Software Defined Vehicle (SDV) is a trending topic and is extensively discussed across the internet and social media. While SDV is not a new term to the industry, and it has been under discussion & development for the last 25 years, the SDV concept gained significant momentum in the market from Tesla, where the automotive industry started believing something called SDV is possible. Everyone is trying to add value to the SDV concept in terms of software. Various groups have formed online to discuss and propose ideas for making SDV a reality. Some groups are proposing best-fit architecture for SDV [1], while others are producing Open-Source software (OSS) [2] and how that is the best fit for SDV. All these groups are not only discussing but also demonstrating their concepts.
The image above shows a Classic OEM car. You can see how complex the connection network between the ECUs is, with various communication protocols like CAN, LIN, MOST, FlexRay, etc. Additionally, due to ever-increasing customer demand for more features, more ECUs are being added to which in term increases complexity in managing data load, weight, cost, and is still difficult to meet the requirement. This complexity is getting difficult for OEMs to manage from the Mfg., maintenance, and cost perspective. These SDVs enable automotive manufacturers to deliver a tailor-made driving experience by augmenting through software, as we see in cell phones and other digital gadgets.[3] [4].
Zonal architecture shown above is one of the most achievable ways to reduce ECU complexity and thereby reduce network connections between various ECUs, which minimizes weight, cost, and complexity compared to Classic OEM architecture. SDV uses a central vehicle computing unit, also known as HPC (High Performance Computer), where all the heavy CPU load calculations, decision making, and cloud communications are managed in one place, thus reducing the load on zonal ECUs. Zonal ECUs are only responsible for capturing data and sending it back to the HPC, which in turn will process data and make decisions that will be communicated to zonal ECUs. The presence of a telematics unit enables cloud connectivity to keep track of vehicle maintenance/diagnostics, which makes the car Internet of Things. Since Zonal architecture includes HPC and cloud connectivity, SDV becomes more software-centric and opens multiple doors for the automotive industry for features and functions. [5] [6].
What are Software-Defined Vehicles?
Software-defined vehicles are vehicles whose features and functions are driven and managed by software rather than traditional mechanical systems and hardware. [7] [8].
In a classic vehicle, the development of the vehicle stops at the Start of Production, whereas in an SDV vehicle evolves continuously after manufacturing, like our smartphones through software updates to enhance performance and experience new features.
The presence of the telematic unit interface with HPC enables the connectivity of the vehicle to everything (V2X), allowing the vehicle to communicate with infrastructure, cloud, and other vehicles. This connectivity enhances the operational efficiency, real-time traffic routing, various infotainment options, active diagnostics, and improves the driving experience. [9] [10] [11].
How SDV Works
SDV brings into service software software-centric architecture that differentiates SDV from classic vehicles. On the other hand, in the classic automotive industry, features are largely fixed and determined at the point of SOP, whereas SDVs harness software to perform all operations, add functionalities and features to enhance user experience. [12] [13].
SDV development continues throughout the life cycle of the vehicle. SDV structure makes it capable of evolving continuously, like we see in smartphones, software updates come at regular intervals to enhance features and fix bugs, if any. OEM continues to get revenue even after SOP. [14].
The architecture of the SDV is designed in such a way as to separate non-essential applications from safety-critical tasks. Non-essential applications manage infotainment and connectivity, whereas the RTOS handles safety-critical tasks. Containerization and hypervisors are deployed in a way that separates various software components. This ensures that updates from over-the-air (OTA) can be implemented without compromising the integrity of safety systems. [15].
Challenges
Despite a promising outlook, there are challenges in the evolution of SDV that must collectively collectively addressed and resolved to see SDV in reality. Automotive industry development differs from other industries due to extremely strict functional safety requirements, hard real-time software requirements, multiple layers of integration, and multiple suppliers involved in development. Another critical issue in SDV is security, since the vehicle is continuously connected to the cloud.
Solutions
There is a need for global automotive communities that should come together to work on distributed regulations, country by country, either create a new one or work on the existing regulations to streamline them to match SDV development.
Tremendous efforts are going on globally to achieve this through DevOps. Zonal architecture is a proposed solution to speed up development activities. This will reduce complexity and hence will reduce the development time.
There are multiple groups worth mentioning, including COVESA, the Zephyr project for OSS hosted by the Linux Foundation, working tirelessly to contribute SDV developments. [16] [17].
Conclusion
The SDV model addresses the automotive industry’s required focus on sustainability and efficiency. SDV constitutes the next rim in the automotive industry of controlling the majority of functions and features through software, which will allow continuous enhancement and upgrades post SOP. [18] [19].
In the future, there may be trends that suppliers only need to provide the software to OEMs in the form of an executable like object file or models, or .a or a2L and OEMs will integrate at their end. Due to this central HPC, the number of ECUs in the vehicle will be reduced to just two digits.
The global market for SDV is projected to grow dramatically from $219.7 billion in 2024 to $489.7 billion by 2030. A compound annual growth rate (CAGR) will remain approximately 18.79% during the period from 2025 to 2032 in the United States alone. [20] [21].
All the SDV challenges can be conquered by fostering collaborative innovation, the industry can build smarter, safer, and more resilient SDVs.