Using the Kalman filter to detect frequency jumps in atomic clocks

We propose an algorithm for the detection of frequency jumps in atomic frequency standards. The algorithm is based on the innovation generated by the Kalman filter. When the magnitude of the innovation is larger than a given threshold, an anomaly is detected. We characterize the performance of our method for the case of a Rubidium clock by estimating, through numerical simulations, the probability of false alarm, the probability of detection, and the latency time on a large number of numerical simulations.