“会学习”运载火箭的制导控制技术  

Guidance Navigation and Attitude Control Technology of“Self-Learning”Launch Vehicle

作　　者：马卫华[1,2,3] 禹春梅 路坤锋[1,2,3] 柳嘉润 司文杰[1,2,3] 李文婷 Ma Weihua;Yu Chunmei;Lu Kunfeng;Liu Jiarun;Si Wenjie;Li Wenting(Beijing Aerospace Automatic Control Institute,Beijing 100854,China;National Key Laboratory of Science and Technology on Aerospace Intelligent Control,Beijing 100854,China;Innovation Center of Aerospace Intelligent Technology,Beijing 100854,China)

机构地区：[1]北京航天自动控制研究所,北京100854 [2]宇航智能控制技术国家级重点实验室,北京100854 [3]航天智能技术创新中心,北京100854

出　　处：《航天控制》2020年第2期3-8,共6页Aerospace Control

摘　　要："会学习"运载火箭具备"边飞边学"和"终身学习"的特征。对于"会学习"运载火箭制导控制技术具备的5个能力需求,即飞行状态在线辨识与感知、制导控制在线重构、经验知识自学习、"终身学习"和箭上强计算能力等进行了分析,研究了支撑"边飞边学"的在线辨识与控制重构、轨迹在线规划、目标在线变更等技术和支撑"越飞越聪明"的数据挖掘、自我学习、持续优化和决策评估等技术。最后,对"会学习"运载火箭的智能控制技术进行思考与展望。The’Self-learning’launch vehicle takes advantage of two characteristics,known as’learning by flying’and’lifelong learning’,and five capability requirements of the guidance and control technology of’self-learning’launch vehicle are summarized.They are online identification and perception of flight status,online reconstruction of guidance and control,self-learning of experience knowledge,’lifelong learning’and strong onboard computation.The key technologies are researched,which are necessary for’learning by flying’by including online identification and control reconstruction,online trajectory planning,online target alteration as well as those necessary technologies for’flying smarter’that include data mining,self-learning,continuous optimization and decision evaluation.Finally,some thinking and prospects for the intelligent control technology of’self-learning’launch vehicle are presented.

关 键 词：运载火箭 智能控制技术 思考与展望 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]