Prediction of air traffic complexity through a dynamic complexity indicator and machine learning models

Abstract

In recent years, there has been an increase in traffic demand. This means that the balance between the capacity of the Air Traffic Control system and traffic demand is affected. As demand exceeds capacity, measures such as the Air Traffic Flow and Capacity Management regulations have emerged to reduce the number of flights in the airspace. Complexity is a topic widely studied by researchers all over the world. For this reason, the objective of this paper is to develop a complexity indicator that can be used to predict complexity of Air Traffic Control sectors with help of Machine Learning models. The structure of complexity prediction is based on different machine learning models predicting operational variables using Random Forest Algorithms, and then predicting the complexity combining the results of the Machine Learning models. With this artificial intelligence application, the objective is to predict a complex variable by structuring the problem and dividing it in simpler models. Thanks to the application of the methodology, the Air Traffic Control service can see which possible flows or sectors will be congested and thus allocate resources optimally, but also simulations of different scenarios can be made to analyse how the operation changes, and thus structure the traffic prior to the operation.