Modeling and Predicting Mixed Errors of Shipboard Measurement Equipments

To have a better understanding of the characteristics of measurement errors of shipboard space tracking equipments so as to improve some traditional data processing methods, cubic spline function is applied as a separating tool to extract the errors from complicated measurement data. By taking traditional time sequence model technology and wavelet frequency division theory into account, based on the feature that errors can be divided in frequency domain, wavelet package decomposition and threshold operation are applied to frequency division modeling and forecasting for measurement errors, thus presenting a non stationary error forecasting method based on wavelet analysis. Simulation and actual data processing results show: wavelet prediction method possesses such features as higher precision, fewer errors, better fitting the real time measured values. Compared with simply applying time sequence prediction method, it proves to be much more efficient which can better predict the mixed non stationary errors in measurement data of space tracking equipments.