Error Model Research Of GPS Based On QAR Data

GPS is one of the important navigation ways of civil aircraft currently, but it can't undertake the aircraft navigation task independently because of the single point positioning accuracy is not high and the differential positioning is difficult to achieve global. Although GPS/IR integrate navigation system has realized the high precision positioning, but the structure of integrated navigation is complex and the cost positioning is high. The high precision GPS independent navigation system being the key of the research. Analyze the error characteristics of the airborne GPS, and set up the error prediction model is helpful to improve GPS positioning precision, promote the development of the airborne GPS. QAR records a large amount of positioning data in the process of actual flight, that provided a large number of reliable sample data for setting up the aircraft GPS error model.Firstly, according to the characteristics of GPS positioning error and QAR data presenting establish time series model of GPS error based on QAR data.Secondly, calculating error data based on the center line of the route and QAR.We find that error data consists of GPS positioning error and the actual flight path deviation by analysis its features. GPS error data was extracted by removing the track error based on least square.Thirdly, error data are checked as non-stationary time series using unite root,data smooth processing has been realized through the difference. The GPS error increment model was established based on the improved AIC and least-square method.Finally, considered the poor prediction accuracy of ARIMA, the combined model of original signal and noise signal of error were set up by using the method of wavelet denoising and wavelet hierarchical reconstruction. The residual analysis confirms the prediction accuracy of the combined model has been greatly improved compared ARIMA model.