采用分段法估算Allan方差中的各噪声系数

Estimation of Various Noise Factors from Allan Variance by Subsection Method

指出采用传统的最小二乘拟合法估算存在于Allan方差中各噪声系数存在的问题,提出采用分段法来估算Allan方差中的各噪声系数,并通过对多组实测陀螺数据进行分析,证明了分段估算法与最小二乘拟合法相比是一种更有效的、科学的噪声估算方法。提出了一种根据数据点的多少自动改变Allan方差积分时间的算法,可在保证估算精度的前提下减少不必要的计算过程,有效地缩短了Allan方差的计算时间。

We found that the estimation of various kinds of fiber optic gyro noise factors from Allan variance by using the least square fitting method has some problems. The subsection method was brought forward to estimate errors from Allan variance and several sets of gyro data have been analyzed. The analyzed result showed that the subsection method was more effective and scientific than the least square fitting method. In addition, an algorithm which can change the integral time of Allan variance automatically based on the length of consecutive data points was proposed. Using this method can reduce some unimportant calculations and shorten Allan variance calculating time effectively, as well as guarantee Allan variance calculating and noise estimating precision.