Resilient timekeeping algorithm with multi-observation fusion Kalman filter

Abstract

Abstract The timescales incorporated into the Primary Frequency Standard (PFS) exhibit excellent stability and accuracy. However, during the dead time of PFS, the reliability of the timescale can be compromised. To address this issue, a resilient timekeeping algorithm with a Multi-observation Fusion Kalman Filter (MFKF) is proposed. This algorithm fuses the frequency measurements from hydrogen masers with various reference frequency standards, including PFS and commercial cesium beam atomic clocks. The simulation results show that the time deviation and instability of the timescale generated by MFKF are improved compared to those with Kalman filtering. The experimental results demonstrate that even within 70 days of PFS dead time the resilient timescale generated by MFKF can operate reliably. Furthermore, it is theoretically proven that MFKF produces a smaller post-covariance than that with single-observation Kalman filtering.