Recent advances in technology have contributed to the deployment of a "de-facto" large GNSS receiver array based on affordable smart devices that are easy to find in the consumer market (dual band smartphones, raw GNSS data recording, new sensors). 

These devices, evolving fast with each new generation, feature an increasing number of capabilities and sensors able to collect a variety of measurements and provide improved GNSS performance. Among these capabilities, Galileo dual band smartphones receivers and Android’s support for raw GNSS data recording represent a major step forward for PNT data processing improvements.

Meanwhile, the growing number of IoT GNSS devices, the expansion of groundbased networks of fixed GNSS receivers and the availability of GNSS spaceborne receivers has resulted in an increasing availability of GNSS products (GNSS Big Data).

Effective crowdsourcing of GNSS data is also flourishing across different science disciplines (e.g. space weather, water vapour measurement, geo-hazard detection). 

Crowdsourced GNSS Big Data repositories provide a unique opportunity to apply innovative Machine Learning techniques to characterise multiple error sources, providing a unique opportunity for identification of singularities and correlations, particularly across the above GNSS-related disciplines.

The objectives of the proposed activity are:

• validate the application of Machine Learning, Big Data and Data Mining techniques to vast amounts of data resulting from diverse GNSS data sources for a set of use cases (e.g. improved tropospheric / ionospheric characterisation and earth's magnetic field mapping);
• set the basis of a unique infrastructure for the identification of singularities and correlations across GNSS related domains (e.g. space weather, water vapour measurement, geo-hazard detection, among others);
• crowdsource and integrate GNSS data also from ground-based IoT sources (e.g. smartphones, smart-cities and wild life tracking devices) and space-based ones, much beyond classical GNSS data collection networks.

The tasks to be performed will include:

• survey and assess existing and potential providers for GNSS crowdsourced data to complement accessible space-based data repositories (e.g. the GNSS Science Support Centre located at ESAC);
• define the implementation approach for a set of use cases, focussing on the areas of Fundamental Physics, Metrology, Earth and Space Science; assess state-of-the-art Big Data Machine Learning models to be applied to the selected use cases;
• implement and validate a GNSS Big Data repository and the Machine Learning algorithms for the selected use cases, deploying community engagement measures to ensure crowdsourcing support;

 
The main results of the activity will provide a better characterisation of received GNSS signals and an improved understanding of GNSS-related scientific domains. 

Results from previous, related ESA activities, such as NAVISP El1-008 “Weather Monitoring based on Collaborative Crowdsourcing”, will be duly considered and assessed. 
 
The proposed activity is supported by ESA’s GNSS Science Advisory Committee (GSAC)

Point of contact:
Benedikt Soja <soja@ethz.ch>