How explicit are the barriers to failure in safety arguments?

Abstract

Safety cases embody arguments that demonstrate how safety properties of a system are upheld. Such cases implicitly document the barriers that must exist between hazards and vulnerable components of a system. For safety certification, it is the analysis of these barriers that provide confidence in the safety of the system. The explicit representation of hazard barriers can provide additional insight for the design and evaluation of system safety. They can be identified in a hazard analysis to allow analysts to reflect on particular design choices. Barrier existence in a live system can be mapped to abstract barrier representations to provide both verification of barrier existence and a basis for quantitative measures between the predicted barrier behaviour and performance of the actual barrier. This paper explores the first stage of this process, the binding between explicit mitigation arguments in hazard analysis and the barrier concept. Examples from the domains of computer-assisted detection in mammography and free route airspace feasibility are examined and the implications for system certification are considered. © Springer-Verlag 2004.