An orbit determination algorithm by means of the satellite-borne GPS data and Kalman filter

Using the Kalman filter algorithm, we have processed on-board GPS data of Shenzhou 4. The research focuses on three problems, namely, the selection criteria for the model error variance matrix of the Kalman filter, the effects of GPS signal interruption or runs of outliers on the recursive filtering, and the method to monitor the filter running status (normal or divergent). The aim is to evaluate the reliability of long-time stationary running of this algorithm used for on-board autonomous orbit determination