Developing a predictive model of controller workload in Air Traffic Management

Workload has long been used as a metric to indicate system performance and operator strain (Moray 1979). The National Air Traffic Services (NATS), Air Traffic Management Development Centre (ATMDC) uses a unique methodology and toolset based on Wickens' Model of Multiple Resources (Wickens 1991) to analyse and predict workload for a given set of circumstances. PUMA (Hook 1993) (Performance and Usability Modelling in ATM (Air Traffic Management)) uses a task analytic approach to analyse workload and infers workload on the basis of observational task analysis, cognitive debriefs (auto-confrontation) and interviews with subject matter experts. A large and detailed model of a concept of air traffic operation is developed and is subsequently used to predict workload. The following paper describes the PUMA workload analysis process and reviews several projects in which PUMA has had an impact in terms of Human Machine Interface (HMI) design, and workload assessment.