论文部分内容阅读
非平稳时间序列的状态空间建模技术被用于陀螺漂移分析。基于平滑先验概念，使用Ｋａｌｍａｎ滤波和ＡＩＣ准则拟合整体模型，时变系数自回归模型（ＡＲ）被拟合为状态空间模型，并且被用于时变谱估计，针对两类不同的漂移分别建立状态空间模型，所得结果更贴近实际情况，目前本计算分析只能离线分。对故障诊断可提供一些依据。
State space modeling techniques for nonstationary time series are used for gyro drift analysis. Based on the concept of smooth priori, Kalman filter and AIC criterion were used to fit the whole model. Time varying coefficient autoregressive model (AR) was fitted to the state space model and was used for time-varying spectral estimation. For two different types of drift respectively The establishment of the state space model, the results closer to the actual situation, the current calculation and analysis can only be offline. Fault diagnosis can provide some basis.