Generating Test Cases for Timed Systems from Controlled Natural Language Specifications

Abstract

Dynamic testing is still the most used quality assurance technique in the automotive industry. There is a need to automate the testing process as much as possible. In this work we focus on the automatic generation of test cases from requirement specifications. To embed the approach as close as possible into existing workflows we start with natural language like specifications, as requirements are still mostly written in natural language. To support this, we specify a controlled natural language for the automotive domain. After acquiring the requirements they are translated into a formal model. The model enables an efficient reachability analysis and allows to describe rich temporal behavior. We then use partial order planning to create positive and negative tests. The resulting test cases are able to handle non-deterministic timing behavior. Furthermore the test cases can be presented in a comprehensible way, so that the reader can validate them.