Gps Standards Of Fitting The Model Error Analysis And Research

In global positioning system localization technology, as a result of its survey localization technology physical mechanism, its planimetric position precision may achieve the very high standard, but its elevation precision compares and plane precision approximately two time. In particular in WGS84 coordinate system to 1954 Beijing coordinate system (in 1980 Xi'an coordinate system), 1956 Yellow Sea elevation system(or 1985 Yellow Sea elevation system), because the WGS84 ellipsoidal height only has the geometry significance but the physical connotation, and the Yellow Sea elevation system normal height both has the geometry significance, and has the earth interior mass density distribution non-uniformity this kind of physical phenomenon. Therefore, at present this kind of transformation is one of international domestic mapping educational world positive exploration popular topics.The widespread application height us the GPS standard in the project practice. Mainly is through the GPS standard fitting in the region of the elevation exceptionally distributes, transforms the ellipsoidal height which measured by GPS to normal height. Its precision mainly is decided by the accuracy of the GPS standard elevation exceptionally fits of the model.At present because the GPS standard fitting model research is insufficient, lacks the contrast research and the precision evaluates analysis on the many kinds of mathematical model. This article according to the reference model error related theory, has inferred function model optimal criterion based on the model error. This article proposed in function fitting model optimal before pushes using the resistance of corvee fitting estimates, Initially judges the data point which includes the thick difference, Screens the optimization the fitting spot, causes the fitting result to avoid the thick almost pollution, thus causes the fitting result resistance of corvee, the precision to be higher.This article uses the MATLAB language to carry on the programming to the multinomial fitting model and the multi- surface function fitting model. Through calculated example has carried on to the above model based on the fitting residual error and based on the mean error comparison. Only has when the parameter numbers of differently fits function to be same, can directly compare the residual error size, otherwise should act according to the smallest of the mean error for the selection most superior model standard. Optimizes in the fitting model foundation in this article carries on the model error compensates, introduced two models errors compensating process, namely the attachment model straggling parameter smallest two rides and the take the model error as the stochastic parameter two rides the disposition, Causes the fitting model to be more precise.The most superior choice of GPS standard fitting model, the fitting model parameter determination as well as the model error compensate will be the key to increases the GPS standard fitting precision.