Model-based testing of real-time embedded systems in the automotive domain
Design decisions that used to be made at the code level are increasingly made at a higher level of abstraction. This shift of focus from implementation to design requires the creation of a consistent, reusable and well-documented specification model. Nowadays, about 40% of embedded system designs are within 20% of functionality expectations. This is partially attributed to the lack of an appropriate approach for functional validation. To improve hybrid system design, this dissertation presents a test method at the model level. The so-called Model-in-the-Loop for Embedded System Test (MiLEST) approach primarily employs a systematic, structured, repeatable, and abstract test specification and concentrates on automation of the test process. A signal-feature - oriented paradigm allows an abstract description of a signal and addresses the problems of the missing reference signal flows as well as systematic test data selection. Numerous signal features are identified while predefined test patterns help build the test specification. Testing then starts in the requirements phase and goes down to the test execution level. MiLEST is implemented in MATLAB/Simulink/Stateflow. Three case studies are presented. They correspond to component, component-in-the-loop, and integration level tests. Moreover, the quality of the resulting test models and test cases are investigated in depth.
Zugl.: Berlin, Univ., Diss., 2008