两种抗差估计在陀螺信号处理中的比较

Comparison of two robust estimations applied to gyro signal

捷联寻北仪中影响寻北精度的主要因素是陀螺随机漂移。由于陀螺信号中不可避免地存在异常观测或粗差,陀螺随机漂移中包含异常噪声(异常数据)。抗差估计通过连续降权和连续减弱影响的方法抵制异常数据,是一种控制异常数据的有效方法。本文给出了两种抗差估计解及其相应原则,利用一个实际算例比较了两种抗差估计控制异常数据影响的能力。计算结果表明:在观测数据中存在异常数据的情况下,结合两种抗差估计各自优势的算法既减少了计算量,又获得了可靠的估计效果。

In the strap-clown gyro north seeker, the random drifts of gyro arc primary factors to degrade orientation accuracy. The random drifts of gyro include abnormal noises （ abnormal data） because abnormal observations or outliers inevitably exist in gyro signal. Robust estimation is an effective method to control abnormal data by descending weights and weakening effects of abnormal data continuously. In the paper, the formulas of two robust estimations were presented and the abilities of controlling abnormal data were compared. The calculations showed that the algorithm combining the advantages of two robust estimations could both decrease calculations and obtain reliable estimates on condition that the abnormal data exist in observations.