Robust Kalman Filtering Model for Dynamic Data Processing of GPS Monitoring Networks

In order to overcome the influence of gross errors in observation vectors on the filtering values of state vectors, based on the laws that the gross errors influence the state vectors and the characteristic that they are shown completely in residual forecasting, a robust Kalman filtering model was derivesd. This model has good robusticity to observation space and design space. According to the computation results of a simulative GPS monitoring network with gross errors,and compared with the results of the standard Kalman filtering model,this robust Kalman filtering model can obtain reliable results of deformation analysis.