Classifying Aircraft Approach Type in the National General Aviation Flight Information Database

This work details the development of the “Go-Around Detection Tool”, a tool for classification of aircraft approach types for the National General Aviation Flight Information Database (NGAFID). The NGAFID currently houses over 700,000 h of per-second time series flight data recorder readings generated by over 400,000 flights from 8 fleet of aircraft and over 190 participating private individuals. The approach phase of flight is one of the most dangerous, and classifying types of approaches as stable or unstable, and if they were a go-around, touch-and-go, or stop-and-go is an especially important issue for flight safety monitoring programs. As General Aviation typically lacks the Weight on Wheels (WoW) technology and many others that exist within Commercial Aviation, there is difficulty in detecting landings and go-arounds as these need to be inferred somehow from the raw flight data. The developed application uses several airplane parameters reported by a flight data recorder and successfully detects go-arounds, touch-and-go landings, and stop-and-go landings as either stable or unstable with an accuracy of 98.14%. The application was tested using 100 student flights from the NGAFID, which generated 377 total approaches. Out of those approaches, 25.73% were classified as unstable. It was found that only 20.62% of all unstable approaches resulted with a go-around, which is far from the ideal 100% goal. Lastly, the application was parallelized and found to have a 9.75x speedup in doing so. The Go-Around Detection Tool can be used to provide post-flight statistics and user-friendly graphs on both an organizational- and individual-level for educational purposes. It is capable of assisting both new and experienced pilots for the safety of themselves, their organization, and General Aviation as a whole.