Robust Estimation Of The Gps Data Processing

In recent years, GPS location measurement has become a major modern geodetic techniques with the rapid development of GPS positioning technology and universal. Many countries have established national and regional GPS geodetic control network by using of GPS technology. China has also established a high-precision GPS control network of the country. In addition, GPS in engineering surveying, topographic surveying, deformation monitoring and other aspects of applications are most widely used. In modern high-precision measurement, the accuracy of GPS data processing requirements are also getting higher and higher.In the GPS location measurement, as a result of observing the signal, the spread of the path and receiver parts, such as the impact of observation inevitably outliers, and the outliers or systematic errors affections are the main reasons for the accuracy of adjustment. At the same time, the correlation between the baseline should be carefully consided. However, in the GPS data processing, if there are some outliers in vectors for high-precision GPS network adjustment, how to handle these outliers of GPS data processing is the key to good or bad.When the measurement dates obey normal distribution, least squares estimation is the best unbiased statistics and it has been used most widely so far. If the measurement dates were interferenced by the outliers distribution, as the least squares estimation does not have the ability to anti-outliers, it is necessary to find the estimates with capacity of anti-outliers from least squares estimation theory and method. Therefore, the robust estimation was introduced to the filed of measurement data processing.First, this paper introduced the principle of GPS positioning technology and error sources, GPS network adjustment method and criterion of the adjustment method, the basic principles and adjustment model of robust estimation. On this basis, we expounded the robust estimation theory about the relevant observations including the function model, the relevant principles of observation. Then, introduced the applications of the language of C++in GPS network adjustment. Finally, designed two types of experimental program through an adoption of a GPS observation network to compare and analysis the different effects with different methods of estimation of the GPS network adjustment. At the same time, this paper put forward using the standard precision accuracy to correct the weight, then adjusted the GPS network. Applied this adjustment method to process measurement included outliers on the basis of equivalent variance-covariance. Function model with coarse poor adjustment of the GPS network would reduce the influnce of outliers to the estimated parameters. In addition, two-factor variance expansion method which variance inflation factor from the improved countdown of the weight function of the IGGⅢcheck program were used in adjustment of the GPS network.This method were proved feasible, and had a good capacity of resistance to outliers and stable efficiency of weight down.