Test-based behaviour model derivation and adaption to enable automated test case scheduling for automated production systems

Abstract

AbstractAutomated production systems require continuous maintenance and validation through regression tests after on-site changes. However, identifying change-affected system parts, selecting relevant test steps, and scheduling them time-efficiently are time-consuming for on-site test engineers, especially in legacy systems lacking updated documentation. This paper proposes an approach to record a system behaviour model during system testing based on the system’s sensor, actuator and internal variable values. The resulting behaviour model is expandable, covers non-deterministic behaviour, includes timing information, more precisely, timestamps, and is usable for test planning. The resulting model is subsequently used for test step derivation based on the system states to be tested and automated test case scheduling.