Sesar-02-2015 - Data Science in ATM
Specific challenge: The ambition of this topic is to explore the potential of the rapidly maturing techniques in complexity science and data science for the ATM domain. Slow adoption of new procedures and automation tools in ATM is an indication that system complexity is no longer mastered; a better understanding of interdependence and feedback mechanisms between components and of parameters that control the overall performance of the socio-technical ATM system of systems is urgently needed.
In addition, massively collaborative environments for information sharing as envisaged for the future ATM system entail many risks. Generic solutions are being researched in computer science which now need to be demonstrated for specific ATM applications.
Scope: The range of research that could be covered in this topic is broad and topics mentioned here are indicative. Research proposals may target SESAR concepts and systems currently being implemented or could address timeframes up to those of the Flightpath 2050 vision document.
Over the last four years, the SESAR-sponsored ComplexWorld network has brought together scientists from complexity science with ATM researchers. Collaboratively they identified four promising areas for complexity research: (1) characterisation of uncertainty in ATM, (2) detection of emergent behaviour, (3) resilient design and (4) development of non-classical performance metrics. In addressing these and building upon the relevant WP-E results, the focus should be on exploiting complexity science’s modelling and analysis techniques such as network theory or agent-based modelling, at the appropriate scales (in time and space).
The availability of ‘big data’ offers a range of research opportunities for data science in ATM. Focussing on automating the extraction of knowledge from raw, heterogeneous and incomplete sources, techniques such as data mining, visualisation, stream processing, learning or scalable analytics may dramatically improve decision support and performance monitoring. A multi-disciplinary approach involving both data and complexity scientist is encouraged.
The advent of concepts such as collaborative decision-making, trajectory exchange and, more generally, system-wide information sharing raise issues not present in the peer-to-peer architectures that prevail in the current ATM systems. Essential operational aspects of SWIM are covered in the mainstream SESAR work programme, but further research is needed to address, inter alia, collaboration in the presence of competition, ‘value’ or ‘optimal amount’ of information, availability and security of cloud architectures in the context of SWIM and the future evolution of the European ATM system.
Expected impact: Successful research in this topic will improve the quality and availability of knowledge for decision making (both strategically and in real-time), will lead to more agile ATM system designs, to a more secure and widely usable SWIM infrastructure and offer benefits for exchanging passengers/freight information seamlessly between different transport modes. These developments will enable a range of performance benefits across the entire ATM system.
Type of action: Research and Innovation Action (RIA).
Further conditions related to this topic are provided in the Technical Specification of the Call.