Data Mining of Air Traffic Control Operational Errors

In this paper we present the results of applying data mining techniques to identify patterns and anomalies in air traffic control operational errors (OEs). Reducing the OE rate is of high importance and remains a challenge in the aviation safety community. Existing studies, which use traditional methods and focus on individual aspects of OEs, are limited to operations at a single facility, or events in a short period of time. A holistic study of historical data available on OEs has not been conducted. We have applied an attribute focusing technique to study 15 years of operational errors at all FAA Air Route Traffic Control Centers (ARTCCs) 1 in the National Airspace System (NAS) in the U.S. We have found ‘interesting’ patterns of common characteristics, anomalies, and changes in trends of operational errors. We interpreted the results with the help of domain experts and plan to do a similar analysis for OEs at other types of air traffic control facilities (towers and TRACONs) as well.