradient(TD3) is designed to find the optimal strategy, and the effectiveness of the proposed autonomous evasion decision-making method is demonstrated by simulation results and comparative experiments.