New ECU Architectures, Open-Source Initiatives Shift SDVs Into Overdrive

From offering advanced safety and convenience features to enabling personalized in-cabin experiences, software-defined vehicles (SDVs) are taking the automotive industry and the global consumer market by storm. Indeed, many car manufacturers increasingly rely on software to differentiate a wide range of sophisticated systems, from infotainment and head-up displays (HUDs) to haptic touch interfaces and advanced driver-assistance systems (ADAS).

According to Pedro Pacheco, senior research director at Gartner, the software will become the main profitability growth driver for automakers as more consumers view digital technology as a purchase differentiator. Tesla, for example, sells monthly subscription services for various features, such as its Premium Connectivity offering with live traffic visualization, video streaming, satellite-view maps, and more. Similarly, BMW monetizes software-driven capabilities, such as advanced driver-assist capabilities and adaptive suspension, via subscriptions in some regions.

Electrification, plus greater levels of automation and connectivity, are transforming what consumers worldwide expect from their cars and how they are designed. As Simon Humphries, the chief branding officer of Toyota, recently noted to The Next Web: “People want control over their own experiences.” To bring these advanced functions and features to life, it’s not uncommon to find up to 150 million lines of software code distributed among more than 100 electronic control units (ECUs), as well as in sensors, cameras, radar, and LiDAR devices.

Read on to learn how automakers are shifting SDVs into overdrive with new ECU architectures, open-source projects, and cross-industry collaboration.  

Streamlining Automotive System Design

When it comes to software, every automotive manufacturer wants to spend less time on what are becoming commodities—such as the core operating system and components connecting the various pieces—and focus on features that will differentiate the brand. For many automakers, the open-source model supports this objective by expediting agile product development for certain systems and functions. As Canonical’s Bertrand Boisseau observes in Automotive World, open source software, with its flexibility and scalability, will be key to building cars fit for the future.

However, the automotive industry first needs to reassess electrical/electronic (E/E) architectures to pave the way for a new generation of SDVs that will more efficiently leverage both open-source and proprietary software to create differentiated experiences. As ABI analyst Dylan Khoo writes in a recent blog post, systems based on traditional E/E architectures aren’t well-suited for SDVs due to the many isolated ECUs they contain. To be sure, these disparate ECUs typically run independently with their own data connections, processing, and software.

That’s why automotive manufacturers are implementing new domain-based and zonal E/E architectures to streamline system design and functionality. For example, domain-based architectures consolidate and link function-specific ECUs to domain controller units (DCUs), which are then logically grouped to manage sub-domain functions and systems such as powertrain, infotainment, ADAS, and passenger comfort.

Similarly, zonal E/E architectures leverage Ethernet-connected zonal controllers to manage different body sections of the vehicle. This localized approach significantly reduces the number of ECUs and cables by implementing a centralized controller for each vehicle section. Indeed, the flexible zonal architecture allows a single ECU to manage multiple vehicle functions across different domains by simultaneously running software from different suppliers on the same silicon.

Virtualizing ECU Testing with Digital Twins
Evolving E/E architectures have spurred automakers to design new low-latency, high-bandwidth ECUs with more advanced processing capabilities. However, it is increasingly costly and time-consuming to physically test complex ECUs and accompanying open-source and proprietary software stacks via traditional methodologies.

Virtualizing ECUs shifts testing from the road to a more efficient, cost-effective, and deterministic environment where results are accurately analyzed in minutes. Creating “digital twins” allows developers to collaboratively identify potential issues and eliminate bugs before they are inadvertently installed in hardware. This virtual paradigm also enables developers to test code changes continuously and almost immediately determine if over-the-air (OTA) firmware and software updates securely download, install, and function as planned. Additionally, advanced features such as silicon lifecycle management (SLM) and in-field analytics can be closely monitored and safely evaluated in simulated, real-world conditions.

Many automotive companies now test and optimize silicon in virtualized environments where every aspect of the ECU and accompanying software stack—including interactions with multiple vehicle systems—are holistically simulated. In short, virtualized ECU testing is playing a crucial role in jumpstarting innovation while accelerating the development of sophisticated automotive systems that deliver more interactive in-cabin experiences and advanced self-driving capabilities.  

Accelerating Automotive Innovation with Open-Source Initiatives
To further evolve automotive systems and accompanying software stacks, leading automakers, semiconductor companies, and cloud leaders are collaborating to develop the Scalable Open Architecture for Embedded Edge (SOAFEE) platform. According to the SOAFEE website, the platform builds on technologies like Project Cassini and SystemReady (which define standard boot and security requirements for Arm architecture) by adding a cloud-native development framework while introducing additional functional safety, security, and real-time capabilities.

Specifically, SOAFEE offers standards-based firmware and security interfaces that enable seamless secure boot and system bring-up across all compliant hardware. SOAFEE also includes a reference framework for existing open standards and specifications to standardize key non-differentiating middle layers, such as the hypervisor, operating systems, container runtime, and hardware abstraction layers. Because of its open-source components, any middleware or software provider can plug into the solution stack, or even provide the elements of the stack. Easily tapping into SOAFEE’s open architecture allows developers and designers to focus on creating differentiated automotive designs and in-cabin experiences.  

Autoware, another cross-industry, open-source initiative, seeks to lower the entry barrier to autonomous driving for various applications such as robot-taxis, cargo delivery, and autonomous valet parking (AVP). According to the Autoware website, the modular Autoware Core is built on the Robot Operating System (ROS) and includes all APIs and functionality required for autonomous driving, from perception and localization to planning and control.

Perhaps most importantly, Autoware Core implements collaborative software engineering practices, including pull request reviews and builds, comprehensive documentation and code coverage, a coding style guide, and well-defined development and release processes. Autoware Universe further accelerates the development of autonomous driving capabilities by making it easy for automakers to design, test, and rapidly integrate new and differentiated functionality for complex use cases such as robot-taxis operating in dense urban environments or high-speed autonomous racing.

Conclusion
Software is fast becoming a profitability growth driver for automakers as consumers increasingly view digital technology as a purchase differentiator. Software powers a wide range of advanced automotive systems, from infotainment and HUDs to haptic touch interfaces and ADAS. To pave the way for a new generation of SDVs, automotive manufacturers are implementing new ECU architectures while exploring open-source projects and expanding cross-industry collaboration.