Renesas Electronics Reduces Development Time for NCAP Front Camera Applications for Automotive with R-Car V3M-Based Solution

• R-Car V3M入门套件可以显著减轻开发强度并加快上市时间

• R-Car V3M now embeds CNN accelerator, enabling high-performance convolutional neural networks with ultra-low power consumption in NCAP front camera applications

Renesas Electronics Corporation (TSE: 6723) today announced the R-Car V3M Starter Kit, which simplifies and accelerates the development of front camera applications, surround view systems, and LiDAR for the New Car Assessment Program (NCAP, Note 1). Based on the R-Car V3M image recognition SoC, the new starter kit provides a low-power and high-performance solution for the growing NCAP front camera market. By combining the R-Car V3M Starter Kit with supporting software and tools, system developers can easily develop front camera applications, helping to reduce development workload and shorten time to market.

Renesas also introduced an enhanced version of the R-Car V3M, which is gradually emerging in the field of automotive applications by integrating a hardware accelerator for energy-efficient convolutional neural networks (CNN, Note 2), enabling functions such as road detection or object classification. The R-Car V3M's innovative hardware accelerator enables CNN to run at ultra-low power levels that cannot be achieved when running on a CPU or GPU.

The new R-Car V3M starter kit, R-Car V3M SoC and supporting software and tools such as the Renesas open-source e2studio integrated  development environment (IDE) are all part of the open, innovative and reliable Renesas autonomy™ Platform, which provides a comprehensive end-to-end solution from the cloud to perception and vehicle control for advanced driver assistance systems (ADAS) and automated driving applications.

“瑞萨电子凭借R-Car V3M入门套件可以显著降低开发强度并加速上市时间，满足快速增长的NCAP前置摄像头市场需求。”瑞萨电子副总裁Jean-Francois Chouteau表示。“随着基于R-Car V3M SoC的入门套件的推出，我们与HELLA Aglaia合作开发出前置摄像头应用软件，本套件为开发一流的前置摄像头解决方案提供了重要支持。

Key Features of the R-Car V3M Starter Kit

Flexible deployment of development boards for a variety of computing applications

The new starter kit is a ready-to-use kit that includes the necessary interfaces and tools. The kit provides core components for ADAS and autonomous driving development, such as 2GB RAM, 4GB eMMC (embedded multimedia memory card, Note 3) on-board memory, Ethernet, display output, and debug interface. The integrated 440-pin expansion port provides system manufacturers with full flexibility to develop dedicated expansion kits for a variety of computing applications, from simple advanced computer vision development environments to multi-camera prototypes such as surround view systems. This flexibility greatly reduces the time required for hardware development while maintaining a large degree of software portability and reusability.

Quick start solution supported by Linux BSP and multiple tools

The R-Car V3M starter kit is already equipped with a Linux BSP, which has been released through elinux.org. Other commercial operating systems will be developed next year.

Codeplay will add OpenCL™ and SYCL™ models to this starter kit in Q1 2018. More tools, sample code, and application notes for computer vision and image processing will follow in 2018.

Renesas has integrated a variety of tools such as the Renesas e 2 Studio tool chain and debugging tools into the R-Car V3M starter kit to reduce development burden and shorten time to market.

Partners, including Cogent Embedded, which develops a 3D surround-view parking assist system, have integrated their field-proven software solutions into the R-Car V3M starter kit to help customers accelerate their development process.

In addition to the R-Car V3M Starter Kit, Renesas has also integrated a highly energy-efficient CNN network for image recognition and image classification on the R-Car V3M SoC. The R-Car V3M Starter Kit uses a high-performance, low-power CNN network in the NCAP camera, and this hardware accelerator cannot be replaced by traditional high-power CPU or GPU architectures. Renesas has completed the IMP-X5 (Computer Vision Processing Subsystem), which consists of an image processor, a programmable CV engine, and a new innovative CNN hardware accelerator developed by Renesas, which is ideal for deploying high-performance CNN networks at ultra-low power. With this new IP, Renesas allows system developers to choose between the IMP-X5 and the new hardware accelerator when deploying CNNs. This heterogeneous approach allows system developers to choose the most efficient architecture based on the required programming flexibility, performance, and power consumption.

Availability

The R-Car V3M SoC is available for order now.

The R-Car V3M starter design software with Linux BSP will be released in limited quantities in the first quarter of 2018. The full version with enhanced software is expected to be available in the third quarter of 2018.

Please refer to the product specification sheet of the product R-Car V3M Starter Kit.

(Note 1) New Car Assessment Program (NCAP):

A government-run automotive safety program that evaluates how new vehicle designs perform against various safety threats.

(Note 2) Convolutional neural networks (CNNs) are deep feed-forward artificial neural networks that have been successfully used to analyze visual images and are increasingly being used in automotive applications such as road detection or object classification.

(Note 3) An embedded MultiMediaCard (eMMC) can be used to boot the operating system. This is much faster than booting over the network.