【摘要】 动力调谐陀螺仪受陀螺仪自身工艺及各种电气因素影响,其随机漂移往往表现出非线性性质,利用线性随机建模方法并不能表现出随机漂移的非线性。作者通过传统的24位置测试获得动力调谐陀螺仪的静态漂移数据,然后对数据进行预处理(如,去除野点,进行数据平滑并利用小波分解提取趋势项),提取出随机漂移数据,并采用非线性AR时间序列方法对处理后的这些漂移数据进行建模。模型适用性检验结果表明,所建立的非线性AR模型可以很好地拟合动力调谐陀螺随机漂移,适用于描述动力调谐陀螺仪随机漂移特性。更多还原

【Abstract】 Random drifts of the dynamical tuned gyro (DTG) commonly show non-linear due to the influences of the gyro’s technical process and various electrical factors,thus its linear random model can not present the nonlinear random drifts.To establish a fitful model,the drift data of DTG are sampled by traditional 24-position test,and then by pre-processing these data,such as removing outliers,smoothing,and extracting trend terms using wavelet decomposition,the random drift data are acquired.Then NAR model is established using these random drift data of DTG by nonlinear auto-regression (NAR) time series method.The model fitness test result shows that the NAR model is suitable for presenting the features of DTG random drifts.更多还原