Research On Data Processing Method Of Dynamic GPS Precise Positioning

Research and analysis on data processing problems of high precise differential positioning using dual frequency dynamic GPS data are disscussed this thesis. The main work and innovation focus on following aspects: including computation of satellite position, error correction, cycle slip detection, lack of observables, ill-conditioning model, development of software and so on.The method of the computation of satellite position using different ephemeris is profoundly investigated, this paper analyse the precision of interpolation method using polynomials of different orders. For alleviating the Runger phenomenon in extrapolation of satellite position, methods of broadcast ephemeris selection that focus on the interpolation of two points are put forward and analysed, at the same time, Chebyshev fitting method is applied for fitting the broadcast ephemeris of navigation satellites for computation efficiency. The impact of the fitting and interpolation methods of ephemeris on the result of differential positioning is intensively analysed. The common methods of cycle slip detection and correction are systematicly compared. Thealgorithm comprehensively makes full use of pseudorange/carrier-phase combination, ionosphere residual combination and M-W combination is put forward to detect and correct cycle slip. Therefore, the shortcomings of the three combinations are remedied, and the actual detection results are better, all cycle slips are detected. Epoch differential location model is used to construct the error-equations, and standard residual is used for cycle slip detection. Compared with the other methods, the method avoid the influence of the measurement errors of position, detect cycle slip in limited time, which leads to better position accuracy, and is able to determine cycle slip. Moreover,the method makes it possible to detect cycle slip just in two epochs.Aiming at the lack of observables, new solution for differential positioning of GPS, which constructs the interpolation observation equations among epochs, is put forward and validated by experiment. The experiment result shows that the interpolation accuracy is in accord with differential solution accuracy that is with presence of observables. The positioning accuracy of different observable combinations is analysed, which is under conditions of different baseline length and considering characteristics of different observable combinations. The software that is based on this paper automatically select optimal observable combination according to baseline length for the solution of position.Analysis on the influence and causations of ill-conditioning is carried out. The method using DOP to diagnose ill-conditioning problem is theoretically deduced, which is combined with other common methods. Ridge estimation is proposed to solve ill-conditioning model problem. Proved by theory deduce, the results of ridge estimation is better than least square method with suitable parameter. The selection of the parameter approximation value is discussed, and the method to calculate prior value by curve fitting is put forward. The influence on positioning precision of different approximation value is experimental analysed,and the experiential Ridge estimation prarmeters is given, positioning accuracy after Ridge estimation is analysed otherwise.Methods of short-term prediction of ionosphere delay using the TEC data released by IGS is analysed. The Wavelet Neural Network on predicting ionosphere is proposed for the whole world. and the experiment result shows better accuracy.Software bases on differential positioning of GPS theory is developed, the algorithms which are put forward in this paper are added in corresponding software modules. Therefore, the precision accuracy of the software is validated and analysed under different sample rates and static or dynamic states.