SystemC integration in SDV simulation environments through FMI-FMU

The shift towards Software-Defined Vehicles (SDVs), where software controls the majority of vehicle functions, has introduced complex challenges in embedded system design. SDVs rely on the seamless integration of diverse subsystems for autonomous driving, connectivity, safety, and entertainment, demanding highly adaptive, scalable, and reliable solutions. Traditional hardware-software design methods, which develop components in isolation, no longer meet the requirements of SDVs that depend on early integration and verification to satisfy strict safety, performance, and real-time constraints. Virtual Prototyping (VP) has emerged as a key approach for enabling early modeling, simulation, and validation of SDV components. Through VP, designers can simulate both hardware and software early in development, identifying potential issues and reducing overall development time. A central tool in VP is SystemC, a C++-based system-level modeling language that supports various abstraction levels, particularly beneficial in automotive contexts where safety standards like ISO 26262 are critical. SystemC facilitates early system-level verification of SDV subsystems, including control units, sensors, and actuators. To realize the full potential of VP, co-simulation across different design domains is essential. This is where the Functional Mock-up Interface (FMI), a tool-independent standard for model exchange and co-simulation, proves invaluable. By enabling the integration of models from different simulation environments, FMI allows for a unified virtual prototype, making it possible to validate SDV dynamics, control systems, and electronics as an integrated system. This thesis proposes a structured and automated framework for integrating SystemC-based virtual prototypes within a Functional Mock-up Unit (FMU) package using the FMI standard. The approach aims to enhance Software-Defined Vehicle (SDV) design through early validation and efficient cross-tool co-simulation, ultimately reducing costs and ensuring adherence to industry standards. Key technical challenges in this integration include handling the sc_in and sc_out ports in SystemC, managing both sc_start and sc_main functions, and ensuring data compatibility between SystemC and the FMI standard. To address these complexities, the proposed framework fully incorporates a SystemC-based hardware component into an FMU package, formatted as a .fmu zip file compatible with both FMI 2.0 and 3.0 standards. A Python master script, leveraging the FMPY library, is used to unpack, analyze, and co-simulate the FMU, providing configurable options for simulation time, step size, and variable management. To ensure reproducibility and scalability, the entire process is encapsulated within a Docker container, which maintains a consistent environment for simulation and dependency management, preserving the integrity of the original SystemC hardware code. Furthermore, performance profiling is conducted to evaluate and optimize simulation execution, providing insights into computational efficiency and ensuring a robust solution for effective hardware-software co-simulation.