Renesas — RZ/V2N Quad-Core Vision AI Microprocessors (MPUs) | Futureelectronics NorthAmerica Site
text.skipToContent text.skipToNavigation
Renesas

 

Renesas RZ/V2N Quad-Core Vision AI Microprocessor (MPU)

Featuring a 2-camera connection and excellent power efficiency

The RZ/V2N is a next-generation AI microprocessor (MPU) in Renesas’ RZ/V series, featuring the proprietary DRP-AI3 accelerator, which delivers up to 15 TOPS AI performance with industry-leading power efficiency of 10 TOPS/W - 10 times greater than previous models. With its balance of low power consumption and high AI inference capability, the RZ/V2N is ideal for Edge AI use cases such as Driver Monitoring Systems (DMS), surveillance cameras, mobile robots, and more.

Built around quad Arm® Cortex®-A55 (1.8GHz) and Arm Cortex-M33 (200MHz) cores, the RZ/V2N includes dual-channel MIPI® CSI-2® interfaces for simultaneous input from two cameras. This dual-camera support enables broader field-of-view and more accurate AI detection, crucial for modern vision systems.

The inclusion of an Image Signal Processor (ISP) ensures high-quality image preprocessing, while high-speed interfaces such as PCIe® and USB 3.2 allow seamless expansion with external devices.

Key Benefits and Features:

  • Top-Class AI Inference Performance:
    With DRP-AI3 and advanced pruning techniques, the RZ/V2N achieves real-time AI processing and up to 3x better inference performance than conventional accelerators.
  • Compact and Power-Efficient Design:
    With a 38% smaller package size than its predecessor (RZ/V2H), and 10x power efficiency improvement, the RZ/V2N supports installation in compact, battery-operated devices.
  • Accelerated Development via SOM Boards:
    Available on a compact SOM board, the RZ/V2N offers a ready-to-use platform that simplifies hardware integration, allowing developers to fast-track prototyping and product deployment.
  • Flexible and Scalable Architecture:
    The RZ/V2N shares CPU cores, the ISP, and DRP-AI3 with the RZ/V2H, enabling reusability of existing designs and reducing time-to-market and development costs across the RZ/V family.
  • Advanced I/O and Peripheral Support:
    Improved internal bus speeds enhance system performance, while PWM Timer resolution and ADC conversion speeds are optimized for precision applications.

With its versatile feature set and superior AI capabilities, the RZ/V2N delivers unmatched performance in vision AI at the edge—helping customers build smarter, faster, and more efficient AI solutions.

Features

  • Cortex-A55 (1.8GHz) x 4 cores
  • Cortex-M33 (200MHz) x 1 core
  • AI Accelerator DRP-AI
  • Arm Mali™-C55 ISP (option)
  • 1.5MB on-chip SRAM
  • LPDDR4/LPDDR4X memory interface
  • Gigabit Ethernet 2ch
  • USB 2.0 interface 1ch (Host/Function)
  • USB 3.2 (Gen2) interface 1ch (Host only)
  • PCIe interface (Gen3/2-lane) 1ch
  • MIPI CSI-2 camera interface 2ch
  • CAN interface (CAN FD) 6ch
  • AD converter 24ch
  • BGA package (15mm x 15mm 840-pin 0.5mm pitch)

Applications

  • AI-Enabled Surveillance Camera
  • RZ/V2N Based Raspberry Pi Single Board Computer
  • IMDT V2H & V2N SOM
  • IMDT V2H & V2N SBC
  • AI Dash Camera
  • Banana Pi High-Performance SoM
  • SOMDEVICES µSMARC RZ/V2N
  • Dashboard camera/driver monitoring system
  • AI monitoring camera
  • AI people counting camera
  • AI visual inspection camera
  • Robot vacuum cleaner
  • Robot lawn mower
  • Smart doorbell
  • Access control

RZ/V2N Block Diagram

 

Hardware: RZ/V2N Quad-Core Vision AI MPU Evaluation Kit

RTK0EF0186C03000BJ

The RZ/V2N AI MPU evaluation kit (RTK0EF0186C03000BJ) is used to evaluate our RZ/V2N quad-core vision AI MPU. The kit includes a CPU board, expansion (EXP) board and two sub-boards. We provide an AI Software Development Kit (AI SDK) as the software development environment for this evaluation kit. It allows for the evaluation of low-power AI inference, video streaming, and other built-in features. It also supports optional functions like the Image Signal Processor (ISP), 3D graphics engine (GE3D), and Trusted Secure IP.