Securing Autonomous Air Traffic Management: Blockchain Networks Driven by Explainable AI
Authors:
Louise Axon,
Dimitrios Panagiotakopoulos,
Samuel Ayo,
Carolina Sanchez-Hernandez,
Yan Zong,
Simon Brown,
Lei Zhang,
Michael Goldsmith,
Sadie Creese,
Weisi Guo
Abstract:
Air Traffic Management data systems today are inefficient and not scalable to enable future unmanned systems. Current data is fragmented, siloed, and not easily accessible. There is data conflict, misuse, and eroding levels of trust in provenance and accuracy. With increased autonomy in aviation, Artificially Intelligent (AI) enabled unmanned traffic management (UTM) will be more reliant on secure…
▽ More
Air Traffic Management data systems today are inefficient and not scalable to enable future unmanned systems. Current data is fragmented, siloed, and not easily accessible. There is data conflict, misuse, and eroding levels of trust in provenance and accuracy. With increased autonomy in aviation, Artificially Intelligent (AI) enabled unmanned traffic management (UTM) will be more reliant on secure data from diverse stakeholders. There is an urgent need to develop a secure network that has trustworthy data chains and works with the requirements generated by UTM. Here, we review existing research in 3 key interconnected areas: (1) blockchain development for secure data transfer between competing aviation stakeholders, (2) self-learning networking architectures that distribute consensus to achieve secure air traffic control, (3) explainable AI to build trust with human stakeholders and backpropagate requirements for blockchain and network optimisation. When connected together, this new digital ecosystem blueprint is tailored for safety critical UTM sectors. We motivate the readers with a case study, where a federated learning UTM uses real air traffic and weather data is secured and explained to human operators. This emerging area still requires significant research and development by the community to ensure it can enable future autonomous air mobility.
△ Less
Submitted 27 April, 2023;
originally announced April 2023.
