In the next few years, automated driving at Level 2+ and Level 3 will arrive in the vehicle class where it is not only technology that counts, but also economy: in the high-volume segment. There, the price pressure is even more pronounced than in the premium brands. It is precisely into this market that the Japanese chip manufacturer Renesas is now making inroads with the latest generation of its vehicle processors.
The SoCs of the R-Car V4H generation are designed to process numerous environmental sensor signals and, depending on the equipment and design, enable vehicles to drive autonomously, overtake, change lanes and perform similar manoeuvres. However, the driver must be able to intervene immediately at any time if electronics are no longer able to find their way around. In addition, the vehicle can park itself within certain limits. Despite all this, the Renesas chip still has enough hardware reserves to conjure up visual effects on the dashboard.
Renesas has placed special emphasis on the fact that the corresponding control computers in the car do not require active cooling: These computers are usually installed inside the habitacle of the vehicle, and car designers do not like noisy fans to cool the electronics.
- Deep Learning Platform speeds projects with Renesas R-Car SoC
- Renesas adds more deep-learning power to R-Car V3H computing platform
- Renesas expands R-Car SoC line to boost performance, integration
The performance as well as the economic handling of the V4H SoCs is achieved by a combination of sophisticated semiconductor technology and architectural tricks. For example, Renesas uses 7nm technology for the chip, together with a selection of hardware accelerators for specific tasks. This enables the device to achieve a deep learning performance of 34 Tera Operations per Second (TOPS). In addition to the lidar and radar sensors required for SAE level 3, the chip also supports up to 16 cameras for a 360° surround view.
Regarding functional safety according to ISO 26262, the SoC development process targets ASIL D systematic capability for all safety relevant IP. The signal processing portion of the R-Car V4H is expected to achieve ASIL B and D metrics for the real time domain.
Among other, the SoC contains four ARM Cortex-A76 cores with a total performance of 49KDMIPS, three lockstep ARM Cortex R52 cores for safety-critical tasks, an Image Signal Processor with parallel processing for machine and human vision and an image renderer and a Graphic Processor Unit (GPU). A selection of automotive-typical interfaces such as CAN, Ethernet AVB, TSN and FlexRay as well as two 4G PCIe interfaces enable data communication with the outside world.
Renesas also provides a dedicated power solution for R-Car V4H based around the RAA271041 pre-regulator and the RAA271005 PMIC. This enables a highly reliable power supply for the R-Car V4H and peripheral memories from the 12V supply of the vehicle battery. These features enable low power operation while targeting ASIL D compliance for systematic and random hardware faults at very low BOM cost. This helps to minimize the effort of hardware and software development while reducing design complexity, cost, and time to market.
An R-Car V4H software development kit (SDK) is also available to perform faster and easier initial device evaluation, and software development including deep learning. The SDK offers full functionality for machine learning development, and optimization of embedded systems for performance, power efficiency, and functional safety. Complete simulation models are available, and the Renesas operating system agnostic software platform enables easier development of the software-defined car.
In case of development from scratch, Fixstars’ Genesis platform enables engineers to evaluate R-Car from anywhere via their cloud solution and can provide quick and easy CNN benchmark results.
Samples of the R-Car V4H SoC are available now, with mass production scheduled for the second quarter of 2024.