Detail

SESAME – A Model-driven Process for the Test Selection of Small-size Safety-related Embedded Software

Embedded software applications are part of our daily lives (home appliances, means of transport, etc.), some of which have safety implications on human beings. It is thus important to trust this type of software. In our industrial context, testing is the most widely-spread technique used for that purpose. The state of practice is divided in two main categories. Explicit test selection techniques that enumerate the test cases to be exercised, but do not offer large coverage of the system’s behavior; and implicit techniques that encompass fully automated test generation techniques, which tend to hide the test selection information from engineers. The problem that we aim at solving in this thesis is the definition of a test selection approach that is capable of helping test engineers to better reach a delimited and verifiable test set with respect to some given test requirements and project specificities taking into account different test stakeholders.

Model-Driven Engineering (MDE) is a part of software engineering in which models are not only used as documentation but also as a building tool. Models become the backbone of the software development process. In this thesis, we define a lightweight model-driven test selection process (called SESAME) based on formal grounds that is particularly adapted to small-size and safety-related embedded software systems. The SESAME process makes a systematic and coherent use of meta models and model transformations, the two fundamental concepts of MDE. Meta models are used to precisely define the different types of SESAME models and model transformations are used to give meanings to these models.

The SESAME process is composed of the four following tasks: specification of the system for testing purposes; specification of a test selection that constrains the system specification; evaluation of the test selection; generation of test cases derived from the constrained system specification.

As a validation of our approach, we apply the SESAME process to a case study taken from an industrial project. This industrial application has shown the applicability of our approach on a small-size safety-related embedded system.