Automatiserad och AI-driven testning för inbyggda system - Ett Robot Framework-baserat testsystem för verifiering av M-Bus- och Modbus-kommunikation i IoT-gateways

Abstract

This thesis investigates how a manual testing process for embedded gateway systems can be automated, and how artificial intelligence can be used as support in such a process. The work was carried out in collaboration with PiiGAB and focused on the company’s IoT gateways, which are used for collecting and forwarding measurement data through protocols such as M-Bus and Modbus. The project resulted in a modular automated test environment based on Python and Robot Framework. The system can execute test sequences, handle external con figuration files, collect and compare measurement data, and generate summarized test reports. Several tests were implemented to verify gateway communication and functionality. The modular structure makes it possible to extend the test environment with additional protocols, meters, and test scenarios in the future. However, the work also shows that reliable automated testing depends on controlled test data, correct con figuration, and verification of the test system itself. AI was evaluated as a possible support tool during the project. It proved useful for information gathering, documentation, idea generation, and general reasoning about test cases. However, it was less reliable when detailed protocol-specific functionality had to be implemented, especially regarding M-Bus structure. Therefore, AI was not integrated directly into the final test application. The conclusion is that AI can support the testing process, but it cannot replace human technical review and verification.