Comparative study of metroplex airspace and procedures using machine learning to discover flight track anomalies

The National Airspace (NAS) is constantly changing and adapting to new and complex challenges, and as a result Next Generation Air Transportation System (NextGen) will need to address these important aspects. These challenges range from increased traffic flow, to reducing environmental impact, to routing efficiency, all while maintaining high safety. The FAA has recently been involved in a number of large scale Metroplex redesigns across the country to enable controllers and pilots to implement more efficient Performance-based Navigation (PBN) procedures on a regional basis. The Houston Metroplex project is one of the first to implement such a large scale change in the NAS, and it significantly changed the traffic flows into the Houston Terminal Radar Approach Control Facility (TRACON) airspace for the two major airports: Houston´s George Bush Intercontinental Airport (IAH) and Houston´s William P. Hobby International Airport (HOU). This paper addresses an anomaly detection approach that has previously been used to detect operationally significant anomalous flights on approach to Denver International, Newark International, LaGuardia International, and John F. Kennedy International airports. The same method is applied to radar track surveillance data to identify anomalies in the airspace before and after the Metroplex procedure change at Houston. The study covers flights traversing through Houston TRACON (I90) and landing at IAH and HOU over a period of 2 years before and after the significant procedure change. Anomalies identified before and after the procedure change were characterized by their safety risk and operational efficiency to determine whether the types of anomalies that were discovered from before continued to exist or if they were eliminated after the procedure change or if new types of anomalies began to appear.