Improving dependability through a deviation analysis on distributed tasks in safety critical systems

This paper describes a method aiming to support dependability in interactive-safety critical systems. It proposes a multidisciplinary approach integrating areas usually considered separately -task modelling and distributed cognition- in order to analyse if possible deviations occurring in the performance of a system might affect its overall dependability. The basic idea is that analysis of interactive systems requires a clear understanding of the information needed to accomplish the tasks, and this information has to be derived from both internal cognitive representations and external representations provided by various types of artefacts. We illustrate our approach also applying it to a case study in the air traffic control domain.