ABSTRACT
Multi-agent dynamic risk modelling is used to assess accident probabilities of air traffic scenarios involving multiple entities, disturbances and their interactions. In this context human performance models should be able to capture a wide variety of hazards. This paper provides an analysis of the coverage of a broad set of hazards that can be attained by model constructs in multi-agent dynamic risk modelling. It shows that multi-agent situation awareness is a prime construct for understanding and analysing hazards in air traffic scenarios.
