Towards an Aspect Driven Approach for the Analysis, Evaluation and Optimization of Safety Within the Automotive Industry 2010-01-0208

An approach will be presented how development projects for safety-related and software-intensive automotive systems can be controlled through the application of model-based risk assessment. Therefore specific control measures have to be developed, which represent the degree of fulfilment of several aspects of safety-related developments. The control measures are evaluated through the analysis of risk-reducing aspects, for which the process of identification and specification is described. Thus, a framework for the creation of a probabilistic and aspect-oriented risk-analysis model (AORA) for safety related projects within automotive industries is currently under development. With respect to the upcoming safety standard ISO 26262 the twofold approach focuses on both, the identification and specification of risk-reducing aspects within the development as well as the application of a probabilistic reasoning model. The emphasis of this paper lays on the first step the process of identification, derivation and specification of risk-reducing aspects based on an industry specific safety standard. Therefore, cross-cutting safety concerns and attributes are analyzed within a literature research in order to derive and formulate risk-reducing aspects that are covered by an industry specific safety standard. The formalization empowers the analyst to estimate the current degree of risk mitigation using an underlying Bayesian Belief Network (BBN), which is a formalized representation of a safety standard. An incremental application of this approach within strongly iterative developments would allow for early and effective control of safety processes and the mitigation of risks due to the comprehensible determination and prioritization of preventive risk mitigation measures. This approach can be applied in order to preventively optimize safety consecutively.