Model-Based Regression Testing of Autonomous Robots

Testing is a common technique to assess quality of systems. Regression testing comes into view, when changes are introduced to the system under test and re-running all tests is not practical. Numerous techniques have been introduced to select tests only relevant to a given set of changes. These are typically based on source code, however, model-based development projects use models as primary artifacts described in various domain-specific languages. Thus, regression test selection should be performed directly on these models. We present a method and a case study on how model-based regression testing can be achieved in the context of autonomous robots. The method uses information from several domain-specific languages for modeling the robot's context and configuration. Our approach is implemented in a prototype tool, and its scalability is evaluated on models from the case study.