基于相空间重构和流形拓扑结构的飞参段划分  

Division of flight parameters based on phase space reconstruction and manifold topology

作　　者：曲晋瑶 徐志凌 窦海学 唐义号[1] QU Jin-yao;XU Zhi-ling;DOU Hai-xue;TANG Yi-hao(Product Division,China Helicopter Design and Research Institute,Jingdezhen 333001,China)

机构地区：[1]中国直升机设计研究所产品部,江西景德镇333001

出　　处：《沈阳航空航天大学学报》2021年第5期73-82,共10页Journal of Shenyang Aerospace University

基　　金：国家自然科学基金(项目编号:61903263)。

摘　　要：针对现有飞参阶段划分研究方法中模板和标签数据获取代价高、数据内在属性研究不够深入、算法参数过多等问题提出一种基于相空间重构和流形拓扑结构的飞参阶段自适应划分方法。该方法首先使用C-C法将反映阶段特征的一维高度数据重构到高维相空间;然后根据不同阶段飞行数据在相空间中呈现出不同流形结构的特点,采用局部切空间排列算法得到重构混沌吸引子轨迹拓扑结构;经处理后,不同阶段数据在拓扑结构中沿不同轴向紧密排列,进而完成阶段划分。通过真实飞机多次连续飞行数据对该方法进行验证,结果表明,该方法可以实现不同轨迹下飞参阶段的自适应准确划分。The Adaptive Partition Method for Flight Parameter Stage(APMFPS)is proposed against the problems of low generalization and high training cost in the current flight phase division research.The method first utilizes one-dimensional height data reflecting phase characteristics to complete phase space reconstruction(PSR),which aims to restore the original system dynamic characteristics.Then,the value of the maximum lyapunov index(MLI)of the reconstructed sequence at different stages is calculated to describe the chaotic characteristics at each stage quantitatively.Finally,different flight phases are distinguished according to the numerical difference in MLI of reconstruction sequences at each stage.Multiple real-life flight data from the aircraft is exploited to verify the proposed method.The result shows that the APMFPS can identify the flight data at different stages quickly and accurately,with strong practicability and generalization.

关 键 词：飞参阶段划分 相空间重构 局部切空间排列 流形拓扑结构 自适应准确划分 

分 类 号：V241[航空宇航科学与技术—飞行器设计]