The automotive industry is in the midst of a digital transformation. Shift-left practices and the rise of software-defined vehicles (SDVs) are accelerating this change. To keep pace, developers need advanced tools, efficient workflows, and robust frameworks that enable them to build, test, and update software reliably. Continuous integration and continuous testing (CI/CT) pipelines have become essential for achieving scalable, efficient, and reliable software validation. When combined with software-in-the-loop (SIL), they enable accelerated, automated testing of virtual electronic control units (V-ECUs), ensuring consistency and rapid feedback across development cycles.
What is SIL?
Software-in-the-loop (SIL) is a simulation-based testing method in which embedded software runs on a virtual platform instead of physical hardware. This approach enables developers to validate software logic, algorithms, and system behavior early in the development cycle. SIL therefore plays a critical role in verifying electronic control unit (ECU) software long before physical hardware is available. This architectural shift also opens new opportunities: Developers can apply modern agile and DevOps methods to update software much faster, without being limited by hardware dependencies or other physical updates.
Why use SIL in CI/CT pipelines?
Modern software enables vehicles to achieve a higher level of safety, comfort, and functionality. For example, advanced driver assistance systems, automated driving functions, and intelligent connectivity provide greater protection, a better user experience, and more convenient operation in everyday driving. In order to reliably develop and test these complex and powerful functions, automotive teams need modern, flexible software development environments; conventional legacy systems quickly reach their limits in this regard. One promising solution approach is the introduction of a cloud-based, centralized platform that bundles global resources and enables any tests to be performed on virtual test benches at any location. This allows users to overcome development and testing bottlenecks and achieve the scalability that is essential for overcoming current challenges.
dSPACE’s industry-leading tool chain for SIL integration testing lets you generate and import your code for the system under test and connect it to other simulation parts, such as plant and environment models or other V-ECUs.
SIL testing allows for the validation of ECU software in a virtual environment without hardware dependencies. CI/CT pipelines offer the benefits of automation, scalability, and early feedback to embedded systems development. When integrated into CI/CT pipelines, SIL enables the following:
- Early detection of issues before hardware becomes available
- Automated regression testing
- Reduced manual effort
- Faster development cycles through continuous validation
- Improved traceability and repeatability
- Reusability of the SIL artifacts in HIL testing with dSPACE solutions
What is necessary to successfully integrate SIL into CI/CT pipelines?
A SIL pipeline can consist of five core components:
- Model or source code
– Based on the V-ECU levels defined by ProSTEP, SIL typically starts with a Simulink® model for Level 0 V-ECUs. For higher levels, it uses production-like artifacts such as C/C++ code or other embedded software components that represent the system under test. - Virtual execution environment
– A simulation platform to execute the software in a controlled environment. - Test automation framework
– Scripts or tools to automate test case execution, data logging, and result analysis. - CI/CT pipeline orchestration
– Streamlined workflows can be easily integrated with version control systems that trigger builds and test executions based on code changes or schedules. This ensures traceability and consistency, thereby accelerating the development cycle. - Reporting and feedback
– Automated result generation with KPIs enables the developers to compare the results and increase productivity.
What does a typical CI/CT pipeline workflow for V-ECUs look like?
A typical automotive CI/CT pipeline workflow for virtual ECU development follows a structured process:
- Code commit & trigger:
The developer writes and validates code locally, then pushes changes to the version control system (e.g., Git). This ensures that all updates are tracked and integrated into the shared repository. The commit triggers the CI pipeline, initiating the automated processes. - Build & simulation setup:
The system builds the virtual ECU (V-ECU) or software components. These are then deployed into the simulation environment, which may include virtualized hardware models or plant/environment models for testing. - Test execution and result analysis: Predefined test cases, such as functional tests, are executed automatically. During this process, logs, KPIs, and performance metrics are collected and analyzed – by generating dashboards, comparing results to baselines, and identifying anomalies or failures. This ensures that new code meets quality standards and does not impact existing functionality.
- Feedback loop: Developers receive detailed reports or alerts via e-mail, dashboards, or integrated tools. These reports highlight pass/fail status, coverage metrics, and any issues that need attention, enabling quick fixes and continuous improvement.
The dSPACE Workflow for SIL Orchestration
The dSPACE SIL automation framework manages the complete V-ECU lifecycle, starting from generation and ending with validation. It leverages a suite of dSPACE tools to ensure a streamlined and automated process.
Key Stages
V-ECU generation with SystemDesk & TargetLink:
These tools enable the creation of accurate and production-ready V-ECUs.
- SystemDesk: Used for architecture modeling and generating V-ECUs from AUTOSAR-compliant models. More: SystemDesk - dSPACE
- Upcoming feature: Copilot agent for V-ECU generation: GitHub Copilot extension agent that developers can directly use in the development environment to generate V-ECUs.
- TargetLink: Generates highly optimized, production-ready C code from Simulink® models for embedded systems. More: TargetLink - dSPACE
- dSPACE also allows a model-as-code workflow for model-based development. More: Video: Model as code – Efficient workflows - dSPACE
Build using VEOS:
- A SIL simulation platform/SIL simulator that runs the generated V-ECUs in a virtual environment. More: VEOS - dSPACE
- Supports integration with plant models and other simulation components for system-level testing.
- VEOS also supports cosimulation with platforms such as QEMU as well as ARM-based high-performance computing (HPC) systems.
CI/CT pipeline integration:
CI pipelines trigger builds and tests on code commits or scheduled intervals. Teams use YAML files to orchestrate these automated processes in a structured and consistent way. The simple, human-readable configurations define the automation logic, ensuring reproducibility and scalability.
Test execution:
Local automated execution with AutomationDesk:
- Executes automated test sequences and logs results.
- Integrates with test libraries and supports regression testing for continuous validation.
- More: AutomationDesk - dSPACE
Scaled executions on cloud with SIMPHERA:
- Enables users to design, execute, and analyze highly realistic, deterministic simulations.
- Reduces the time required for testing by executing simulation scenarios in parallel in the cloud.
- More: SIMPHERA - dSPACE
Result analysis:
Logs and signal traces are collected and post-processed to evaluate pass/fail criteria.
Integration backbone:
SIL workflows demand seamless connectivity between modelling tools, simulation platforms, and automation frameworks.
- At the core of this dSPACE SIL-CI/CT pipeline integration lies a scripting layer powered by Python and dSPACE APIs, which orchestrates interactions across diverse tools.
- Python and dSPACE APIs allow seamless integration with CI/CT systems like GitLab CI, making it possible to trigger builds, simulations, and tests automatically after code commits.
Conclusion
A well-implemented CI/CT pipeline leveraging SIL capabilities empowers automotive companies to improve software quality and accelerate development. The solution enables scalable testing for various projects in the cloud. The test efficiency is improved by less manual intervention and faster feedback loops. Open interfaces and a robust architecture make maintenance, updates, and integration with external systems easier. Collectively, these attributes position the approach as a strategic enabler for future-ready automotive software development.
