Bayesian network model of aviation safety: Impact of new communication technologies on mid-air collisions

This article presents a method of estimating the risk of a mid-air collision. The proposed method is an enhancement of the traditional aviation safety model - Integrated Safety Assessment Model (ISAM) - developed by the Federal Aviation Administration (FAA) and EUROCONTROL. ISAM is a mix of event-based models and fault trees that identifies causes of 35 different types of aviation accidents. While useful for conceptual understanding of accidents, the model does not handle human-technical or inter-system interactions. These drawbacks are especially evident when assessing safety impact of new communication, navigation and surveillance technologies since they rely on pilots and controllers. We propose a method of analyzing the impact of new technologies in aviation by presenting a case study of the Data Communication system - a new technology developed by the FAA used for communication between pilots and controllers. The method builds upon ISAM and leverages a Bayesian Network to estimate safety risk. The results indicate that the implementation of Data Comm can reduce the risk of collision by 25 %. In addition, if a collision has occurred, it is 10 million times more probable that the likely culprit is an error in human communication rather than a failure of communication equipment.