Aircraft Navigation Systems Safety Assessment via Probabilistic Model Checking

The safety assessment process is a mandatory step in the development and certification of safety-critical systems such as the ones in the aerospace industry. In this work we show how Probabilistic Model Checking, a Formal Verification method, can help to assess the safety of navigation systems for a civil commercial transport category aircraft. The process involves a top-down approach identifying functions and its respective failure modes. Each failure event is associated with a hazard level, with an inverse relationship between the maximum acceptable probability and the event severity. Fault Tree Analysis (FTA) is the most commonly used method to quantify each events probability, but probabilistic models are also accepted as means of compliance demonstration, as per ARP-4761. Results show that the use of Probabilistic Model Checking as a means to complement non-formal methods is valuable, where we were able to evaluate the probability of several failure modes described in FAAs AC 20-138D, making use of Continuous-Time Markov Chains (CTMCs) with up 4.3 million of reachable states and 60 million of transitions.
isbn: 978-303086972-4
Volume: 12952
Pages: 465-480