Automated risk-based testing by integrating safety analysis information into system behavior models
The development of safety-critical software-intensive systems requires systematic quality assurance on all stages of the development process. Executable development artifacts are validated against the system specifications. Riskbased test approaches enable the distribution of test effort in a specific way to cover critical system parts, functions, and requirements. The development process of safety-critical systems usually implies analysis activities for determining and understanding hazards and risks. Moreover, it requires a systematic design of the system structure and behavior based on the specification. For achieving a high degree of automation of test case derivation, existing formal models from the risk analysis and system design phases are combined. The approach presented here focuses on integration of fault trees into statebased behavior models. Therefore, fault trees are analyzed and their elements are assessed for their validity and significance for the test modeling. The approach systematically transforms the relevant fault tree elements like single critical basic events, system states, or sequences of events into elements of the state-based behavior model. The resulting model enables the automated generation of test cases considering risk-based test purposes such as the coverage of critical states, transitions, or sub-models. The feasibility of the approach is shown in a small case study.