Relation Between Test Coverage and Timed Automata Model Structure

Abstract

This paper deals with problematics of structure of Timed Automata models suitable for Model-Based Testing of automotive systems. Previous experiments, primarily focused on the environmental models, have shown that their structure does not significantly affect the coverage speed of testing process. However, similar questions regarding the observer part of the system model remained open. This paper analyzes those remaining questions and focuses on uncovering possible relation between an observer model structure and the quality of generated test sequences according to multiple criteria. Goal of presented experiments is to compare multiple modeling approaches and discover which one is most suitable for automotive systems.