Home Applications Septentrio The application of machine learning enab…

The application of machine learning enables the prediction of EGNOS performance

The European Space Agency's (ESA) NAVISP programme has funded the development of a performance prediction model for EGNOS using machine learning techniques.EGNOS, or the European Geostationary Navigation Overlay Service, is the European Union's Satellite-Based Augmentation System (SBAS) that provides critical services to a wide range of domains, including aviation. This new model is designed to simulate the output of the EGNOS Central Processing Facility (CPF), monitor the integrity of the system and generate navigation information and corrections for users.

Mickael Dall'Orso, technical officer, mentioned that traditional EGNOS performance prediction models are simpler and based mainly on satellite and Range Integrity Monitoring Station (RIMS) geometry, whereas the new machine learning model takes into account more variables such as multipath differences and space weather. Bastiaan Ober, project coordinator at IntegriCom in the Netherlands, said that they were particularly interested in predicting the uncertainty in the Gridded Ionospheric Vertical Error Indicator (GIVEI) and the User Differential Range Error Indicator (UDREI). These metrics describe the state of the EGNOS estimated ionospheric delay and satellite health parameters, respectively.

The project partners, including IntegriCom, GeoX Hungary and Iguassu Software Systems Prague, were provided with 5.5 years of EGNOS data, including key variables such as space weather. They used 70 per cent of the data to train the models and 30 per cent to test the models.Ober concluded that the machine learning models significantly outperformed the standard macro models, improving the overall performance predictions of EGNOS, both in terms of UDREI and GIVEI values, as well as in terms of protection levels and availability.

In the field of high-precision GNSS positioning, Septentrio has developed strong support with its mosaic GNSS modules. mosaic series modules, such as the mosaic-X5 and mosaic-H, ensure high-precision positioning in a wide range of environments with their multi-constellation, multi-frequency tracking capabilities, as well as advanced anti-jamming and anti-spoofing technology, AIM+!
. These modules support all current and future GNSS satellite signals, making them a reliable choice for future signals and services. mosaic-H modules are particularly suitable for applications where directional capability is required, such as automation and navigation systems, and feature a dual-antenna design that provides precise heading and pitch angles or heading and roll angle.

Ober also suggested ways to improve the model, including the use of more accurate space weather forecasts, noting that there would be additional computational costs associated with the implementation of such a model, but that such costs should not be a limiting factor for applications such as aviation. He expressed his gratitude to ESA for helping them to understand more about how the EGNOS system works and for allowing them to use the high-end EGNOS simulator.

Stefano Binda, ESA NAVISP Element 1 Project Manager, mentioned that this is the third in a series of projects with the same title but different content. The fast growing global PNT market is a huge opportunity for the European economy and they need to make the most of it, which is why they will continue to support similar work. Meanwhile, Septentrio, a world leader in GNSS high-precision positioning solutions, has had its mosaic GNSS module integrated into NXP's V2X solution, which provides powerful centimetre-level positioning capabilities for advanced driver assistance systems (ADAS).

PREVIOUS：A trial of Jammertest 2024 is underway in Norway NEXT：The operation of SAMVA eVTOL flights at a Vertiport with EGNOS-enabled navigation capabilities