ABSTRACT
In the domain of embedded systems, models are often used either to generate code, possibly after refinement steps, but they also provide a functional view of the modeled system that can be used to produce black-box test cases, without considering the actual details of implementation of this system. In this process, the tests are generated by appling given test selection criteria on the model. These test cases are then played on the system and the results obtained are compared with the results predicted by the model, in order to ensure the conformance between the concrete system and its abstract representation. Test selection criteria aim at achieving a reasonable coverage of the functionalities or requirements of the system, without involving a heavyweight human intervention.
