Research On Interpolation Algorithm For GPS Tropospheric Delay Based On GA-BP

There are many factors that influence the positioning accuracy by using GPS. Tropospheric delay is one of the main errors of GPS positioning and there is a close relationship between the tropospheric delay and the pressure, temperature, humidity and other meteorological factors of the station. There is a close relationship between the tropospheric delay and the angle of satellite signal’s path and when altitude angle is 90 degrees the zenith tropospheric delay error caused by is up to 2-3 m and when altitude angle is 10 degrees the error is caused by up to 20 m. With the requirements of engineering, the demand on the precision of GPS positioning is higher and higher. When the height difference of short baseline is large or the meteorological elements are flexible, the residual tropospheric delay is still the main error source of precise surveying by using the difference method. If you use the GPS tropospheric delay correction model, meteorological parameters of the station are needed and it is inconvenient for us. The static data of ten AHDLCORS stations are chosen as the training data of GA-BP model and the data of the eight remaining CORS stations as testing data. A new interpolation method of zenith tropospheric delay is established by using GA to optimize the weights and threshold value of BP neural network. The input parameters are the latitude, longitude and altitude.According to the data, when the training data are selected to the right geographical position, the accuracy of GPS tropospheric delay can be up to mm and it is higher reliability and stability when using GA-BP model. The results show that GA-BP model is a method of relatively high prediction accuracy in the prediction of regional zenith tropospheric delay.