Skip to Main Content
This paper aims to describe a new methodology specifically designed for testing measurement and diagnostic software. A black-box procedure allows the user to verify whether the functional requirements of a software module under test are fulfilled. The robust experimental design techniques and statistical theories implemented to generate the software input test sets are described in detail. The reliability of the testing methodology is estimated by applying it to diagnostic software in wide use throughout the automotive industry. The results of validation tests carried out on data from a car engine measurement system are reported and analyzed.