Functional test data generation for Simulink-like models

Abstract

Abstract Embedded systems are increasingly present in many electronic devices and is often related to critical applications. Therefore, the need for a well planned and executed testing procedure is even higher. We intend to contribute in this area by presenting an experimental evaluation of the pairwise combinatorial approach as a technique for test data generation applied specifically to Simulink-like models. In particular, we have applied our strategy to the generated source code of several models. Furthermore, a testing tool was developed to assist in the test data generation process. We show that there is no statistical significant advantages of the proposed approach over random generation of test data, but when used together they yield better results. The feasibility of the experimental results indicate that efforts can be employed in order to obtain a testing strategy integrated within a testing environment.