Air traffic complexity indicators & ATC sectors classification

It is a widely held view that complexity is a key factor that significantly affects the work of an air traffic controller, which, in turn affects capacity. A better understanding of what makes the controllers´ work complex will improve current and future air traffic management (ATM) capacity, analysis, airspace planning, and future air traffic control (ATC) development. This paper describes our approach to develop a macroscopic model that will give us an automatic and non-subjective method to classify sectors according to their complexity. The first step was to identify the complexity indicators. We combined ATC operational advice with statistical analysis to compile a list of relevant complexity indicators. Clearly, these indicators, their influence and interaction vary amongst sector types. Hence, our next step was to classify our sectors into a small number of homogenous groups, or clusters, to arrive at the sectors´ typology. We used two approaches to classify the sectors. The first was based on a K-means classification and the second was by descendant hierarchical clustering - divisive segmentation. Our study shows that our model gave us a meaningful typology and understanding of our sectors´ complexity and that we can improve future controller workload and sector capacity predictions at a macroscopic level.