AD-R1M

Open Mobile Robot Platform Reference Design

Introduction

Figure 1 AD-R1M Open Mobile Robot Platform

The AD-R1M Open Mobile Robot Platform Reference Design is a modular, extensible, and fully open-source framework developed to accelerate the design, prototyping, and deployment of autonomous mobile robots.

This reference design integrates key hardware and software components including motor control, sensor fusion, localization, navigation, and communication into a cohesive platform that supports rapid development and experimentation.

The platform supports two processing configurations:

• Raspberry Pi 5 with ADI Kuiper2 for standard deployments

• NVIDIA Jetson AGX Orin for GPU-accelerated workloads and visual SLAM

Features

• Open-source hardware and software design

• ROS 2 Humble compatible with modular node architecture

• Differential drive mobile base with encoder feedback

• Integrated ADI IMU (ADIS16470) for precision localization

• ADI Time-of-Flight camera (EVAL-ADTF3175) for 2D/3D perception

• Battery Management System with CAN-based monitoring

• Support for NVIDIA Jetson AGX Orin for GPU-accelerated workloads

• Zephyr RTOS firmware on embedded motor and BMS controllers

• CRSF/ELRS remote control with safety killswitch

• Gazebo simulation environment included

Applications

• Mobile robotics research and development

• Autonomous navigation and SLAM algorithm development

• Warehouse and logistics automation prototyping

• Educational robotics platforms

• Multi-robot setup management research

Hardware Components

The AD-R1M platform is built from modular ADI reference design boards:

Component	Board	Description	Documentation
Motor Control	ADRD3161-01Z	CAN-based motor driver with encoder interface (x2)	ADRD3161-01Z
Power Distribution	ADRD4161-01Z	Robotics perception compute carrier for Raspberry Pi	ADRD4161-01Z
Battery Management	ADRD5161-01Z	BMS board with CAN interface (3S or 12S variants)	ADRD5161-01Z
IMU	ADIS16470	High-precision 6-axis IMU for localization	imu_ros2 docs
ToF Camera	EVAL-ADTF3175	Time-of-Flight depth camera for perception	adi_3dtof

System Applications

SLAM Mapping

Figure 2 SLAM mapping with AD-R1M

Nav2 Navigation

Figure 3 Nav2 stack navigation with AD-R1M

Figure 4 AD-R1M demo navigation with elevator payload

Multi-robot Fleet

The AD-R1M supports multi-robot deployments using ROS 2 Zenoh middleware and namespacing for swarm robotics management.

Figure 5 Multi-robot control with ROS 2 Zenoh middleware

Gazebo Simulation

Full simulation environment for development and testing without hardware.

Todo

Add Gazebo simulation demo GIF

Capability Expansion

NVIDIA Jetson AGX Orin + cuVSLAM

For GPU-accelerated workloads and visual SLAM, the AD-R1M supports NVIDIA Jetson AGX Orin integration with NVIDIA Isaac ROS cuVSLAM and Intel RealSense cameras.

Key Components:

• NVIDIA Jetson AGX Orin - High-performance edge AI compute

• NVIDIA Isaac ROS - GPU-accelerated ROS 2 packages

• cuVSLAM - CUDA-accelerated visual SLAM

• Intel RealSense D455 - RGB-D camera for visual odometry

Figure 6 AD-R1M with NVIDIA cuVSLAM visual odometry

Figure 7 Gazebo simulation with cuVSLAM

The Jetson AGX Orin is powered by the ADRD5161 BMS and communicates with the AD-R1M via Ethernet using ROS 2 middleware wrappers.

BMS power connection

RPi Ethernet

AGX Orin power & Ethernet

See the AGX Orin cuVSLAM Guide for full setup instructions.

GMSL Camera Integration

The AD-R1M platform can be extended with ADI GMSL camera boards for enhanced perception capabilities in robotics and autonomous systems:

Board	Description
AD-GMSL716MIPI-EVK	GMSL2-to-MIPI deserializer for multi-camera vision systems
AD-GMSL717MIPI-EVK	GMSL2 serializer for Raspberry Pi cameras
AD-GMSL792MIPI-EVK	GMSL3/2 quad deserializer for surround-view applications

Image Processing

The ad_r1m_image_processing ROS 2 package provides real-time image processing capabilities for GMSL camera streams:

Floor Segmentation - Uses FastSAM (Fast Segment Anything Model) to identify traversable floor regions, enabling safe navigation path planning.

Figure 8 Floor segmentation with FastSAM

Object Detection - YOLO-based real-time obstacle detection with bounding boxes and class labels for obstacle avoidance.

Figure 9 Object detection with YOLO

See the ad_r1m_image_processing package for setup and configuration.

Getting Started

Todo

ADD ASSEMBLY/OTHER GUIDES

Full documentation and guides are available in the project repository:

• Quick Start Guide

• Hardware Setup Guide

• Software Guide (ROS 2, Docker)

• Use Cases and Examples

ADI ROS 2 Ecosystem

The adi_ros2 meta-repository streamlines the use of ADI packages within ROS 2 by providing a single entry point for all ADI-supported packages. It includes CI scripts for building system dependencies from source, centralized within a Docker wrapper for reproducible builds. See the official documentation for getting started.

Related ROS 2 Packages:

Package	Description
imu_ros2 (GitHub)	ROS 2 driver for ADI IMUs (ADIS16470, ADIS16475, etc.)
adi_3dtof_adtf31xx	ROS 2 driver for ADI Time-of-Flight cameras (ADTF3175D)
adrd_demo_ros2	AD-R1M platform packages and documentation

Resources

• AD-R1M Repository: adrd_demo_ros2

• ADI ROS 2 Meta-repo: adi_ros2

Related Products

Hardware Modules:

• ADRD3161-01Z - Motor control board

• ADRD4161-01Z - Robotics perception compute carrier

• ADRD5161-01Z - Battery management system

Sensors:

• ADIS16470 - High-precision 6-axis IMU

• ADTF3175D - Time-of-Flight depth camera

GMSL Camera Boards:

• AD-GMSL716MIPI-EVK - GMSL2-to-MIPI deserializer

• AD-GMSL717MIPI-EVK - GMSL2 serializer

• AD-GMSL792MIPI-EVK - GMSL3/2 quad deserializer

Supporting ICs:

• MAX32662 - Microcontroller (ADRD4161)

• MAX96716A - GMSL2 deserializer (AD-GMSL716MIPI-EVK)

Help and Support

For questions and more information about this product, connect with us through the Analog Devices EngineerZone.