Microservice integration testing with hardware-in-the-loop in CI/CD pipelines

Abstract

Mechanical products are increasingly evolving into software-driven systems with expanded responsibilities, as exemplified by the rapid advancements in the vehicle industry and the rise of autonomous driving technologies. Cyber-physical systems (CPSs), such as vehicles that rely on both hardware and software, present significant testing challenges due to their distributed nature and the necessity for real sensor input data. Traditional testing methodologies are suboptimal for CPS, as they typically do not accommodate the integration of both simulation and hardware testing in a distributed environment. This thesis aims to evaluate the feasibility and effectiveness of a containerized, microservice-based testing framework. The framework is designed to support simulation, data replay and hardware test levels and to be integrated into a continuous integration and continuous deployment (CI/CD) pipeline. The proposed framework was implemented and tested within the context of the TME290 Autonomous robots course at Chalmers University of Technology. This involved the development of hardware-in-the-loop (HIL) rigs, as well as the creation of test execution and interpretation software. The framework’s capabilities were assessed through the execution of various test scenarios. The integration of the testing framework into the course demonstrated its suitability for simulation, data replay, and hardware test levels on a distributed system. Furthermore, it was successfully integrated into a CI/CD pipeline. The findings suggest that a microservice-based architecture can effectively be used for the integration testing of CPS within a continuous integration environment. This approach enhances the reliability and efficiency of testing processes for autonomous systems, offering a promising solution to the challenges associated with traditional CPS testing methods.