Ground control uses the Departure Sequence Planner (DSP) to optimally plan departures within the operational constraints. The DSP uses an estimated Variable Taxi Time (VTT) to calculate an aircraft’s Target Take-Off Time (TTOT). If the VTT is underestimated flights will not make the determined TTOT, while an overestimation requires the air traffic controller to tactically hold an aircraft. These consequences result in delays, capacity losses, additional workload, and uncertainty. This uncertainty inhibits further operational optimizations. The VTT is currently predicted based on the average taxi times for different gate-runway combinations, wake turbulence categories, deicing procedure, and simplified runway configuration. The goal of this project is to improve VTT predictions by employing machine learning techniques and additional data sources such as traffic density, weather, aircraft type, and more.
Graduated: June 2021