基于灰色时序模型的激光陀螺漂移数据滤波

Filtering of RLG′s drift data based on grey-time series model

为了有效抑制机抖激光陀螺(RLG)输出数据中的随机漂移,提出了采用新陈代谢GM(1,1)灰色模型与时间序列模型融合的灰色时序建模新方法。依据所建模型对激光陀螺的漂移数据进行Kalman滤波,并利用Allan方差法对建模滤波前后的陀螺数据进行分析对比。结果表明:该方法抑制激光陀螺随机漂移的效果优于传统的时序建模后Kalman滤波的方法,有效降低了激光陀螺的量化误差、角度随机游走、零偏不稳定性、角速率随机游走、速率斜坡;相对于传统方法,对量化误差的改善尤为明显。

In order to restrain the random drift of mechanically dithered RLG′s output data effectively,a new method of grey-time series modeling was proposed,which has integrated the metabolic GM（1,1） model and time series model.Kalman filter was used to filter the RLG′s drift data based on the model which was built by using the new method.The data of mechanically dithered RLG before and after modeling and filtering was analyzed by the method of Allan variance.The results show that the effect on restraining the random drift of RLG′s output by using the new method is better than that of traditional time series modeling and succedent Kalman filter.The new method decreases all random error terms in RLG effectively,which include quantization noise,angle random walk,bias instability,rate random walk and rate ramp.Compared with the traditional method,the meliority on quantization error of the new method is obvious.