Research on AT1 Algorithm Improvement Based on Prior Information

The real-time clock synchronization control system is based on the reference of the atomic time scale, making the performance of the atomic time scale a crucial factor influencing the results of frequency control. The Atomic Time Scale (AT1) algorithm is a short-term iterative method for calculating the time scale, known for its good real-time performance. However, when applied to calculate the time scale for atomic clock ensembles with significant frequency drift, the AT1 algorithm may exhibit divergent trend terms, leading to a deterioration in long-term stability. In response to this issue, this paper proposes an Improved AT1 Algorithm based on Prior Information (IAAPI). The article first utilizes system prior observations, employing a least squares polynomial fitting to correct initial values. Subsequently, in the dynamic iterative process, the prior information is used to dynamically adjust frequency changes, forming a forecasting model. The results are then obtained accordingly. The paper compares these results with Weight Average Timescale (WAT) and conventional AT1 algorithms. The findings indicate that the IAAPI algorithm reduces the fluctuation range of the time scale compared to the AT1 time scale, with a significant decrease in trend terms. Moreover, long-term stability is significantly improved. In comparison to the WAT algorithm, the short-term fluctuation is smaller, and the long-term and short-term stability of the resulting time scale surpasses that of the WAT time scale.