Abstract：Due to the deficiency of the traditional MAD method in detecting the gross clock error, we proposed a method of detecting and processing the gross error of clock data based on wavelet algorithm. By using the multi-scale analysis capability of wavelet transform, the data with gross errors are decomposed into the low-frequency wavelet coefficients and the high-frequency wavelet coefficients of each layer, and the gross errors are detected and eliminated at different time scales. Using the BDS clock bias data provided by CODE (Center for Orbit Determination in Europe), we analyze the performance of different wavelet functions and decomposition scale. Compared with the MAD method, we find that after preprocessing using wavelet analysis, the data has obvious advantages in clock prediction, and the average prediction precision increases by 10%.
