Renesas RZ/V2H Robot RDK
Robotic Development Kit for high-performance AI vision and robotics on the Renesas RZ/V2H MPU.
4x Cortex-A55 @ 1.8GHz | 2x Cortex-R8 @ 800MHz | 80 TOPS DRP-AI3 | 10 TOPS/W efficiency | ROS 2 Jazzy
View on Renesas.comOverview
The Renesas RZ/V2H Robot Development Kit (WS125-V2HRDKREFZ) accelerates AI vision and robotics development on the RZ/V2H MPU. The RDK delivers up to 80 TOPS (sparse) AI inference performance via Renesas' DRP-AI3 accelerator at an industry-leading 10 TOPS/W power efficiency, alongside multi-core CPU flexibility and simultaneous multi-OS operation. The kit ships with Yocto 5.1 (Styhead) and Ubuntu 24.04 support and includes ROS 2 Jazzy for robotics workflows. Quad Cortex-A55 application cores, dual Cortex-R8 real-time cores, and a Cortex-M33 for system management unify vision AI and real-time motor/sensor control in a single platform — purpose-built for next-generation robotics, drones, and edge AI vision systems.
Specifications
| Specification | Value | Notes |
|---|---|---|
| Application CPU | 4x Arm Cortex-A55 @ 1.8GHz | |
| Real-Time CPU | 2x Arm Cortex-R8 @ 800MHz | |
| System CPU | 1x Arm Cortex-M33 @ 200MHz | FPU + DSP |
| AI Accelerator | DRP-AI3 (up to 80 TOPS sparse) | |
| AI Efficiency | 10 TOPS/W | |
| Image Processing | DRP (HW-accelerated OpenCV) | |
| GPU | Mali-G31 (select variants) | |
| Memory | LPDDR4/DDR4 | |
| Storage | eMMC, xSPI, SD card | |
| High-Speed I/O | PCIe, USB 3.2, 2x Gigabit Ethernet | |
| Camera | MIPI CSI-2 | |
| Display | HDMI (via expansion board) | |
| OS Support | Yocto 5.1 (Styhead), Ubuntu 24.04 | Simultaneous multi-OS operation |
| Robotics Stack | ROS 2 Jazzy | |
| Debug | Onboard JTAG, DIP switches | Flexible power architecture |
Use Cases
- Robotics development and deployment with ROS 2 Jazzy and real-time motor/sensor control on the Cortex-R8 cores
- UAV/drone flight control systems combining edge AI vision with deterministic real-time control
- AI-enabled vision processing for autonomous machines, AGVs, and AMRs
- Industrial vision systems requiring hardware-accelerated OpenCV alongside neural network inference