Datawise is a project between NATS and Cranfield University investigating a baseline proof of concept methodology for using machine learning methods and applying them to understand how data exchange between the air and ground can be optimised using federated learning approaches. Federated learning offers a novel approach to explore how data can be exchanged between actors in an ecosystem and this project is applying this in the context of the future Air Traffic Management ecosystem.  

The Air Traffic Management (ATM) “ecosystem” has a number of players - current airspace users, air traffic control providers and management entities. This will evolve in the future as new airspace users enter, with a corresponding increase in the amount of data exchanged, collected, stored, processed and used to address operational needs.  

To prepare for how the ATM ecosystem will evolve with new airspace users, increasing capacity and greater integration, we need to understand how data would be exchanged and managed from an air traffic control perspective.  There may be innovative ways in which data can be exchanged in the future between all the various players, and how the expected increase in the volume of data can be managed to optimise operations and leverage the capabilities of the infrastructure.  

The project involves investigating suitable machine learning algorithms which could be applied to effectively handle increasing data volumes transmitted with the ATM ecosystem (including aircraft, drones, future airspace users, ground-based air traffic control systems/tools) to use the communications infrastructure optimally.  

While this project is mainly exploratory, it will aim to produce a baseline methodology and early proof of concept prototype which can be used for further research in an air traffic management context. 

The project was one of the chosen projects to receive support from NATS at the first Research Collaboration Conference in 2019.