一种飞行参数缺失数据的预测估计方法  

An Estimation Method of Lost Data for Flight Parameters

作　　者：张玉叶[1] 李明珠[1] 刘建东[1] 张国栋[1] 

机构地区：[1]海军航空工程学院青岛分院,山东青岛266041

出　　处：《测控技术》2015年第7期42-45,共4页Measurement & Control Technology

摘　　要：飞行参数数据在记录过程中受到各种干扰因素影响,容易产生部分数据的丢失。为了得到完整又符合实际飞行状态的飞行参数数据,需要对缺失数据的参数数据进行预测估计。飞行参数数据中包含十几个参数,若将所有参数用于每个动作的飞行评估,复杂度太高。考虑到不同飞行状态下,不同的参数相关度不同,采用主成分分析选择相关度较大的若干参数。这些参数数据构成多元时间序列,因为飞行参数具有连续性和平稳性,所以利用多元时序建模综合多个参数相关性进行建模分析,实验证明该方法能够准确地估计缺失数据。Influenced by some kinds of interference in the process of recording,the flight data are easy to lose.In order to obtain the time series data which can reflect the practical flight state,the lost flight parameter data should be predicted and estimated.There are a dozen parameters in flight data.The computation complexity will be too high if all parameters attend to evaluate fight operation.Given that the relevance between different parameters of different acrobatic maneuver is different,principal component analysis is adopted to choose some parameters which have higher relevance.And the data of chosen parameters constitute a multivariate time series.Because the flight parameters have continuity and stability,multivariate time series model can be used to do modeling and analysis,combined with the relevance of some relevant parameters.The test proves that the method can improve the accuracy of lost data estimation.

关 键 词：飞行参数数据 主成分分析 多元时间序列建模 数据估计 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]