Assessment of the effect of GNSS sampling rate on GNSS/INS relative accuracy on different time scales for precision measurements.

• The abstract were revised to clearly explain the proposed improvement method. • The general role of Allan variance in the application of GNSS/INS was revised. • The matrices used in the optimal estimation were explained in detail. GNSS/INS relative accuracy representing navigation error variation or instability on different time scales has been increasingly required in precision measurement. However, research on the effect of GNSS sampling rate on GNSS/INS relative accuracy is limited. We mainly focus on the assessment of the effect of GNSS sampling rate on GNSS/INS relative accuracy on different time scales through Allan variance, and propose to improve the relative accuracy by adjusting GNSS sampling rate and applying backward smoothing. The field test results show that the time-correlation of GNSS measurement noise caused by high-rate sampling should be dealt with; increasing the GNSS sampling rate can improve the short-term position accuracy in the real-time mode; the backward smoothing can mitigate the weakness of low-rate GNSS sampling in the post-processing mode. This work can provide guidance for the reasonable selection of GNSS sampling rates to meet the demands of GNSS/INS relative accuracy. [ABSTRACT FROM AUTHOR]