Aircraft-Based Complexity Assessment for Radar Controllers in the Multi-Sector Planner Experiment

Abstract

New technologies and automation tools under development in Next Generation Air transportation System (NextGen) will change controllers tasks, roles, and responsibilities. However, controllers cognitive complexity will remain one of the limiting factors on system capacity. To better understand cognitive complexity in future Air Tra c Control (ATC) environments, an investigation of cognitive complexity factors was performed for radar controllers in the Multi-Sector Planer (MSP) II experiment. A large amount of work in ATC complexity has been performed to identify factors and in uences that make an air tra c situation more or less complex. Summaries of these studies can be found in the review papers. Most of the complexity factors identi ed can be grouped into two categories: the distribution of aircraft in the air tra c situation and properties of the underlying structure in a sector. Indicators for the distribution of aircraft in the air tra c situation can be measured directly with parameters, such as tra c density, the proportion of aircraft changing altitudes, and number of con icts, etc. However, indicators for the properties of the underlying structure in a sector cannot be readily calculated, for example, sector shape, the con guration of airways, and the impact of restricted areas of airspace, etc. A number of quanti able metrics based on the identi ed complexity factors have been proposed to describe ATC complexity or the limit of controller workload. For instance, Dynamic Density is intended as an objective measure to identify situations that are complex enough such that centralized control would still be required in the concept of Free Flight. Kopardekar et al. de nes Dynamic Density as the collective e ect of all factors, or variables, that contribute to sector level air tra c control complexity or di culty at any given time. Multiple metrics related to Dynamic Density have been proposed using various sets of variables representing complexity factors. These metrics formulates the relationship between complexity factors and controller indicated complexity level. Four popular Dynamic Density metrics are examined by Kopardekar and Magyarits . Twelve complexity factors with high weightings from the four metrics have been identi ed and incorporated into one single metric. Further study indicates that the Dynamic Density metric performs better than aircraft count. However, using Dynamic Density also has its shortcomings. Factors weightings are applicable only to the sector in which the data are collected and validated. Some other complexity models also attempt to capture intrinsic complexity factors. For example, Delahaye and Puechmorel use factors derived directly from the locations and speeds of aircraft. In this study, cognitive complexity was measured through controller’s evaluation of the complexity contribution of each individual aircraft. A few studies have proposed complexity metrics based on aircraft count. However, no complexity measure used in the past has the ability to explicitly assess each aircraft’s contribution to controller’s cognitive complexity to support the development of those aircraft-based complexity metrics. An aircraft-based complexity assessment method was developed to obtain aircraft speci c information. In this method, experiment participants were asked to identify speci c aircraft in a tra c