A Scalable Software-in-the-Loop Test Methodology for Early Integration Testing Using Virtual ECUs

Abstract

The increasing software complexity of modern vehicles requires scalable test methods that enable early system integration testing independent of physical hardware availability. While Hardware-in-the-Loop (HiL) testing is well established in industry, it is limited by high costs, long setup times, and late hardware availability. Software-in-the-Loop (SiL) testing provides faster feedback cycles but is commonly limited to function-level validation. This paper presents a scalable SiL test methodology that enables early integration testing through the simultaneous execution of multiple virtual electronic control units (V-ECUs). The approach is based on production-oriented V-ECUs integrated into a synchronized SiL environment and supports automated test execution and continuous integration. The methodology is evaluated in an industrial case study using a prototypical SiL test system with six V-ECUs on the dSPACE VEOS platform. In total, 200 existing HiL test cases were analyzed for transferability to SiL. The results show that approximately 30% of the test cases can be executed under current conditions, enabling early detection of integration and communication issues. Limitations arise from hardware-dependent behavior, timing deviations, and incomplete virtualization. The results provide practical guidance on shifting integration test scopes from HiL to SiL in early development phases.