相关数据的Fisher信息增益及在导弹试验鉴定中的应用(英文)

Fisher Information Gain of Data with Applications in Test Evaluation of Missile

在数据处理的应用领域 ,观测数据对系统参数估计提供了多少信息量是一个令人关心的问题 .基于共轭正态分布 ,利用统计推断的结论 ,研究了相关的观测样本可以提供多少统计信息的问题 .此结论可应用到航天测控领域中的试验鉴定中 .可以充分利用制导工具误差造成的弹道主动段的遥外弹道差数据 ,借助折合算法 ,将弹道差数据转换为关机点数据 .再通过关机点和落点偏差的关系得到相应的落点偏差 ,通过度量相关观测点提供的Fisher信息量作为权重 ,融合这些落点偏差信息 。

In the applied fields of data processing, the measurement of statistical information gain when adding samples is important. Based on common normal (conjugated) distribution, this paper investigates the relation between the Fisher information gain and the addition of sampled data relevant to previous samples. By transforming relevant value of data into added weight, the relevant sample can also be used in statistical inference. This conclusion is applied to ballistic missile test evaluation by converting trajectory difference between telemetry data and trajectory tracking data, which is mainly produced by guidance instrumentation systematic error of missile, into impact point deviation for solving evaluation problem of small-sample impact point in practical test despite of using simulation.