Estimation and analysis of GPS inter-fequency clock biases from long-term triple-frequency observations

Usually, the difference between the satellite clocks computed with L1/L2 and clocks computed with L1/L5 is defined as inter-frequency clock bias (IFCB). It is critical to correct its L5 time-variant portion in the GNSS triple-frequency precise positioning. Using two years of observations from more than 100 stations worldwide, we use the epoch-differenced method to estimate IFCB for all available 12 GPS BLOCK-IIF satellites, and analyze its short-term and long-term variations. The experimental results indicate that the IFCB variations are clearly consistent for two satellites located in the same orbital plane, which perhaps means that the variations of IFCB are dependent on the orbital plane. We found that the IFCB of each Block-IIF satellite shows repetition characteristics over two years. The annual repetition cycle of 352 days of IFCB is consistent with the GPS year 351.4 days may originate from the rotation of satellites around the earth. GPS triple-frequency uncombined PPP is carried out using 9 globally distributed MGEX stations from June 1 to 30, 2018. The experimental results indicate that compared to the PPP solutions without IFCB corrections, GPS triple-frequency PPP can achieve an accuracy of 2.2, 3.8 and 11.4 mm in the north, east, and up components after correcting IFCB, which is an accuracy increase in 31.3%, 17.4%, and 13.0%, respectively. The average RMS of the phase posteriori residuals for each frequency is also reduced significantly, especially 79.1% for L5 frequency.