| Business | Period | Project Coordinator | Client |
|---|---|---|---|
| Industry | January 2021 -December 2022 | RINA Consulting - Centro Svilluppo Materiali S.p.A. |
ESA - European Space Agency |
GNSS infrastructure has been growing significantly in recent years, in the space segment as well as on ground. Millions of Internet-of-things (IoT) devices, including smartphones, use GNSS for positioning.
Due to the large number of devices, IoT data offer great potential for GNSS science exploitations, with unprecedented spatio-temporal resolution. However, access to IoT data for scientific purposes is currently limited and the data processing challenging.
In the project “CAMALIOT: Application of Machine Learning Technology for GNSS IoT Data Fusion” the innovative software infrastructure involved data ingestion, processing and analysis service, implementing IoT components, Machine Learning (ML) models and pipelines. Through an online platform, users have direct access to GNSS Science Data from different IoT sources, benefitting from ML potential to deliver new products. With a modular design, the platform architecture focuses on ease-of-use and flexibility for seamless development and deployment of Big Data and ML pipelines, and adaptability to future changes in the fast-moving world of IoT.
The validation has been conducted through two use cases addressing Ionosphere and Troposphere characterisation. The feasibility of estimating ionospheric and tropospheric products starting from smartphone GNSS raw data was demonstrated with two experimental campaigns in Italy. Both types of products were validated against those derived from geodetic GNSS receivers, and, in the case of the ionospheric Total Electron Content (TEC), against state-of-the-art ionospheric maps.
In the framework of CAMALIOT, RINA was in charge of the following main activities:
The combination of IoT, Big Data and Machine Learning technologies in the GNSS fields is ready to implement new services and products for different applications fields.
Despite having been made accessible since 2016 on Google’s Android-equipped devices, raw GNSS data that are actually usable for scientific purposes (i.e. with carrier phase, and on two frequencies) are still difficult to access.
The feasibility of estimating ionospheric and tropospheric products starting from smartphone GNSS raw data was demonstrated, albeit only with the Xiaomi MI 8 Pro device.
The analysis related to the troposphere characterization underlined the importance of the use of meteorological data together with classic mathematical models for the achievement of a useful model.
The two performed experimental smartphones acquisitions campaigns validated the Big Data platform infrastructure and its modularity enabling the data gathering from a restricted set of devices and computing a forecast that is comparable with the actual state of art.
