Spatiotemporal Data-Driven Simulation and Clustering of Ground Operations of Aircraft for Comprehending Airport Jams and Collisions

摘要

Increasing air traffic volume poses challenges of safe airport ground operations, especially in jammed areas where higher risks of collisions arise. Unfortunately, operators lack detailed spatiotemporal data for predicting jams and related collision risks based on airport layout and flight schedules. Without those correlations, operators could only qualitatively assess the airport operational conditions based on their experiences. Detailed historical data set, such as ASDE-X, provide the potential of quantifying the correlations between aircraft motions, jams, airport layout, and other environmental conditions. Such detailed data could form a reliable basis for spatiotemporal simulation and prediction of collision risks. This paper focuses on establishing a quantitative spatiotemporal data-driven simulation framework capable of predicting airport jams and ground collisions, with a focus on clustering jams for predicting locations of high collision risks. The results revealed three clusters of jams across LAX airport during a day when a collision occurred by using simulations that models four typical collision scenarios synthesized from historical data. The overall conclusion is that the proposed framework could help to reveal how clusters of airport ground traffic jams occur and influence the safety and efficiency of airport operations.