多项式信号加自回归噪声模型的参数估计

The Parameters Estimation of Polynomial Signs with AR Noise

本文提出一种新的多项式信号加自回归ＡＲ噪声的组合模型统计处理方法．方法采用形式处理和分离参数的技术克服了测量数据中平稳相关噪声ＡＲ序列未知的难点，避免了传统差分方法中残差数据非零均值和非ＡＲ属性的缺点，相应算法为线性求解，简便、实用、运算速度快、求解精度高，可同时获得数据真实信号的估值及相应的误差估计。本方法可直接用于飞行器电子测量数据的实时处理。

A new AR mixed model statistics method dealing with electric measure data of aircraft is presented. The method overcomes the difficulty that the AR noise in observation data is unknown. The algorithm is linear and has fast computer speed and so can be applied to the timed data processing. The simulated results show that the new method has higher accracy than the usual difference method.