Test Suite Augmentation for Reconfigurable PLC Software in the Internet of Production

Regression testing is an established technique used to attest the correctness of reconfigurations to PLC software. After such a reconfiguration, a test suite might not be adequate to ensure the absence of regressions, requiring the derivation of new test cases to uncover potential regressions. This paper presents a combination of state-of-the-art symbolic execution algorithms for test suite augmentation, an indispensable part of regression testing. Test generation is guided towards the changed behavior using a technique known as four-way forking. The old and new PLC software are executed in the same symbolic execution instance to account for the effects of the reconfiguration and increase the chances of generating difference-revealing test cases. The prototypical implementation is evaluated using domain-specific benchmarks such as the PLCopen Safety library and the Pick and Place Unit, exposing the limitations in applicability and effectiveness of the used techniques for safeguarding PLC software subject to frequent reconfigurations as found in cyber-physical production systems.