Bradford Renduchintala Centre for Space AI develops system to combat air traffic cyber attacks
A team of Bradford researchers in the Bradford-Renduchintala Centre for Space AI has developed an AI-based system to combat cyber-attacks against aeroplanes and air traffic control.
Cyber attacks, including ransomware, malware and bot attacks, are considered a major threat to aeronautical systems the world over.
Prof Fun Hu, inaugural director of the Bradford Renduchintala Centre for Space AI, is leading a research team at Bradford to develop AI-based solutions to thwart such attacks.
Prof Hu said: “The project aims to define an intelligent and secured aeronautical data link communications network, based on software defined networking (SDN), augmented with AI, to predict and prevent safety services outages, to optimise available network resources and to implement cybersecurity functions to protect the network against digital attacks. The University of Bradford leads the Cyber Security work package in the SINAPSE project.”
Prof Hu recently presented the team’s work, Machine Learning and Cyber Security: Robust Protection Against Digital Attack, to the EU SESAR webinar ‘Innovative Solutions for ATM Resilience’, as a part of the SINAPSE project.
SINAPSE started in May 2020 and is funded under the EU H2020 Single European Sky ATM Research Joint Undertaking (SJU), with a remit to develop solutions to cyber-attacks directed at air traffic control and aeroplanes.
Prof Fun Hu, pictured above, added: “We developed a framework to tackle potential digital attacks on aeronautical communication networks using federated learning. While the prototype produced encouraging results, a lot still needs to be done to ensure the framework be efficiently and realistically applied to the real aeronautical environment.
“Currently we rely on open datasets to test our framework. The next stage of development will involve testing the algorithm on real traffic and attack patterns of the aeronautic communications applications.”
The Bradford team designed an innovative cyber security framework using collaborative machine learning, considering a hybrid-SDN architecture that maps onto the Future Communication Infrastructure, to provide digital and secure communication capabilities for integrated communication, navigation, and surveillance.
SDNs create and control a virtual network or control a traditional hardware through software, while traditional networks use dedicated hardware devices, for example router and switches, to control network traffic.
A prototype was implemented to assess the machine learning-based security mechanism using publicly available attack data. Preliminary assessment results showed good results.