Performance Analysis Galileo Satellite Clock And Clock Bias Prediction Algorithm

In various satellite navigation and positioning systems,the accuracy of time measurement determines the accuracy of navigation and positioning.The on-board atomic clock is not only an important part of navigation satellites,but also maintains the overall operation of the system.Analysis of the performance of the on-board atomic clock can grasp the operation status of the satellite clock,and it is also an important prerequisite for improving the accuracy of the clock difference modeling prediction.With the advent of passive hydrogen atomic clocks,it is considered to be the most suitable spaceborne atomic clock in the space environment.The passive hydrogen clock carried by the Galileo satellite navigation system has been in operation for 2 years.It is of great significance to study the long-term performance,and it can also provide reference and reference for the passive hydrogen clock of the Beidou three generations of satellites in China.In this paper,the Galileo system clock bias products are used to systematically study and analyze the long-term performance of the passive hydrogen clock and the clock difference pre-processing and forecasting methods.?1?A clock difference preprocessing method based on sparse Bayesian algorithm is proposed to recover the missing clock data.By setting an independent hyperparameter for each weight,the sparseness of the model is increased.It not only avoids the phenomenon of over-learning in machine learning but also reduces the use of kernel functions and shortens the learning time.The effectiveness of the method for restoring missing data is verified by simulation analysis of the missing data of different clock errors.?2?The study analyzed all passive hydrogen atomic clocks carried by the Galileo system.Changes in four indicators of residual error,frequency accuracy,frequency drift rate and frequency stability by clock difference model.It was found that the model residuals of the IOV satellite and FOC type were kept between 1 and 2 ns.The accuracy is in the order of 10^-12,and the individual satellites reach the order of 10^-13.The frequency drift rate of the IOV satellite is on the order of 10^-17,while the FOC satellite is mostly in the order of 10^-16.The stability of IOV and FOC satellites has reached the order of 10^-14,and its stability is better than 10,000 seconds and stable in thousands of seconds,indicating that the long-term stability of the Galileo passive hydrogen clock is better.?3?An improved grey model satellite clock error prediction algorithm is designed.The algorithm can avoid the process of identifying parameters by the background value of the traditional gray model.And introduce a new variable,reconstruct the initial value,and improve the prediction accuracy.The short-term and long-term prediction performance of the model is verified by the clock data of different clock types.The results show that the model is more suitable for the prediction of passive hydrogen clock and cuckoo clock,but the prediction accuracy of cuckoo clock is not high.The accuracy of long-term forecasting is more superior,and the accuracy is improved by about 20%compared with the traditional gray model;The short-term forecast performance is comparable to that of the traditional gray model.When the clock data fluctuates greatly,the forecast accuracy is better than the quadratic polynomial model.The influence of different modeling data quantities on the accuracy of improved model prediction is analyzed.It is found that the forecasting accuracy of Cuckoo Clock decreases with the increase of modeling data.When the ratio of modeling data to the amount of data to be predicted is 25%³5%,passive hydrogen clock and cuckoo clock have the highest prediction accuracy.