Satellite navigation (GNSS) signals provide accurate and continuous position, navigation, and timing (PNT) services to citizens. These services are necessary to critical infrastructures, such as accurate timing for stock market, electricity transmission, banking and security information systems, reliable positioning for aviation, wireless communications, accurate localization for emergency (112) personnel and logistic chains, and transport.

There will be an outage in critical services when GNSS signals are unavailable in the area. To prepare Finland to be resilient against such a mass disruption, the quality of navigation signals should be continuously monitored. Any anomalies should be immediately detected, isolated, and reported to authorities so that back-up solutions can take over seamlessly.  We will explore the potential of emerging Machine Learning methods (such as Deep Learning) for GNSS to renew the concept of GNSS situational awareness.

The Finnish Geospatial Research Institute has developed a GNSS-Finland concept for GNSS fault detection. Signals are recorded at FinnRef permanent GNSS reference stations across Finland to detect the general location of any disruptions. Using the big data acquired with GNSS-Finland, we will develop completely new Deep Learning (DL) methods to compute future trends to detect signal anomalies, assess the continuity of location information, and forecast critical failures in positioning and timing information. If a disruption is eminent in the immediate future, monitoring is enhanced by temporary stations, and back-up options are triggered to continue uninterrupted PNT.

For the improved resilience of timing service, robust back-up to critical time provision systems will be developed. Back-up solutions for positioning have already been widely studied, but this is one of the first studies to investigate the timing backup.