Basile, Francesco
(2019)
Multi-frequency GPS/Galileo combinations for quick Precise Point Positioning in urban environments.
PhD thesis, University of Nottingham.
Abstract
High-end applications, such as surveying, geodesy, and mapping, require positioning accuracies in the order of centimetres or even millimetres. Global Navigation Satellite Systems (GNSS) enables such high accuracy levels by exploiting very precise carrier phase measurements. While, in the past, high accuracy GNSS was dominated by differential techniques, such as Real Time Kinematic (RTK), in recent years the interest in Precise Point Positioning (PPP) has considerably grown. Thanks to its high computational efficiency, no need for any synchronous measurements from nearby reference stations, and homogeneous positioning quality on a global scale, it is widely used for many applications characterized by open sky conditions, e.g. precision agriculture. Unfortunately, the main downside of PPP, i.e. its long (re-)convergence time, strongly limits its use in constrained and transient signal environments associated with urban areas.
The modernisation process of the GNSS will positively affect the PPP solutions thanks to the higher number of satellites in view broadcasting multi-frequency ranging signals. Indeed, previous studies identified the number and geometry of the observed satellites as one of the factors impacting on the PPP (re-) convergence time. It was observed that, when multi-constellation GNSS measurements are employed, the PPP solutions are greatly improved, especially in particularly masking environments. At the same time, many publications showed that triple-frequency observables benefit ambiguity fixing, but not the float solutions.
In this PhD thesis, the PPP float solutions based on multi-frequency GPS and Galileo measurements recorded on-board kinematic platforms in urban areas are investigated, in terms of precision and convergence time. In addition, the specific benefit of the Galileo AltBOC modulated signal E5 on PPP is also analysed. This research has been conducted in five stages.
Firstly, GPS+Galileo PPP solutions were compared against GPS only, and Galileo only PPP solutions. The current state of the Galileo constellation and its PPP products, as well as of the GPS next generation constellation, is still not mature for a clear evaluation of multi-frequency multi-GNSS PPP. Hence it was decided to employ simulated data.
The second part of this research is aimed to identify the GPS and Galileo two-frequency ionosphere-free combinations that guarantee the best accuracy and convergence time in PPP. Both for GPS and Galileo, this is the combination between L1/E1 and L5/E5a.
Then, the GPS, Galileo, and GPS+Galileo PPP solutions were compared in open sky conditions. In a multipath-free environment, the positioning solution based on Galileo measurements, on average, converges to an accuracy of better than 10 cm in 2.5 minutes horizontally and 6.5 minutes vertically. As a comparison, it takes 4 and 10.5 minutes to have an average error, respectively, in the horizontal and vertical GPS PPP solution smaller than 10 cm. This has been justified by the lower noise in the Galileo pseudoranges. On the other hand, in a multipath-rich site, where the quality of the GPS and Galileo pseudoranges is comparable, GPS offers better PPP performance. For example, the horizontal RMS error in the GPS PPP solution converges in 20 minutes, while the average convergence time for the horizontal Galileo solution is as large as 33.5 minutes. In all cases, the GPS+Galileo solutions are more accurate and take less time to converge below 10 cm of error than the single-constellation configurations. Indeed, in a multipath rich site, the RMS horizontal error in the GPS+Galileo PPP solution takes only 16.5 minutes to go below 10 cm and reaches a final accuracy of about 9.6 mm. In contrast, the horizontal accuracy in the GPS solution is about 1.3 cm.
The fourth part analyses the performance of GPS+Galileo PPP in urban environments. Although, there are areas where satellite availability is good enough to compute a GNSS positioning solution, the slow PPP (re-) convergence may not allow high accuracies. In the scenario that was simulated, once the receiver loses track of most of the satellites, the horizontal GPS+Galileo PPP solution never re-converges below 50 cm for 95% of the time. Nonetheless, the results showed that GPS+Galileo PPP easily enables sub-meter level horizontal accuracy for most of the time.
Finally, to fulfil the needs of higher accuracy applications, more refined positioning algorithms, aimed to reduce the PPP (re-)convergence time, are proposed. These are based on triple-frequency ionosphere-free combinations. Unlike the most successful methods that can be found in literature, they do not rely on any external ionospheric information. Therefore these algorithms, not
exploiting the spatial correlation in the ionospheric delay, but only the temporal correlation, for a given level of ionosphere activity, guarantee homogeneous performance with increasing distance of the receiver from the network of reference stations that was used to compute the PPP corrections. In urban areas, while the horizontal GPS+Galileo PPP solution based on the traditional ionosphere-free combination is not able to re-converge below 20 cm, in the first method, a re-convergence time of a few tens of seconds can be obtained by employing smoothed-ionosphere corrected, low-noise pseudorange combinations. The second approach uses triple-carrier ionosphere-free combinations that minimize the noise in the ambiguity observable to shorten the first convergence of the float Galileo PPP horizontal solution from 20.43 minutes, adopting the two-frequency ionosphere-free combinations, to 8.58 minutes.
Actions (Archive Staff Only)
 |
Edit View |
Tools
Tools
![[thumbnail of Thesis_final.pdf]](https://eprints.nottingham.ac.uk/style/images/fileicons/application_pdf.png "Thesis_final.pdf")