GNSS Phase Positioning Without Ambiguity Resolution

The aim of this work is to introduce an accurate positioning method relying on GNSS carrier-phase measurements at high rate. In this approach, the position of an antenna with respect to a reference antenna is estimated using the GNSS carrier-phase difference between the two antennas. We show that this phase difference evolves following a multiple linear-circular regression model. The parameters of this model are the three coordinates of the rover antenna in the reference antenna frame. Then, we propose a regression estimator of these parameters that is shown on synthetic data to reach the Cramér-Rao Lower Bound. This estimate does not require phase ambiguity resolution. An experimentation on real data shows that the proposed method allows reaching the millimeter accuracy for static positioning using GPS-L1 signals only and with integration times of a few milliseconds.