www.embedded.com, Oct. 24, 2024 –
The convergence of several technological advancements, such as Artificial Intelligence, software, and electric vehicles, has deeply transformed the automotive industry from its traditional principles.
A Software-Defined Vehicle (SDV), is a game changer. That's the reason why the market size of SDVs is expected to register a compound annual growth rate of 22.1% between 2023 and 2032.
But what does the architecture of an SDV look like? Let's review the parts that power SDVs and understand how they're shaping the future of transportation.
Architectural Shifts in SDVs: From Legacy to Dynamic
In traditional automotive designs, vehicles have several ECUs, and each of these is responsible for one specific function. For example, you have an ECU for the radio, another for the brakes, another for the infotainment... you name it.
The result is that the car has a very complex web of wiring and hardware that it's not easy to maintain, and that is rigid: it doesn't allow for quick updates when it's needed, and it can go obsolete in very little time. Besides, the driver can't personalize it as much as they might need to.
SDVs architectures evolved past this, and they've been designed for flexibility. The control is centralized, and it separates hardware from software. If you integrate the software into cloud and edge computing, SDVs can be updated easily. This approach also offers more control for the user and advanced functionalities.
The Importance Of Centralized Architectures For SDVs
With a centralized architecture, you replace the traditional ECUs with a few high-performance computers, and these manage several functions at the same time. These computers can run complex software stacks; the best part is that they can be updated remotely. The advantages are clear: less hardware, less weight, and lower energy consumption.
Another good aspect of a centralized architecture is the possibility to perform complex processing tasks. For example, ADAS, or Advanced Driver Assistance Systems, require high amounts of data processing power. It is the same with autonomous driving. Centralized systems can manage this data in real-time, something crucial for these kinds of vehicles.
These high-performance computing units (HPCUs) are built on advanced processors capable of handling data from various sensors (e.g., LIDAR, radar, and cameras) and making split-second decisions.
Zonal Architectures: Simplifying Complexity through Regionalization
If there's one crucial difference between SDVs and traditional vehicles, it's zonal architectures. Zonal systems group vehicle functions in different areas of the car, such as the rear, the front, or the cabin. Each of these areas has a central hub, and this hub communicates with sensors and actuators within its region while keeping a connection to the central control unit.