Dr Simon Birnbach - University of Oxford

UK IC Postdoctoral Research Fellowships 2024

A key challenge for maritime domain awareness (MDA) is automatically detecting and classifying ships. Due to limitations of the two technologies underpinning MDA — the automatic identification system (AIS) and synthetic aperture radar (SAR) satellite imagery — current solutions are inaccurate and often require manual intervention. AIS is a cooperative tracking system that provides precise and frequent updates of a vessel’s location and identity. But, due to the cooperative nature of AIS, ships can choose to stop participating in the system or even falsify messages that mask their location or identity. As a non-cooperative system, SAR can detect ships even if they want to remain hidden. However, SAR has its own disadvantages, including low image resolution and infrequent updates of each location.

This project will design a system that securely fuses SAR and AIS data, enhancing the performance of both technologies without increasing the attack surface. The basis for this system will be a thorough security analysis that systematises attack behaviour, as well as possible ways for an attacker to avoid detection or otherwise mislead the system. I will use this security analysis to design an attacker-aware system that understands its own limitations regarding commonly confused vessel types, and the impact of data quality on classification. My system will assign confidence scores to detections and classifications, allowing the system to autonomously determine when additional data sources are required. Ship detections derived from SAR data will then be used to cross-verify data obtained from AIS. When recent SAR data is unavailable, the system will increase trust in AIS messages based on their RF signal characteristics. Specifically, I will design techniques that: (1) use RF transmitter fingerprinting to verify if the signal is originating from a legitimate transmitter, and (2) leverage RF location verification methods to test the validity of AIS position claims.

Personal website

LinkedIn