A novel network approach to study communication activities of air traffic controllers

Air traffic controllers play critical roles in the safety, efficiency, and capacity of air traffic management. However, there is inadequate knowledge of the dynamics of the controllers' activities, especially from a quantitative perspective. This paper presents a novel network approach to uncover hidden patterns of the controllers' behavior based on the voice communication data. We convert the time series of the controllers' communication activities, which contain flights' information, into a time-varying network and a static network, referred to as temporal network and time-aggregated network, respectively. These networks reflect the interaction between the controllers and the flights on a sector level, and allow us to leverage network techniques to yield new and insightful findings regarding regular patterns and unique characteristics of the controllers' communication activities. The proposed methodology is applied to three real-world datasets, and the resulting networks are closely examined and compared in terms of degree distribution, community structure, and motifs. This network approach introduces a “spatial” element to the conventional analysis of the controllers' communication events, by identifying connectivity and proximity among flights. It constitutes a major step towards the quantitative description of the controller-flight dynamics, which is not widely seen in the literature.