Air Traffic Configuration Modelling and Dynamic Airspace Allocation using Discrete-Time Markov Chains

This paper introduces the use of discrete-time Markov chain models to represent the spatial-temporal behaviour of air traffic configurations near terminal areas. The airspace is first discretised using a rectangular grid and the occupancy status of each grid cell over each time interval is recorded. Each unique grid occupancy combination is then vectorised to define a unique state. Using the Baum-Welch algorithm, the transition probability between the set of possible states is then determined. Using real air traffic data, the approach is used to model the air traffic behaviour around a large Australian airport. The model is evaluated using hold out data and applied to the novel application of dynamic airspace allocation for integrated unmanned aircraft operations.