On the potential of undifferenced and uncombined GNSS time and frequency transfer with integer ambiguity resolution and satellite clocks estimated

The use of global navigation satellite systems (GNSS) has been a competitive way to provide high-precision and low-cost time and frequency transfer results. However, the traditional GNSS method, the precise point positioning (PPP), is usually based on the ionosphere-free (IF) combination, which is not flexible when applying multi-frequency scenarios. In addition, PPP relies on precise satellite clock products with an accuracy of tens of picoseconds, limiting the time and frequency transfer performance. More importantly, achieving integer ambiguity resolution (IAR) is challenging, which makes high-precision phase observations underutilized. To achieve a better time transfer performance, we must consider all those factors from the GNSS end. In this contribution, a new GNSS time and frequency model at the undifferenced and uncombined (UDUC) level is first derived. In the UDUC model, the satellite clocks are estimated together with other parameters, and the integer ambiguities are resolved in the double-differenced (DD) form for their reliable estimation. Our numerical tests suggest three major findings. First, with integer ambiguities resolved, the UDUC model with satellite clocks fixed showed a 20% to 50% improvement compared with the UDUC PPP model. Second, with IAR and satellite clocks estimated, the proposed UDUC model shows a 10%–40% improvement over the model with satellite clocks fixed. Third, with integer ambiguities resolved and satellite clocks estimated, GPS, Galileo, and BDS-3 all have the potential to achieve frequency transfer in the low-mid 10 ^{- 17} range for averaging times within one day.