LEO navigation systems have gained significant attention in recent years as an opportunity to enhance and complement existing GNSS systems relying on MEO satellites. To achieve these goals, the LEO satellites shall broadcast their ephemeris to allow the users locating them in the space at any point of time. These ephemerides are a parametrisation of the underlying orbits, and the nature of the LEO orbits is substantially different from the MEO one. Therefore, although the GPS/Galileo ephemeris model is a good starting point, it needs to evolve to cope with LEO orbit dynamics. This paper addresses the selection and justification of parameters to be included in the LEO navigation message, in order to ensure high-accuracy ephemeris.

To identify the relevant ephemeris parameters, the temporal evolution of the instantaneous LEO orbital elements has been analysed, characterising this evolution into linear, quadratic and harmonic trends, complemented with a Fourier analysis to characterise the harmonic frequency and power. Additionally, a software tool has been developed for LEO ephemeris computation. This tool is capable of fitting ephemeris to a given input orbit for different subsets of parameters, allowing to systematically test the most convenient combinations to minimise the ephemeris fitting error.

If the adequate parameters are added on top of the GPS/Galileo ephemeris model, the fitting error tends to reduce. Beyond the ephemeris parametrisation, other factors, such as the length of the fitting interval, significantly influence the achievable accuracy. As shown in the figure below, for a reference orbit at an altitude of 510 km, the best results in terms of Signal-In-Space Ranging Error (SISRE) at Worst-User-Location (WUL) are in the order of 1 cm for a 21-parameter set and a fitting interval of 10 minutes, or even close to 1 mm for a 5-minute interval and a 19-parameter set.

Multipath interference significantly degrades GNSS positioning in automotive applications, posing challenges for the safety and reliability of autonomous driving and ADAS systems. Employing a patented supercorrelation algorithm, the S-GNSS receiver [1], developed by Focal Point Positioning under an ESA NAVISP project, offers a unique solution to multipath interference in challenging automotive environments by filtering incoming signals by angle of arrival in software processing alone. This ensures only multipath-free line-of-sight signals are processed in the navigation solution.

In this work, we will present initial experimental results in urban and under foliage scenarios with an S-GNSS receiver, areas where traditional processing struggles to provide high accuracies.

Preliminary analysis reveals significant improvements in GNSS robustness and accuracy, hinting at the full potential of supercorrelation technology in a final ADAS solution. Ongoing investigations will quantify these gains under different conditions, with the full results to be presented at the conference. These findings pave the way for reliable GNSS navigation in the toughest automotive environments, ultimately propelling the safe and efficient deployment of autonomous vehicles.

[1] Garcia, J.G.; van der Merwe, J.R.; Esteves, P.; Jamal, D.; Benmendil, S.; Higgins, C.; Grey, R.; Coetzee, E.; Faragher, R. Development of a Custom GNSS Software Receiver Supporting Supercorrelation. Eng. Proc. 2023, 54, 9. https://doi.org/10.3390/ENC2023-15423

With the aim of enhancing autonomous orbit determination (OD) capabilities of lunar navigation satellites, this paper proposes a visual processing technique called Crater Matching, based on machine learning. Specifically, the paper analyzes the performance results obtained from different architectural implementations of convolutional neural networks (CCNs) and the hardware used to identify an on-board feasible solution. The research focuses on achieving real-time processing and on-board execution of the crater matching algorithm, ultimately enhancing the OD autonomy of navigation satellites in lunar environments

Most modernized civil navigation signals offer a data and a pilot channel modulated on the same carrier frequency (e.g. GPS L1C/L2C/L5, Galileo E1 OS/E5a/E5b, BDS B1C/B2a). While the data channel needs to be demodulated to obtain ephemerides, clock corrections, system time, etc., the pilot channel exhibits only predictable symbol transitions to facilitate tracking of the signal’s time of arrival (ToA). Nevertheless, receivers may choose to perform ToA tracking based on a combined tracking (X) of pilot and data channels, pilot channel only (P) or even data channel only (D). While the latter is rarely used in practice, X-tracking offers lower tracking noise than P-tracking, since it combines the power of both channels.

Recent studies have reported that receivers may experience different pseudorange biases depending on whether D-, P-, or X-tracking is used [1,2]. This causes problems for the harmonized usage of receiver network data such as from the International GNSS Service (IGS), since the network essentially needs to be partitioned into separate groups of stations depending on which tracking mode is used [1]. In particular, differential biases between the P- and X-tracking receivers will affect orbit determination and time synchronization (ODTS), hence many applications in the field of surveying, timing, and geodesy. The above references provide an exhaustive overview of differential data-pilot (D-P) as well as differential combined-pilot (X-P) biases of modernized GNSS signals. Measurements were conducted with a set of commercial GNSS receivers.

While such efforts greatly facilitate the usage of IGS network data, the causes and individual contributors for the D-P bias are not yet fully understood. Interestingly, the largest D-P biases (+/- 3 ns) have been observed for GPS L5 signals, which is surprising in view of the similarity of the L5 data and pilot power spectral density (PSD). Moreover, there is reason to assume that the bias is receiver-specific as well as satellite-specific. For the Galileo E1 Open Service, where pilot and data components have slightly different PSD, the reported D-P biases are smaller but still noticeable (up to 0.2 ns), but are not consistent across different receivers. Naturally, the receiver acts as a black box in measurement campaigns such as the above mentioned. Receiver parameters such as correlator spacing, front-end bandwidth, or other implementation details (e.g., type of discriminator) are not usually disclosed by manufacturers, but may have a notable impact on the observed bias.

In this work, we present a model for the D-P and X-P differential biases. Our preliminary results suggest that the observed biases can be mostly traced back to four parameters, two of which are satellite-specific and two of which are receiver-specific. At the satellite, an offset between data and pilot stream as well as the difference in effective chip duration between data and pilot channel were identified as the main causes for D-P biases. These findings are underpinned with measurements from DLR’s 30 m dish high-gain antenna located at Weilheim, Germany. At the receiver, front-end bandwidth and correlator spacing have a noticeable impact on the D-P biases. Record & replay of real-world data to a software receiver and commercial receivers will serve to validate the proposed model.

[1] O. Montenbruck, P. Steigenberger, J. M. Sleewaegen, “Data+pilot biases in modern GNSS signals”, GPS Solutions (2023) 27:112.

[2] J. M. Sleewaegen, F. Clemente, “Quantifying the pilot-data bias on all current GNSS signals and satellites”, IGS Workshop 2018, Wuhan.