Abstract: At present, the residual chisquare detection method based on extended Kalman filter has been widely used in GPS satellite fault detection independently by the receiver. However, this method has some problems such as dependency on mathematical model and detection delay. A GPS satellite fault detection method based on wavelet analysis was proposed. The wavelet analysis was used to analyze the GPS data in the time and frequency domain, and the data could be processed as multiscale analysis. Abnormal points were used to determine the occurrence of failure. The simulation results show that the proposed method is efficient, sensitive, concise and easy to implement. It′s useful to the reliability and stability of navigation system.

Key words: wavelet analysis, GPS , fault detection, extended Kalman filter