Modeling aircrew information management for estimation of situational awareness using dynamic Bayesian networks

Abstract

Situational awareness (SA) and its related metrics are often used to rate the performance of flight crews, especially for flight safety purposes. In the context of a Ph.D. thesis on data mining, this paper presents the principles followed to model and measure SA using Bayesian networks, focusing on the work performed to validate different data discretization methods. A simulation environment has been implemented to conduct experiments that monitor pilot activities, with special attention to information management. The outcome of simulated flights are datasets that contain relevant data, including control actions of the pilot, information queries, navigation information and environment and flight parameters. Large amount of data are generated during the experiments, driven by the complex and powerful flight data relationships provided by System Wide Information Management standards. Thus, dynamic Bayesian networks are specially applicable for this research due to their suitability to learn probabilistic dependencies from data.