Real-Time Atomic Clock Anomaly Detection and Processing Based on Generalized Likelihood Ratio Test

Abstract

As a core part of the navigation satellite payload, the performance of spaceborne atomic clocks directly decides the accuracy of GNSS Positioning Navigation Timing (PNT) service. Through continuously monitoring of clock bias, fast anomaly detection and processing can be realized, and thus, ensuring that satellite navigation system can meet the requirements of various end users. The paper combines principles of Generalized Likelihood Ratio Test (GLRT) with polynomial clock model based on clock types, and then a new method of real-time atomic clock anomaly detection and processing was put forward: firstly, anomaly detection was carried out based on the real-time GLRT detector, then short-term clock bias forecast and extrapolation was conducted by using the polynomial clock model, and based on steps above, eliminate the anomaly points and replace it with extrapolation values. Simulation examples was given based on GPS satellite clock bias provided by IGS, the results show that the method can detect and handle atomic anomalies fast and effectively (especially abnormal frequency jumps), and it can guarantee the integrity of clock bias.