Weather has been closely associated with flight safety since the beginning of aviation history, with many aviation accidents and incidents occurring due to unusual weather conditions near airports. The current aviation unusual-weather detection systems use the weather information provided by ground meteorological observation stations in the airport vicinity. They can only detect unusual weather conditions near the ground surface because all the ground meteorological observation stations are at almost the same elevation. Consequently, this Paper aims to augment the current system by collecting vertical weather information in the air from a new data source, namely, Automatic Dependent Surveillance-Broadcast systems. This Paper evaluates the correlation between actual aircraft Automatic Dependent Surveillance-Broadcast data and unusual weather conditions based on several unusual-weather detection algorithms and then uses machine learning to establish the system prototype to detect aviation unusual weather. These unusual-weather detection algorithms have been verified in this Paper to distinguish between normal and unusual weather conditions. The developed machine-learning model can provide accuracy rates of close to 98% for detecting aviation unusual weather conditions. Finally, we develop a system prototype graphical user interface with which users can interact through graphical icons and indicators. Users can directly get information regarding aircraft and weather from this system prototype. The testing of the developed models is validated by the actual Automatic Dependent Surveillance-Broadcast signals broadcast by several aircraft recorded at the airport.
All Science Journal Classification (ASJC) codes
- Aerospace Engineering