Optimizing High Precision RTK/PPK GNSS Algorithms Using Real-World Data

Jean-Baptiste Lacambre, Tim Barford,Nicolas Oudart, Patrick Lieffering, Guirec Morvant, Benoit Guyot

Abstract:	In optimizing and qualifying a new RTK GNSS algorithm, there are typically three challenges to overcome: 1. Acquisition of sufficient, high-quality data in a variety of signal conditions (open-sky, partial open-sky, obscured, etc.) 2. Efficiently building a trustworthy ground-truth reference against which to compare the GNSS output 3. Determining the true double difference for each satellite visible by the receiver Those three challenges are almost never tackled altogether. Today, most RTK algorithms are assessed using the number of RTK fixes, or in the case of more well qualified solutions, the number of correct RTK fixes validated by an external reference, but the performance and reliability for that reference is almost never assessed. The challenge in optimization is that to do so one must not only understand the performance of the algorithm in its current state, but also understand what the component impacts are on the final performance. By addressing the three challenges listed above, this paper provides a methodology that overcomes those three challenges altogether, with a focus on challenge #3: determining true satellite-by-satellite double differences in order the better identify algorithmic impacts on system performance. The intent of this paper is to detail the collection of a robust database of multi-sensor data, in a variety of satellite conditions, under an assortment of dynamic conditions to make reference data easily accessible. This reference data provides the ability to calculate the true double difference satellite-by-satellite along a real-world dataset, rather than using simulated data. The second intent is to present insight to database usage for the optimization of GNSS algorithms. In order to accomplish this optimization, the paper explores machine learning methods, outlier detection improvements, and key performance indicators accessible from reference comparison, along with pointers to additional algorithm blocks which provide statistically significant contributions to improving real-world positioning performance.
Published in:	Proceedings of the 35th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2022) September 19 - 23, 2022 Hyatt Regency Denver Denver, Colorado
Pages:	1746 - 1759
Cite this article:	Lacambre, Jean-Baptiste, Barford, Tim, Oudart, Nicolas, Lieffering, Patrick, Morvant, Guirec, Guyot, Benoit, "Optimizing High Precision RTK/PPK GNSS Algorithms Using Real-World Data," Proceedings of the 35th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2022), Denver, Colorado, September 2022, pp. 1746-1759. https://doi.org/10.33012/2022.18500
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In