Towards safer skies: A probabilistic approach to risk assessment of new aviation safety products

摘要

The Aviation System Risk Model (ASRM) is both an analytic method and a decision support tool to assist decision makers with risk assessments of new aviation safety products. The approach uses a Bayesian Belief Network (BBN) framework to provide a causal structure for aircraft accident modeling. In particular, the modeling approach explores the probabilistic interdependencies among individual, task/environmental, and organizational factors leading to an accident. The causal models are "quantified" through knowledge engineering sessions with multiple subject matter experts. Then a portfolio of new aviation safety products, such as synthetic vision, advanced weather and communication sensors, among others, are assessed for their projected impact upon relative safety risk compared to a baseline period. The ASRM is explained with one of the models developed for a certain aircraft accident type known as Controlled Flight Into Terrain (CFIT).