Improving the Accuracy of GPS Positioning with an Exponential Stochastic Model

Due to the difference of received satellite´s observational error distributions relative to the receiver in satellite positioning, the least square estimation is not the best linear unbiased estimation and this causes bias in positioning result. The weight least square method is adopted in order to satisfy the best linear unbiased estimation and improve positioning accuracy. Weight factor model was constituted which has exponential relation to the signal intensity of received satellites. The coefficient of the exponential model was determined through actual test of the received data. In this study, the positioning accuracy of the weight algorithm with the exponential model was analyzed. Results show that the positioning bias is 27.51% of no weight positioning algorithm in max bias measure and the computation of positioning algorithm has almost no increase.