|
Abstract:
|
Aviation safety professionals are often interested in the control actions a flight crew must make to generate an observed flight trajectory, because aviation safety regulations generally apply to airlines, aircraft, and pilots which indirectly act on flight controls. Systems designed to enable air traffic control primarily measure position and velocity over the ground, rather than measurements of the aircraft state that provide insight to safety regulations. Due to the wide availability of surveillance data used for air traffic control, there are significant benefits to algorithms that can estimate flight dynamics and control inputs based on surveillance data. Some inverse simulation approaches have been proposed by the aeronautical engineering community, but this paper examines the application of several statistical learning techniques using full-flight simulator recordings to estimate flight dynamics parameters from Automatic Dependent Surveillance - Broadcast data.
|
