A new regional tropospheric delay correction model based on BP neural network

Tropospheric delay is an important error source in the Global Navigation Satellite System (GNSS) positioning, navigation and timing. The global empirical models are unable to provide sufficient accuracy for the precise positioning with the increasing demand of precision. In this paper, we utilize the UNB3m model to determine accurate temperature, pressure and relative humidity which can be used to calculate zenith total delay of local area, and use GA-BP model to correct the residual errors by taking the estimated tropospheric delay from BERNESE 5.2 as a reference, then develop a new regional tropospheric delay model of “UNB3m+GA-BP” based on BP neural network. The numerical example by using Hong Kong GNSS data shows that the UNB3m+GA-BP model has improved the accuracy obviously compared to the UNB3m and GTP2 model. The accuracy of the proposed model is about 1.1 cm without systematic error. UNB3m+GA-BP model can better describe the spatial variation of regional troposphere and is suitable for real-time regional tropospheric delay correction.