陀螺寻北仪中测量信号的在线建模与滤波

针对光纤陀螺寻北仪中光纤陀螺(FOG)和加速度计的随机误差,采用改进型二阶自回归AR(2)模型,在线建立了光纤陀螺和加速度计随机误差模型。根据该模型,建立了FOG陀螺寻北仪的12阶Kalman滤波器,实现了两个FOG和两个加速度计测量信号在寻北过程中的实时滤波。仿真、Allan方差分析与寻北试验结果表明:FOG信号中随机游走、零偏不稳定性、变化率随机游走、变化率斜坡和量化噪声五项噪声源误差系数都小于滤波前的二分之一;在减小光纤陀螺和加速度计测量信号中的随机误差,提高其精度的同时,FOG寻北仪的寻北误差减小了0.3 mil。

On-line modeling and filtering of measured signals in FOG north-seeker

Aiming at the random errors of FOG and accelerometer in the FOG north-seeker,an improved AR(2) model is designed,and by which the random error models of FOG and accelerometer are built in real time.Based on the models,a 12-order Kalman filter is set up,and the random errors of two FOGs and two accelerometers are filtered in the FOG north-seeker's north-seeking process.The results of simulations,Allan variance analysis and north-seeking experiments show that these error coefficients,including the random walk,the bias instability,the change rate's random walk,the change rate's ramp and the quantification noise in the output signals of FOG;are about one-half as less as those when the random errors of FOGs and accelerometers are not filtered.In addition,the random errors of FOGs and accelerometers are reduced and their precisions are effectively improved by this modeling and filtering method,and the north-seeking error of the FOG north-seeker is reduced by 0.3 mil.