The Role of Model-based Design in Development of a Software Defined Vehicle
What is a Software Defined Vehicle?
Today’s automotive consumers demand seamless connectivity, intuitive infotainment systems, and advanced safety features. Original Equipment Manufacturers (OEMs) are actively exploring unique business opportunities, particularly in remotely updating software Over the Air (OTA) to address software issues, introduce new features on the fly, and enhance vehicle performance post-sale. The ongoing technology megatrends of automated driving, electrification, and connectivity all necessitate a substantial investment in software. This shift is transforming an industry that has traditionally been mechanically oriented, significantly elevating the importance of software and giving rise to the Software Defined Vehicle.
A software-defined vehicle (SDV) can be defined as a vehicle where the main value is derived from software.
How is development for a SDV different from a traditional vehicle?
SDVs are significantly different from traditional vehicles in the following ways:
- Changing vehicle E/E architectures: With increased software, there is a need for more computing power. However, adding computing power following the traditional method using single-function Electronic Control Units (ECUs) is inefficient and introduces complexity for development and updates. Hence, new architectures are required to consolidate and centralize the computing power into a small number of vehicle and zone controllers, moving away from distributed architectures with dozens of control units and limited communication bandwidth.
- Adopting service-oriented architecture: In order to realize Over the Air (OTA) updates, a static signal-based software architecture needs to give way to a more modular, loosely coupled Service Oriented Architecture (SOA). This has the advantages of modularity and interoperability between different types of software components providing the architecture needed for seamless software updates.
- Hardware and software abstraction through middleware: The SDV architectures abstract the software from the underlying hardware to enable software updates through a middleware, which consolidates development effort but aids scalability and reusability.
- Moving to Agile from Desktop to Cloud: Development is migrating to the Agile process, such as Continuous Development and Continuous Integration, which are required to satisfy the demand for rapid, secure OTA software updates. Additionally, cloud-based development eases provisioning of the tools to a geographically distributed workforce, provides computing power and enables streamlined access to fleet data.
How is Model-based Design evolving to support development for an SDV?
Model-Based Design is the systematic use of models throughout the development lifecycle. Model-Based Design has been instrumental in developing embedded software over the last two decades, enabling the development and delivery of real-time deterministic software in compliance with ASPICE and ISO 26262 standards. Model-Based Design creates a digital thread, “shifts-left” the vehicle development, and allows the opportunity to boost productivity through automation.
Model-Based Design workflows by MathWorks have evolved to enable SDV teams to speed up product delivery while meeting automotive requirements through early validation, software reuse, and tool integration.
Design, Simulate, and Deploy Signal-Based and Service-Oriented Applications on Middleware:
Engineers can leverage MathWorks solutions to author software applications for signal-based and service-oriented architectures, including AUTOSAR Classic and Adaptive. Developers can move existing components between AUTOSAR classic and adaptive platforms by regenerating code – saving effort and maximizing reuse. They can also integrate these applications with commercial or in-house middleware and generate production C/C++ code.
See how Zeekr Intelligent Technology uses Model-Based Design to develop SOA applications for in-vehicle OS
Automate Processes and Scale from Desktop Computers to the Cloud:
Cloud technology enables scaling by speeding up software builds and simulations, processing large data sets, and facilitating collaboration of distributed software teams. Using MathWorks solutions, engineers can integrate with CI/CD systems (like Jenkins®, Gitlab® Actions, and Azure® DevOps), process cloud-based data in systems (such as AWS® S3 and Azure Blob); scale simulation to cloud, accelerate training of neural networks on GPUs, and develop collaboratively with cloud-based software repositories.
See how AWS – Cloud-Native Development and Model-Based Approaches for Software-Defined Vehicles
Shift Left Software Integration with Virtual Vehicle Simulation:
With frequent OTA software updates, testing software using prototype hardware is no longer feasible. Integration testing also needs to be automated to reduce lead time for software releases. Shifting software integration to Model-In-the-loop (MIL) and software-in-the-loop (SIL) simulation integrated with a continuous integration pipeline ensures early and continuous testing. Engineers can use MathWorks solutions to automate the assembly of virtual vehicle models, build virtual ECUs, and deploy simulations into continuous integration pipelines.
Conclusion
The SDV requires a paradigm shift in the development of software, which will also require an evolution of the tooling landscape. MathWorks investments in architecture definition, deployment on middleware, desktop-to cloud transitioning, and advanced virtualization techniques can help the automotive industry to develop functionally safe software while also accelerating the pace of next-generation SDVs.