Principal Investigator R Hansman
The growing availability of massive aviation data creates an opportunity for developing analytical tools that can be useful for post-event efficiency assessment, monitoring and alerting and real time decision support in the air traffic management system. Data types in the US National Air Transportation System (NAS) include planned and actual demand (e.g., airline schedules, wheels off times), air traffic control facility state (e.g., weather information, airport arrival rates) and planned and actual aircraft operations (e.g., flight plans, surveillance tracks). Many of the existing data sets are large volume (e.g., over 200 GB of weather data and 1.2 GB of flight track/plan data per day), large velocity (streaming continually) and large variety (different formats, data types, etc.). These characteristics make rapidly-evolving “Big Data” techniques particularly attractive to handle and convert the system raw data into useful information for system improvement.
The goal of this research project is to develop and demonstrate an interpretable, scalable and generalizable “Big Data” analytic framework for identifying patterns of air transport system behavior at various spatial and temporal scales. As an example, current research efforts include the characterization of NAS-wide network delay propagation dynamics at the strategic level and the characterization of air traffic flows in the transition/terminal airspace at the tactical level using data mining algorithms. The outcomes of the framework can be used to assess operational performance, identify and mitigate inefficiencies, identify operational best practices and generate inputs for descriptive and predictive models.