Research on An Improved Time Scale Algorithm of Time-keeping Atomic Clocks

In this article, quadratic exponential smoothing combined with VC filter, a new time scale data fusion algorithm is proposed. The frequency drifts of hydrogen masers are predicted by the use of the minimum deviation method, so that the time difference prediction becomes more accurate. The time scale is constructed by the modified quadratic exponential smoothing method based on the data of hydrogen masers, and the cesium atomic clocks time scale is established by the weighted mean method, and the VC filter is designed to combine the advantages of long-short-term stability of the two time scales to improve the comprehensive time scale performance. The experimental results demonstrate that the proposed approach produces a comprehensive time scale has a stability of 1.60e-15 for one hour and 3e-15 for one day, and it is superior to the performance of time scale generated by three classic methods: ALGOS, AT1, and Kalman filtering.