PROJECT ID

SINAPSE

PROJECT TYPE

Exploratory research

FLAGSHIP

Not applicable

STATUS

Completed

SESAR PROGRAMME

SESAR 2020

PROJECT DURATION

2020-05-01 > 2022-10-31

TOTAL COST

EUR 853 300

EU CONTR.

EUR 853 300

GRANT ID

892002

PARTICIPANTS

Altys Technologies, Frequentis, Enaire, University of Bradford

Safe air traffic control and capacity management rely on careful planning of complex airways and user preferences.  Artificial intelligence (AI) and machine learning (ML) can assist human decision-making and help to optimize airspace configuration, tactical planning, and post-operation analysis. 

SINAPSE research focussed on the aeronautical communications network (ATN), the global infrastructure used for safety communications between aircraft and air traffic controllers (ATC), airlines or manufacturers. The ATN  comprises all systems that assist aircraft from departure to landing. To enhance this resource, the project developed an intelligent and secure ATN  network architecture design based on a software-defined networking (SDN) architecture model augmented with artificial intelligence. This was used to train federated AI models for the benefit of the whole ATN community and enabled data to be shared widely — while respecting data privacy. 

The intelligent internet protocol (IP) network improved air traffic control awareness by predicting safety services outages: For example, the trained ML model was able to predict events that disrupt ATN controller pilot datalink communications (CPDLC), such as provider abort issues, occurring today in Europe on the current ATN network. As the proposed methodology is agnostic from the underlying network technology, the AI-driven monitoring techniques are future-proofed to support upgrades. 

The network also employed ML to predict the probability of transmission errors in satellite links used for voice and data telecommunication services with air traffic control. Additionally,  the ML uses deep neural network (DNN) algorithms and federated learning and strengthens cybersecurity by the use of intelligent intrusion detection and prevention (IDPS) applications.  The collaborative cyber security model results in collective learning and the data footprint is shared among all the users. Data privacy is ensured as only the AI footprint is shared, thanks to the federated learning architecture.

 

Benefits

  • Connectivity
  • Rapid, automatic adjustment of services
  • Robust IP network for network aircraft
European Union
SINAPSE - Software defined networking architecture augmented with Artificial Intelligence to improve aeronautical communications performance, security and efficiency intermodal