Traditionally, fault- or event-tree analyses or FMEAs have been used to estimate the probability of a safety-critical device creating a dangerous condition. However, these analysis techniques are less effective for systems primarily reliant on software, and are perhaps least effective in Safety of the Intended Functionality (SOTIF) environments, where a failure or dangerous situation occurs even though all components behaved as designed. At BlackBerry QNX we are evaluating the appropriateness of Bayesian Belief Networks for predicting defects in embedded software. This paper describes our approach and reports on early results from our research.
展开▼