LEO augmentation in wide-area PPP-RTK positioning & time transfer

Abstract

Low Earth Orbit (LEO) satellites have been widely discussed for augmenting GNSS-based Positioning, Navigation and Timing (PNT) services. Their low orbital altitudes lead not only to a much stronger signal strength, but also a faster speed and geometry change that delivers a fast convergence of, e.g., Precise Point Positioning (PPP). Compared to PPP, the Precise Point Positioning – Real-Time Kinematic (PPP-RTK) enables the Integer Ambiguity Resolution using satellite phase biases processed by the network. The additional ionospheric information provided by the network is often responsible for the fast IAR and convergence of the positioning results. For wide-area PPP-RTK, the interpolation of ionospheric delays becomes challenging. In the Undifferenced and Uncombined (UDUC) model, the users thus turn to the so-called ionosphere-float mode and estimate the ionospheric delays by themselves. This leads to a long convergence time, and the positioning results become sensitive to the observation model strength, i.e., the number of GNSSs used, the static/kinematic mode of the users, and the IAR fixing rates.

In this study, we show that the LEO augmentation enables fast convergence for wide-area PPP-RTK positioning without needing any ionospheric corrections and interpolations. Using a sampling interval of 30 s and a network of 1000 to 2000 km, the 1-sigma percentile lines for both the horizontal and up coordinates can converge to 1 dm in 1-2 min in both the static or kinematic modes, applying both GPS-only or multi-GNSS scenarios, and in both the ambiguity-fixed and float modes. The fast convergence is insensitive to the observation model strength. In addition to that, the 90% percentile lines of the horizontal and up coordinates also quickly converge to 1 dm in 2-4 min, indicating the benefits of LEO augmentation to the integrity of the positioning.

In addition to positioning, this study also introduces the method for long-based time transfer based on UDUC wide-area PPP-RTK. For a baseline of 800 to 900 km, a 7-day standard deviation (STD) of about 0.06 ns was achieved in the static mode, and the Modified Allan Deviation (MDEV) is below 5e-16 at an averaging time of 1e5 s. With the LEO augmentation in addition, the convergence time is greatly shortened and becomes insensitive to the observation model. One can expect a fast convergence to 0.5 ns within 1-2 min in both the kinematic and static mode, using single- and dual-GNSS observations, and in ambiguity-fixed and -float modes.