航空科学技术2025，Vol.36Issue(7)：67-76,10.DOI:10.19452/j.issn1007-5453.2025.07.009
基于SD3算法与轨迹预测的无人机空战决策研究
UAV Decision-making in Air Combat Based on SD3 Algorithm and Trajectory Prediction
摘要
Abstract
UAV autonomous decision-making is a key technology to enhance UAV autonomy in air combat.Aiming at the problem of UAV autonomous decision-making in air combat,a UAV decision-making method based on deep reinforcement learning is developed in this paper.The model takes the real-time situation data of air combat as input and outputs UAV maneuver commands.The reward function is established based on the air combat situation and UAV air-to-air missile attack envelop model.The residual gated recurrent unit(GRU)model is introduced to predict the trajectory of the target UAV.The Softmax deep double deterministic policy gradients(SD3)algorithm is cited so that the UAV can choose the appropriate maneuver decision.The simulation experimental demonstrates that the proposed method make accurate maneuvers,gain advantages in air combat,and improve the autonomy of UAV in decision-making.关键词
无人机/自主机动决策/轨迹预测/深度强化学习/SD3算法Key words
UAV/autonomous maneuver decision/trajectory prediction/deep reinforcement learning/SD3 algorithm分类
航空航天引用本文复制引用
金文霄,周同乐,陈谋,朱豪杰..基于SD3算法与轨迹预测的无人机空战决策研究[J].航空科学技术,2025,36(7):67-76,10.基金项目
国家自然科学基金(62203217) （62203217）
江苏省基础研究计划自然科学基金(BK20220885) （BK20220885）
中国高校产学研创新基金无人集群协同智能项目(2021ZYA05001)National Natural Science Foundation of China(62203217) （2021ZYA05001）
Jiangsu Province Basic Research Program Natural Science Foundation(BK20220885) （BK20220885）
Unmanned Swarm Cooperation Intelligence Project of China Higher Education Institution Industry-University-Research Innovation Fund(2021ZYA05001) （2021ZYA05001）
