航空科学技术2025，Vol.36Issue(8)：1-10,10.DOI:10.19452/j.issn1007-5453.2025.08.001
基于深度强化学习的飞行控制方法研究现状及发展
Research Status and Development of Flight Control Methods Based on Deep Reinforcement Learning
摘要
Abstract
With the rapid development of aviation technology,the demand for intelligent flight control systems is increasing.Deep reinforcement learning,as a cutting-edge artificial intelligence technology,provides new ideas for solving complex control problems in aircraft and has important theoretical and practical value.This article delves into the application of deep reinforcement learning in flight control,analyzes its basic principles,and explores controller design methods and performance evaluation.Research has found that deep reinforcement learning can effectively handle nonlinear and uncertain problems in aircraft control,and achieve efficient control by learning the optimal flight strategy through the interaction between intelligent agents and the environment.This article also analyzes the main types of deep reinforcement learning algorithms and examines the control effects of several algorithms.Although deep reinforcement learning methods have unique advantages in dealing with complex flight control problems,they also face challenges in training efficiency and generalization ability.关键词
深度强化学习/飞行控制系统/人工智能Key words
deep reinforcement learning/flight control system/artificial intelligence分类
信息技术与安全科学引用本文复制引用
卓博群,杨文骏,李建康..基于深度强化学习的飞行控制方法研究现状及发展[J].航空科学技术,2025,36(8):1-10,10.基金项目
航空科学基金(201905018003) （201905018003）
陕西省秦创原创新"科学家+工程师"队伍建设基金(2023KXJ-075) Aeronautical Science Foundation of China(201905018003) （2023KXJ-075）
Shaanxi Province QinChuangYuan"Scientist+Engineer"Team Construction Fund(2023KXJ-075) （2023KXJ-075）
