Overview

The RDK X5 AI Development Board, developed by D-Robotics, is a high-performance edge computing platform designed for robotics, autonomous systems, and embedded AI vision applications. Powered by the Sunrise 5 System-on-Chip (SoC), the RDK X5 integrates advanced AI processing, robust connectivity, and versatile I/O options in a compact 85mm × 56mm form factor. This article provides a detailed technical overview of the board’s capabilities, specifications, and practical applications for developers and engineers.

Key Features

The RDK X5 is engineered to support real-time AI inference, machine vision, and sensor fusion at the edge, eliminating reliance on cloud computing. Its standout features include:

  • Sunrise 5 SoC: Combines 8 Arm Cortex-A55 CPU cores (1.5GHz), a 10 TOPS Brain Processing Unit (BPU), and a 32 GFLOPS GPU for efficient multitasking, deep learning inference, and graphics processing.

  • Dual 4-lane MIPI-CSI Camera Interfaces: Supports stereo vision for 3D mapping, depth estimation, and visual odometry, ideal for robotics and spatial AI.

  • NodeHub Integration: Access to over 200 plug-and-play AI and robotics modules for rapid development of features like SLAM, object tracking, and voice recognition.

  • Connectivity: Wi-Fi 6, Bluetooth 5.4, and Gigabit Ethernet with Power-over-Ethernet (PoE) for flexible networking in embedded and industrial environments.

  • Versatile I/O: Includes 4 USB 3.0 ports, a 40-pin GPIO header (Raspberry Pi-compatible), CAN FD, HDMI 1.4, and MIPI-DSI for extensive peripheral integration.

  • Flash-Connect USB-C: A single USB-C interface for power, flashing, debugging, and display output, simplifying development workflows.

  • Compact and Robust Design: Passive cooling with a pre-installed heatsink supports operation from -20°C to +60°C, suitable for drones, robots, and space-constrained systems.

Technical Specifications

Processor and Memory

  • CPU: 8 x Arm Cortex-A55 @ 1.5GHz

  • BPU: 10 TOPS (INT8) for real-time AI inference

  • GPU: 32 GFLOPS for hardware-accelerated graphics

  • Memory: LPDDR4 (4GB or 8GB configurations)

  • Storage: 1 Gbit NAND flash, MicroSD card slot (UHS-I mode)

Camera and Display Interfaces

  • Camera: 2 x 4-lane MIPI-CSI (v2.1) supporting high-speed video input and stereo vision

  • Display: 1 x 4-lane MIPI-DSI (v1.2), 1 x HDMI 1.4 (up to 1080p@60fps)

I/O and Connectivity

  • USB: 4 x USB 3.0 Type-A, 1 x USB 2.0 Type-C (device), 1 x USB 2.0 Micro-B (UART debug)

  • GPIO: 40-pin header with SPI, I²C, I²S, PWM, UART (3.3V)

  • CAN: 1 x CAN FD interface for industrial/automotive communication

  • Networking:

    • Wired: 1 x RJ45 Gigabit Ethernet with PoE

    • Wireless: Dual-band Wi-Fi 6 (802.11ax), Bluetooth 5.4, onboard antenna, IPEX connector

  • Audio: 3.5mm stereo jack, Cadence HiFi 5 DSP (voice wake-up, PDM, I²S)

Power and Physical

  • Power Input: 5V/5A via USB-C or GPIO pins

  • Power Output: 5V and 3.3V for external components

  • Dimensions: 85 x 56 x 20 mm

  • Operating Temperature: -20°C to +60°C

  • Cooling: Passive with pre-installed heatsink

Software Ecosystem

The RDK X5 runs Ubuntu 22.04 LTS, providing a stable Linux environment optimized for embedded AI and robotics development. It supports popular frameworks like TensorFlow, PyTorch, and ONNX, enabling seamless deployment of machine learning models. The NodeHub platform accelerates development with over 200 modular components for tasks such as SLAM, path planning, and sensor fusion. Additionally, the board supports ROS2 (Robot Operating System) and RDK Studio, offering tools for rapid prototyping, debugging, and deployment.

Example Workflow

  1. Setup: Flash the RDK X5 OS to a MakerDisk microSD card using the Flash-Connect USB-C interface.

  2. Development: Use Python or ROS2 to integrate NodeHub modules for vision or navigation tasks.

  3. Testing: Connect compatible CSI cameras (e.g., Raspberry Pi 8MP Camera Module V2 or IMX219 Binocular Camera) for stereo vision applications.

  4. Deployment: Leverage Wi-Fi 6 or PoE-enabled Ethernet for headless operation in real-world environments.

Applications

The RDK X5 is tailored for edge AI applications requiring low latency and high reliability. Key use cases include:

  • Autonomous Robotics: Real-time SLAM, obstacle avoidance, and path planning for mobile robots and drones.

  • Industrial AI: Smart inspection, predictive maintenance, and object detection in manufacturing.

  • Smart Surveillance: On-device face recognition, human detection, and license plate recognition.

  • AI Education and Research: Hands-on learning with real hardware for deploying AI models.

  • IoT Edge Devices: Local deep learning inference for smart sensors and controllers.

Case Study: AI-Guided Mobile Robot

A notable project integrates the RDK X5 with the OpenAI API to create a mobile robot with natural language interaction. The Sunrise 5 SoC processes sensor data and controls hardware, while Wi-Fi 6 enables internet connectivity for API calls. The robot uses dual MIPI-CSI cameras for visual perception and NodeHub modules for navigation, demonstrating the board’s ability to handle complex, AI-driven tasks on-device.

Compatible Hardware

CSI Cameras

  • Raspberry Pi 8MP Camera Module V2

  • 5MP OV5647 Camera Module

  • 8MP USB Camera with Housing

  • 8MP IMX219 Binocular Camera

  • 8MP Sony IMX219 Camera Module

Packing Options

  • Mainboard Only: RDK X5 (4GB or 8GB RAM)

  • Mainboard + Basic Kit: Includes USB-C power adapter, HDMI cable, USB-C cable, and 64GB MakerDisk microSD

  • Mainboard + Complete Kit: Adds RDK X5 Camera Module RS800W and metal case

Connectivity Setup

To connect the RDK X5 to a home network:

  1. Configure Wi-Fi 6 settings via Ubuntu 22.04’s network manager or command-line tools like nmcli.

  2. For wired connections, use the Gigabit Ethernet port with PoE for power and data in industrial setups.

  3. Bluetooth 5.4 can pair with peripherals like sensors or controllers.

Note: Wi-Fi and Ethernet settings must be manually configured to match your network.

Development Tips

  • Camera Integration: Use dual MIPI-CSI ports for stereo vision tasks. Ensure cameras are compatible with MIPI CSI-2 v2.1.

  • Power Management: For mobile applications, use a USB-C PD adapter (5V/5A) or battery via GPIO pins.

  • Thermal Considerations: The passive heatsink supports fanless operation, but ensure adequate ventilation in high-temperature environments.

  • NodeHub: Leverage pre-built modules to reduce coding time for common robotics tasks.

Conclusion

The RDK X5 AI Development Board is a versatile, high-performance platform for edge AI and robotics. Its powerful Sunrise 5 SoC, dual-camera support, and extensive I/O options make it ideal for developers building autonomous systems, industrial AI solutions, or IoT devices. With Ubuntu 22.04, NodeHub, and ROS2 support, the RDK X5 streamlines development, enabling rapid prototyping and deployment. Whether you’re a researcher, educator, or professional, the RDK X5 provides the tools to bring intelligent, real-time systems to life.