Improved Grey Model and Application in Real-Time GPS Satellite Clock Bias Prediction

It is difficult to hold the performance of space-borne GPS satellite atomic clock because of the property of high-frequency, sensitivity and impressionable, it is accord with that of grey model (GM), we can look on the process of satellite clock bias (SCB) as that of grey system.Real-time and reliable SCB prediction is a key to implement real time GPS precise point positioning (PPP). Firstly, based on the limits of quadratic polynomial (QP) and conventional GM, a modified GM is put forward, and taking GPS SCB for example, analyzed SCB prediction with different sample interval, the relationship of GM exponent coefficient (EC) and prediction precision, comparison precision to that of QP; finally, to test the reliability and validation of the modified GM, we analyzed the prediction precision with the modified GM, It is showed that the modified GM is validation to predict GPS SCB.