Determination and analysis of the optimal prediction method of GNSS system time offset

Currently, GNSS compatibility and interoperability are one of the most significant problems in multisystem navigation and position. If the prediction precision of system time offset is enhanced, the multi-system user's requirements can be satisfied, about positioning, timing, and so on. And the development of GNSS compatibility and interoperability will be promoted. To increase the prediction precision of system time offset, this essay studies three prediction algorithms of system time offset(Kalman filter, end-point and quadratic polynomial). And then, the prediction precision is analyzed. Firstly, based on research on the feature of GNSS system time offset, three kinds of prediction principle are analyzed. And secondly, the initial values of predictions are determined, in order to reduce the uncertainty and error caused by it. Finally, test three prediction methods with real-time measurement data, and then the optimal method is ensured. Additionally, the real-time measurement data are provided by a platform that monitors GNSS system time offset at National Time Service Center (NTSC). The result shows as follows. Firstly, for all types data, the precision of real-time prediction is best. And the precision of Kalman filter is higher, approximate to 0ns. Meanwhile, the prediction consistency of quadratic polynomial is best. Finally, when the results of all the prediction parameters are summarized, the Kalman filter prediction methods is optimal.