SBAS Interference

April 1, 2009 By: Christophe Macabiau, Guillaume Lamontagne, J.C. Guay, Jean-Luc Issler, Laurent Lestarquit, Marie Nouvel-Malicorne, Matthieu Sihrener, Olivier Julien, Olivier Nouvel, René Landry Jr. GPS World

Worst-Case Scenario

GPS C/A codes present cross-correlation peaks that can potentially cause false acquisitions. They can also create tracking errors and C/N0 degradation, not in traditional GPS tracking situations, but for applications where signals are received with a low Doppler difference.

Several applications present a low relative Doppler, thus causing more frequent Doppler collision and this type of interference. One occurs with the use of ranging signals from geostationary satellites (for example, satellite-based augmentation systems, or SBAS) when two or more of these satellites are in line-of-sight of a fixed- or low-dynamic receiver such as a control station, reference station for differential positioning, or any ground user with little or no motion. The tracking error increases strongly with more geostationary satellites — and multiple SBAS have been and continue to be developed with overlapping interoperative regions. Indeed, it is possible that all the signals will interfere at the same time, or sequentially, one after another. The SBAS C/A code interference could also be seen for mobile receivers such as airplanes, moving along a zero-relative Doppler line between two SBAS satellite longitudes.

Pseudolite use can also induce low dynamic signals, in applications involving a pseudolite and a receiver onboard a geostationary satellite, and pseudolites and fixed- or low-dynamic receivers.

The interference mechanism is physically the same as for multipath error. Actually, multipath can be viewed as C/A code self-interference: the cross-correlation peak of the reflected signal contributes to the correlation function and causes an offset of the zero-crossing of the discrimination function, known as the tracking error. In C/A code interference, the cross-correlation peak of the interfering code creates the same effect. Normalized cross-correlation peak values are either +63/1023 or –65/1023. The error can reach 18 meters for a half-chip spacing correlator when two quasi-stationary C/A codes are received with the same power level. If the cross-correlation function of desired and interfering codes has a secondary peak within the receiver's chip spacing of the relative code offset, there is potential for a tracking error.

A necessary condition for such interference is a relative Doppler between the two interfering signals lower than the receiver code loop bandwidth, typically lower than 1 Hz. The signed amplitude of the interfering peak is a function of the relative phase, and if the relative phase moves too much during the code-loop integration period, the interfering peak (and the tracking error) will be filtered out. We refer to this condition as a Doppler collision, or as having a quasi-stationary code, because from the receiver point of view, the received codes will not move relative to one another. This means that the cross-correlation peak of the interfering signal will not move relative to the main auto-correlation peak of the desired signal, and the interference will result in a lasting bias.

For common GPS applications, a Doppler collision smaller than 1 Hz is quite a rare phenomenon. Doppler due to satellite motion is high (from –4.5 kHz to 4.5 kHz for standard use), giving a low probability of such a Doppler collision; even if it happens, it will last a few seconds. For common GPS applications, C/A code interference rarely leads to a noticeable tracking error.

However, with the development of GNSS systems using geostationary satellites such as EGNOS, WAAS, GAGAN, MSAS, BEIDOU-1, or IRNSS, the dynamics are much lower, in the range of a few meters/second, corresponding to a few 10s of Hz at L1 frequency. Doppler collisions will occur twice a day and last several minutes. Thus, interference due to PN-code code cross-correlation will occur, and can cause false acquisition, C/N0 degradation, and tracking errors. Our results using data collected by WAAS receivers in Canada show the impact of several factors, including the navigation message, on the tracking error.

The current EGNOS inhibits the GEO ranging function, and therefore pseudorange measurement errors made on the GEO signal have no impact today on EGNOS performance. Moreover, in more than 10 years of operation, including the EGNOS test bed, no message loss due to the C/A code interference has been observed or reported on EGNOS. Our work here intends to: