Model-Based Testing for Embedded Systems

260 Model-Based Testing for Embedded Systems

For symmetry, a test (a, B) is called negatively significant for a requirement if every

system behavior F that does not pass the test universally fulfills the requirement R. This

is formally expressed by

¬ptu(F,(a,B)) ⇒R(F).

Of course, in practice, a significant test is only achievable for very simple requirements.

Among other things, testing can be driven by fault models rather than by requirements.

Fault-based tests can be designed and run whenever there is knowledge of a class of systems.

Typical examples include limit-value testing or stuck-at-1 tests. The idea is to identify those

situations that are t