Software-Defined Vehicles: Must-Have Components of a Top-Performing Car

The global market of software-defined vehicles (SDVs) is projected to expand at an impressive CAGR of 22.1% within the next seven years. What features set these cars apart? In this article, Andersen’s experts explain what areas are a must to consider when building a robust SDV solution.

The definitions of SDV vary—here is our take

Today, car owners expect their vehicles to offer the same seamless experiences as their smart devices. Software in cars must be regularly updated to enable the latest features like intelligent navigation and voice controls. It should also be synchronized with drivers’ smartphones, so that they can remotely control their cars, lock and unlock them, change climate settings, schedule maintenance appointments, and interact with their vehicles from their sofas, literally.

Software in SDVs is designed to enhance their features, tailoring them to user preferences. When defining an SDV, two aspects are crucial: the transparency of execution and consolidation of hardware. The former means that you, as a developer, can execute your software either in the vehicle, in the cloud, or else at the edge. This can be achieved through containerization and service-oriented architecture. The latter, centralization of hardware, is realized through reducing the number of electronic control units in cars from a hundred-plus to less and less.

SDVs: must-have components

Below, we will explain what areas must be taken special care of when designing and developing an SDV solution:

Surrounding infrastructure

This includes telecom infrastructure, connected roadside hardware from sidewalks, traffic lights, etc., 3D prints of cities, server-side systems of OEMs, and more. Vehicles use the data from this infrastructure to communicate with each other, learn, and upgrade. Software updates are also performed based on these insights.

China's investments in smart roads are an illustrative example of a well-organized road infrastructure that integrates vehicles with the road and the cloud. In the next 15 years, the country plans to allocate $300 billion to developing and testing roadside infrastructure in Beijing, Shanghai, Shenzhen, and other major cities for seamless autonomous driving.

Communication between vehicles and infrastructure through technologies like vehicle-to-everything, vehicle-to-car, and AI contributes to a safer environment on the roads. This communication can take place kilometers before reaching a dangerous situation. AI can then adapt the car’s ADAS to the situation created by the V2X.

The cloud environment

In SDVs, it's not just about the vehicle’s hardware. It's also about what kind of cloud is behind it. Nowadays, most of the car services are available thanks to cloud services. OEMs have their own cloud platforms, integrating their services and offering a unified experience. The platform connects the car to the edge of the backend. The cloud contributes to the enhanced maintainability, flexibility, and portability of software that comes in the form of containers.

Artificial intelligence

OEMs like XPENG, Honda, and others embed AI and machine learning into software to upgrade functionalities such as ADAS, security event analysis and knowledge management, navigation, voice controls, and infotainment.

For example, ​XPENG’s AI Valet Driver navigation can remember routes up to 100 km long. The OEM has also launched a model called MONA M03, which stands for “made of new AI.” It operates without LiDAR, with AI controlling the ADAS. Based on what it learns from the input, data, and previous drives around the city, AI can make better decisions than the driver.

Gen AI also enables natural interaction with the vehicle through voice controls. Thus, Mercedes-Benz enhanced its MBUX Voice Assistant, allowing drivers to seamlessly control dozens of central screen features by talking to the car.

The security layer

The smarter our vehicles become, the more vulnerable they are to cybersecurity risks. Connectivity, the cloud infrastructure, and personalized features are based on knowing the user. Therefore, the port of defense number one should be GDPR and obtaining user consent to share and process their data. The security of SDVs is regulated by ISO/SAE 21434:2021. Implementing key management systems, certificates, and service authentication inside the vehicle contributes to robust security. The use of open-source software can cause concerns, as, in 99% of the cases, it's implemented in vehicles. However, most of the algorithms in cryptography are open source. When we think about this, we realize that using open-source software from trusted sources and relying on professional companies to support it is secure.

Continuous Integration/Continuous Delivery

Building software for automotive effectively comes down to CI/CD and ensuring continuous compliance. The required checks should be embedded in the pipelines to guarantee superb quality. A robust CI/CD environment allows developers to integrate more and more systems into it so that they can increase the speed without losing the quality.

To sum up

The most interesting part of SDVs is envisioning and building them so that they’re still relevant in ten years. How can we do this? That’s still the question on everyone’s mind. The answer may lie in making the hardware as transparent as possible with respect to the software. This will allow offloading old functionalities and rolling out new services and features via containerization, app stores, etc. Leveraging connectivity and the cloud will help us on this journey.