论文部分内容阅读
Allan方差法分析随机误差时存在2点缺陷,一是长相关时间下估计值震荡较大,二是无法跟踪信号的动态变化。本文融合总方差法和动态Allan方差法的思想和优势,提出了动态总方差法。首先使用窗函数截取原始数据,然后对窗内数据进行延拓,对延拓后的数据进行总方差分析,得到原始信号的局部随机特性,随着窗函数的滑动可以得到原始信号的随机变化特性。经验证,动态总方差法同时解决了Allan方差法存在的2个缺陷。最后设计了半球谐振陀螺(HRG)的线振动试验验证此算法的有效性,分析结果表明动态总方差法在分析精度和使用的数据量上具有优势。但是动态总方差法又存在计算量大、分析速度慢的问题,因此本文又推导出了动态总方差法的递推公式,从而给出了动态总方差的快速算法。
Allan variance analysis of random errors, there are two defects, one is the long-term correlation estimates larger shocks, the second is unable to track the dynamic changes of the signal. This article combines the total variance method and the dynamic Allan variance idea and advantages, put forward the dynamic total variance method. Firstly, the window data is used to intercept the original data, then the data in the window is extended, and the total variance of the extended data is analyzed to get the local random characteristic of the original signal. With the sliding of the window function, the random variation of the original signal can be obtained . It has been verified that the dynamic total variance method simultaneously solves the two defects of the Allan variance method. Finally, the vibration test of hemispherical resonant gyroscope (HRG) is designed to verify the effectiveness of the proposed algorithm. The results show that the dynamic total variance method has the advantages in the analysis accuracy and the amount of data used. However, the dynamic total variance method has the problems of large amount of calculation and slow analysis speed. Therefore, the recursive formula of dynamic total variance method is deduced in this paper, which gives a fast algorithm of dynamic total variance.