捷联惯性组合误差模型相关性分析方法  

Correlation analysis method for error model of strapdown IMU

作　　者：刘璠[1] 魏宗康[1] 

机构地区：[1]北京航天控制仪器研究所,北京100039

出　　处：《中国惯性技术学报》2014年第5期567-571,共5页Journal of Chinese Inertial Technology

基　　金：民用航天专业技术预先研究项目(D010101);国防基础科研项目(B030080021)

摘　　要：在实际应用中通过捷联惯性组合误差模型可以将测量值转换成载体导航解算的需要量,而误差模型中的误差系数一般通过标定试验获得。以往认为误差模型中各误差系数相互独立,但通过实际计算可以发现并非如此。为了考量各系数相关性关系,提出了一种相关性分析方法。首先,提出了捷联惯组加速度计输出方程。之后,运用统计学的相关方法计算出了各系数的相关系数和协方差矩阵,并推导出了测量值和误差系数的分布特性之间的关系。最后,通过加速度计组合的实际测试,分别计算了不包含和包含拟合残差的统计数值。通过与输出值进行比对,验证了相关性分析方法的正确性,并表明拟合残差对于统计特性的计算具有重要作用。By using the error model of strapdown inertial measurement unit (SIMU), the measurements can be translated into the needed variable, and the error coefficients in the error model are usually obtained by calibration tests. In the past, the parameters were thought to be statistically independent, but it is proved wrong by our calculation. In this paper, a method for analyzing the correlation of parameters is studied. Firstly, the error model of the accelerometer of IMU is given. Then the matrix of covariance and correlation of parameters are calculated out with the knowledge of statistical method. The distribution relation between the measurements and error parameters is deducted. At last, with a real test of IMU accelerometers, the statistic values with and without fitting residues are computed, respectively. The method is proved to be valid by comparing with the statistic value of measurements, and the fitting residues are shown to play an important role in the calculation of statistical properties.

关 键 词：捷联惯性测量组合 误差模型 相关性 协方差矩阵 

分 类 号：U666.1[交通运输工程—船舶及航道工程]