| CPC G06F 1/08 (2013.01) | 6 Claims |

|
1. A real-time sliding ultrashort-term forecast model algorithm based on frequency data and phase data, executed by a processor of a computer apparatus, and the real-time sliding ultrashort-term forecast model algorithm based on frequency data and phase data comprising:
converting clock error phase data from a satellite navigation system into the frequency data;
iteratively processing the frequency data to eliminate frequency outliers from the frequency data, comprising:
calculating a standard deviation value of the frequency data, wherein a calculation formula for calculating the standard deviation value of the frequency data is as follows:
![]() wherein sigma represents the standard deviation value of the frequency data, ave represents an average of the frequency data, fi represents one of the frequency data, and n represents a total number of the frequency data;
eliminating a target frequency data fm0 farthest from the average ave of the frequency data from the frequency data to satisfy a condition fm0=MAX(fabs(fn−ave)) to thereby obtain updated frequency data, and updating the sigma based on the updated frequency data to obtain an updated sigma;
designing different thresholds according to the updated sigma and different satellite orbit types;
determining whether fabs(fm0−ave) is greater than a corresponding one threshold of the different thresholds;
in response to fabs (fm0−ave) being greater than the corresponding one threshold of the different thresholds, considering that fm0 is a frequency outlier; and
continuing iterating a next frequency maximum point fm1 until no epoch is beyond a corresponding threshold;
performing a real-time sliding clock error forecast to update a forecast epoch, comprising:
calculating a root mean square error of fitting residuals as a basis for setting a threshold range; and
determining whether a fitting residual of the forecast epoch is beyond the threshold range;
in response to the fitting residual of the forecast epoch being beyond the threshold range, eliminating data of the forecast epoch; and
in response to the fitting residual of the forecast epoch being not beyond the threshold range, sliding forward by one epoch, and performing fitting and forecast again; and
using a real-time clock error forecast value obtained through performing the forecast to correct a clock frequency deviation, for thereby improving forecast accuracy, real-time performance and data stability of a satellite clock error.
|