ION GNSS 2012
Session D1: Emerging GNSS (Galileo, COMPASS, QZSS, IRNSS)

Title: Ionospheric Delay Estimation Using Galileo E5 Signals Only
Author(s): O. Julien, Ecole Nationale de l´Aviation Civile, France; L. Lestarquit, J-L. Issler, Centre Nationale d´Etudes Spatiales, France
Date/Time: Wednesday, September 19, 2012, 8:57 a.m.
Room: Grand Ballroom West (Renaissance)

Future Galileo open signals, E5 (E5A/E5B) and E1 OS, were designed so that they can bring significant improvements to most of the users compared to the current GPS L1 C/A signal performances. Receivers will thus be able to track the different signals with a lower tracking noise and a lower multipath susceptibility, and an increased resistance to interferers, consequently providing cleaner code and phase pseudorange measurements. This enhancement was obtained thanks to, among others, the use of higher code chipping rates (10.23 MHz for E5A and E5B), innovative modulations (BOC, ALTBOC, MBOC) and the use of a pilot channel in parallel with the traditional data channel. The use of the 3 Galileo open signals together can bring further obvious improvements such as (1) a more accurate and robust ionospheric delay estimation, (2) improved ambiguity resolution performances (in terms of success rate and time to fix), (3) potential tropospheric delay estimation, and (4) frequency diversity against potential intentional or unintentional jammers. These different points were backed up by many different investigations and papers from different user communities needing high precision and reliable positioning, showing a great interest in a triple-frequency Galileo (and GPS) receiver. Based on this triple frequency baseline, it is however important when it comes to sensitive applications, to consider degraded modes since it might impact the expected behavior of the receiver. A typical example is the loss of one frequency and it is thus important for a triple-frequency Galileo receiver to consider the loss of any of the E5A, E5B and E1 signal and its consequence on required performances. This article specifically focuses on the event of the loss of the Galileo E1 OS signal. This situation is of particular interest because it means that the receiver is left with measurements coming exclusively from E5A and E5B signals, which are spectrally very close and thus not ideal for high precision positioning. Many different figures of merit are to be investigated in this degraded mode scheme to fully assess how the receiver can cope without significantly losing any of its performance. However, this article will only focus on the ionospheric delay estimation using only E5A and/or E5B signals. The motivation behind this investigation is to show that for a triple frequency Galileo receiver, whatever the jammed band, it is always possible to estimate accurately the ionospheric delay affecting pseudorange measurements and thus keep the high accuracy positioning ability of the receiver. Moreover, an extension of this conclusion is the potential use of the E5 band alone for precise positioning applications. The authors have already presented initial results in 2009 related to this study that investigated the use of an ionospheric delay estimation process based on a Kalman Filter (KF). This KF was based on code and phase geometry free combinations (using Galileo E5A and E5B measurements), jointly with a simplified local model of the Vertical Total Electron Content (VTEC) to represent the ionospheric delay of any visible satellite. This simplified VTEC model was based on the estimation of the VTEC at the ionosphere pierce point relying on the estimation of 3 parameters:  the estimated VTEC at the user zenith  the estimated latitude and longitude VTEC gradients The initial results were promising since the ionospheric delay estimation error standard deviation was at the decimeter-level even for low elevation satellites. However, these results were obtained using simulations assuming that the true ionospheric delays could be represented using the Klobuchar model. The present papers provides 2 significant improvements to this earlier work:  the true ionospheric delays are now represented using the NeQuick model, that is able to more realistically represent strong VTEC variations  7 local ionosphere models (based on up to 5 states to estimate) have been tested in the KF to improve the estimation process.
The new results shows that the new local ionosphere models bring significant improvements in terms of robustness and accuracy of the estimation process. It is also seen that depending upon the user location (and mostly its dip latitude), different local ionosphere models are preferable.

In the first part, the article will present a first analysis of the typical VTEC variations, based on true ionospheric maps and NeQuick-simulated ionosphere maps, in order to understand what elements have to be considered in order to define a simple local ionosphere model. In the second part, the reference estimation process based on a KF using Galileo E5a and E5b measurements and a basic local ionosphere model will be described. Then, the new proposed local ionosphere models will be presented. Then the simulation tool will be described. Finally the different techniques will be tested and compared in order to elect the most promising one, with the ultimate goal of assessing its performance to maintain high precision positioning only based on Galileo E5 signals.

Previous Abstract Return to Session D1 Next Abstract