Distributed AI for enhanced security in satellite-aided wireless navigation (RESILIENT)

Wireless navigation and satellite-based positioning solutions are important parts of information and communications technologies nowadays; not only do they enable improved and efficient location-aware communications, but they are essential components of sensing the environments, enabling context awareness, ensuring precise time-tags and geo-tags to big data, and supporting a variety of future services related to safety-critical applications such as terrestrial, maritime, and aerial transportation, as well as autonomous and automated transportation, eHealth, workers safety, etc. The space technology industry in Finland is currently developing at a fast pace; many of these companies are relying on GNSS signals and their utilization in various geospatial applications with significantly benefiting from improved AI-based security solutions and new location privacy approaches. 

At the same time, cyber attacks in satellite frequency bands (e.g., L and S bands) are increasing, such as spoofing, meaconing, and jamming, especially since cheap GNSS transmitters can be now emulated via Software-Defined Radio approaches and spoofers are now feasible to implement with of few hundred EUR. In addition, ensuring user location privacy is more challenging than ever, as many wireless communication applications nowadays are heavily relying on geo-tagged and time-stamped data.

RESILIENT will address the challenges of ensuring cyber-security and privacy for civilian location-based applications, offering increased robustness and resilience of satellite-based positioning signals to various types of intentional and non-intentional interference. The project leverages the latest wireless communication developments to enhance the field of satellite-based positioning and navigation. Specifically, we will develop hybrid Artificial Intelligence (AI) approaches and statistical model-driven solutions to address two major cyber-security issues which are interference (jamming and spoofing) detection and localization as well as to provide novel methods to ensure user location privacy in collaborative applications. The envisioned solution consists of a privacy-preserving distributed approach that leverages a multitude of collaborative users and will advance AI in the fields of federated learning and unsupervised on-line adaptation.

The results in this project would be beneficial in terms of enabling the development of safer automated driving solutions and of better disaster risk management systems, ensuring an improved resilience of Radio Frequency (RF) infrastructure relying on geo-tagged data, offering an increased protection of user location data privacy employed in a variety of location-based services, including autonomous driving, eHealth, Ambient Assisted Living, etc. The increase in robustness, security, and resilience is aimed to be achieved under power and energy constraints, such that the devised solutions can work on battery-operated sensors and mobile devices.

The Consortium team consists of Tampere University and University of Vaasa from Finland, and Northeastern University from Boston, Massachusetts, USA.

Keywords: navigation, location-based communications, sensing, GNSS, interference mitigation and localization, spoofer detection, cyber-security, privacy, federated learning, deep neural networks, artificial intelligence, heat maps, wireless localization, resilient positioning