论文部分内容阅读
针对高精度光纤陀螺(FOG)随机误差,在分析时间序列相关函数的基础上,得出其时间相关性呈较弱现象,若采用传统方法辨识模型,则容易造成误判。在此情况下,使用适用性更广泛的AIC准则分别确定AR及ARMA模型的最佳阶数,继而采用卡尔曼滤波算法,最后以滤波后零偏稳定性改善情况来选择作为FOG随机误差的数学模型。验证结果表明:ARMA模型和AR模型均适合作为FOG随机漂移数学模型,且均通过模型适用性检验,而ARMA模型效果更佳。
Aiming at the random error of high precision fiber optic gyroscope (FOG), based on the analysis of correlation function of time series, it is found that the time correlation is weak. If the traditional method is used to identify the model, it is easy to cause misjudgment. In this case, we use the more generalized AIC criterion to determine the optimal order of AR and ARMA models respectively, and then use Kalman filter algorithm. Finally, we choose the mathematical error of FOG as the improvement of the filtered bias stability model. The verification results show that both ARMA model and AR model are suitable as FOG random drift mathematical models, and both of them pass the test of model suitability, while ARMA model is better.