Modern vehicles contain numerous software and hardware components that must work together to ensure the safety of the system even in the event of a fault. Verifying the corresponding safety mechanisms is a complex task: manually created tests often only cover some of the possible fault scenarios, their creation and maintenance is time-consuming, and particularly rare, complex fault cases often remain untested.
Systematically testing worst-case scenarios
Under the project management of Tiziano Munaro, a scientist in the Model-based Systems Engineering competence field, TeFoSa has demonstrated how test processes can be made significantly more efficient: Automated fault injection tests are used to introduce specific faults into software and hardware components in order to comprehensively test the performance of the safety mechanisms. “Our goal was to be able to systematically identify even the most complex error scenarios and thus strengthen confidence in the safety of modern vehicles,” explains Munaro.
It became clear that error modes in a cyber-physical system (CPS) can have different effects depending on the parameterization. For this reason, in order to reliably ensure functional correctness, accuracy, and timing security, so-called “worst-case” scenarios must be identified in a high-dimensional parameter space. TeFoSa developed a methodology for this purpose that treats test case generation as an optimization problem.
Evaluating automated test procedures in a practical setting
The developed procedures were tested in a practical setting for two security mechanisms of TTTech Auto MotionWise Automotive Middleware, with TTTech Auto providing the industrial environment. The results suggest that automated test case generation not only reduces the effort required to identify potential vulnerabilities, but also increases confidence in their absence.
TeFoSa thus makes an important contribution to the further development of methods for continuous software development in the automotive industry and provides valuable impetus for future test strategies for complex vehicle software. At the same time, the project demonstrates the benefits of automated testing approaches for functional safety.
