Anomaly Clock Detection Algorithm for a Robust Clock Ensemble

摘要

The Anomaly Clock Detection Algorithm is intended to facilitate a monitoring system to better manage a clock system, insuring continuity of operations and limiting the expected degradation of performance due to contingencies and anomalies. A novel ensemble of five Rubidium Atomic Frequency Standards (RAFS) is proposed onboard the future Global Navigation Satellite System. All clocks are in phase and on frequency all the time with advantages of relatively simple, robust, fully redundant, and improved performance. It allows reliable anomaly detection and smooth switch-over of master and redundant clocks using the same clock model. The algorithm is designed on the basis of the clock output phase/frequency measurements to detect the 'Feared Events' such as phase jump, frequency jump, and frequency instability, which affect various atomic clocks. The proper detection strategy has been determined in view of the detection capability and computation efforts. Algorithm parameters, such as the sliding window width as a function of the time constant, and the minimum detectable levels have been optimized and verified as the trade- off of the detection capability, response speed, and detection confidence. The verification results under numerous test cases in the presence of anomalies that occurred in various RAFS, as well as the clock ensemble, are demonstrated.