The project started by identifying the main data sources on ATM performance and assessing the validity, quality, and geographical and temporal resolution of each dataset. This work produced three main outcomes: a set of Performance Data Factsheets characterising each data source, a Performance Data Guide which links ATM performance data with the sources where such data can be found, and the INTUIT Data Repository, which allowed the project partners to share the datasets used for the INTUIT research work. These outcomes are documented in deliverable D2.1.
Taking this work as a starting point, a combination of literature review and stakeholder consultation allowed the identification of a list of relevant research questions at the intersection of ATM performance modelling and data science, documented in D2.2. Based on a combination of factors, including the relevance of the research question, the expected impact of the results, the availability of sufficient data and the potential of data science to advance the state-of-the-art in that particular field, a subset of these research questions was selected to be investigated in the form of three Case Studies (CS):
- CS-1: Effect of unit rates on airline route choices and impact on ATM performance. The goal was to develop new models able to predict airline route choices between different ODs in order to evaluate the performance trade-offs arising from these decisions (e.g. cost efficiency vs environment). The proposed approach has shown significant potential to improve the understanding of route choices, and it can be applied to pre-tactical traffic forecast.
- CS-2: Sources of en-route flight inefficiency. This case study, conducted in collaboration with the SESAR ER projects AURORA and APACHE, investigated the causes of inefficient routes in the European Network and their effects on performance by means of a machine learning algorithm that predicts en-route flight efficiency, in order to isolate the contribution of different factors and stakeholders.
- CS-3: Multi-scale representation of ATM performance indicators. This case study aimed to disaggregate traffic data and performance indicators at ACC and sector level and to model the relationship between these variables at different scales (e.g. what is the influence of sector configuration on the aggregated performance of an ANSP?).
In order to explore the datasets used for the case studies, different visualisation and visual analytics tools were developed. This work is documented in D3.1 Visual Analytics Exploration of Performance Data. The performance modelling work is documented in D4.1 Performance Metrics and Predictive Models. Finally, the new visualisations and modelling techniques were integrated into an interactive performance monitoring and management dashboard, including: (i) a multi-objective optimisation engine to find the Pareto optimal solution for a set of KPIs. The tool allows the estimation of the effects of a particular setting of unit rates and helps in the selection of the best setting of unit rates to optimise the trade-off between flight efficiency, cost efficiency and capacity, (ii) a flight efficiency monitoring dashboard to identify and evaluate the causes of flight inefficiency in a particular ACC. The results of this work are documented in D5.1 Performance Monitoring and Management Toolset and D5.2 Performance Monitoring and Management Toolset Evaluation Report.
