Information Technology Reference
In-Depth Information
reliable automatic test case generation tools grounded on solid foundations. The
European FP7 project MOGENTES serves these demands.
MOGENTES stands for Model-based Generation of Tests for Dependable Em-
bedded Systems and its goal is to significantly enhance testing and verification
of dependable embedded systems by means of automated generation of ecient
test cases relying on development of new approaches as well as innovative inte-
gration of state-of-the-art techniques. In particular, MOGENTES aims at the
application of these technologies in large industrial systems, simultaneously en-
abling application domain experts (with rather little knowledge and experience
in usage of formal methods) to use them with minimal learning effort.
The industrial partners in the project identified UML as their future modeling
paradigm and hence, require test case generation tools to process UML models.
This need conflicts with the requirement that our test case generation technique
has to be build on solid foundations, because UML lacks a standard formal
semantics. However, a formal semantics is essential for our testing techniques
based on precise fault-models and formal notions of conformance. Therefore, we
decided to treat UML as a front-end modeling language and translate it to a
formal back-end formalism on which our test case generators will work on.
In this paper, we give insights into this translation process. A car alarm system
serves as a running example. Section 2 presents the UML model of the car alarm
system including the technique to express the testing interface in UML class dia-
grams. Then, Section 3 presents and motivates our back-end formalism, namely
Object-Oriented Action Systems (OOAS), a formalism well-suited for expressing
object-oriented models of embedded systems. This section also presents a further
level of semantic mapping: the behavior of state-rich OOAS is interpreted as a
series of controllable and observable events. It is this event-level on which our test
case generators work. This gives us the advantage that we can base our formal
testing approach on the existing testing theory on labeled-transition systems.
Next, in Section 4 we discuss our semantic mapping, including the translation
of non-trivial UML state charts with nested states, parallel regions and time
triggers. In Section 5 we discuss the case study. Finally, in Section 6 we draw
our conclusion and give an outlook on future and related work.
2AUML-Modl
We use a very simplified car alarm system as an example for discussing the
concepts and issues of the transformation of UML models to action systems.
The example is taken from Ford's automotive demonstrator within MOGENTES
and the main purpose of this rather simple example within the project is to test
and validate the test-case generation work flow on a basic level. Notice that we
are dealing with black-box testing here, as, e.g., Ford wants to test components
provided by an external partner based on the requirements that were given to
this company.
Before we can generate any test-cases, we need to build a model from the
requirements. For our simplified car alarm system (CAS), we were given the
following three textual requirements.
