空天防御2023，Vol.6Issue(4)：51-57,7.
基于近端策略优化的制导律设计
Guidance Law Based on Proximal Policy Optimization
摘要
Abstract
The design of guidance law is critical in the interception system.The accuracy of the commonly used variable structure guidance law decreases while intercepting complex manoeuvring targets,and chattering occurs frequently.This paper has proposed a guidance law design method based on near-end strategy optimization.The guidance problem of intercepting manoeuvring targets was abstracted as a Markov decision process,and a reward function evaluating miss distance and line-of-sight angular rate chattering was applied.Comparative experiments show that the interception effect of the guidance law based on near-end strategy optimization and continuous output performs more effectively and can successfully restrain the chattering phenomenon in the sliding mode guidance law,thus providing a significant research prospect and potential application value.关键词
制导律/强化学习/滑模控制/近端策略优化Key words
guidance law/reinforcement learning/sliding mode control/proximal policy optimization分类
信息技术与安全科学引用本文复制引用
李梦璇,郭建国,许新鹏,沈昱恒..基于近端策略优化的制导律设计[J].空天防御,2023,6(4):51-57,7.基金项目
国家自然科学基金(61973254) （61973254）
西北工业大学硕士研究生实践创新能力培育基金(PF2023044) （PF2023044）
