Yahboom ROSMASTER M3 ROS2 AI Large Model Robot Car w/ Mecanum Wheels for Orin Nano/NX SUPER, RDK X5, Pi 5

Overview

The Yahboom ROSMASTER M3 is a ROS2 robot car platform designed for Jetson Orin Nano/Orin NX SUPER, Raspberry Pi 5, and RDK X5. It integrates multimodal AI (text/vision/voice) with SLAM navigation, and features a Mecanum wheel chassis with a pendulum-style independent suspension structure for 360° omnidirectional movement. Depending on configuration, it supports optional single/dual TOF LiDAR and uses a DaBai DCW2 depth camera for 3D vision applications.

Key Features

• AI multimodal large language model applications: semantic understanding, speech dialogue, and scene understanding
• Dify workflow development platform support for developing and deploying large-model workflows
• Dual-model inference architecture with dynamic feedback inference and conversation interruption support
• LiDAR + encoder + IMU (gyroscope) fusion for mapping and navigation; supports multiple mapping algorithms
• DaBai DCW2 depth camera: depth image + point cloud for 3D vision mapping, measurement, and recognition
• Professional-grade Mecanum wheels + pendulum suspension to reduce wheel slip impact on encoder recognition and reduce odometer error
• Integrated RGB headlights/LED strip with flowing, breathing, and marquee lighting effects; customizable colors/brightness
• AI vision stack support: OpenCV / MediaPipe / YOLOv11; includes functions such as gesture recognition, QR code recognition, pose estimation, image segmentation, and object detection
• Multi-robot formation and interconnection control: multi-robot navigation and dynamic obstacle avoidance on the same map; multiple robots controlled by one host

Specifications

Version Options (Configuration Selection)

Note: Only the Ultimate version is configured with Dual T-mini Plus LiDARs.

Controller Selection Suggestions (Reference)

To improve large-model operation smoothness and functional results, selecting Jetson Orin Nano/NX SUPER is recommended. If choosing a version without a board, prepare a Raspberry Pi 5 with at least 8GB of RAM.

Performance Reference (Functional Case Test Comparison)

Functions (LiDAR / Depth Camera / Vision)

LiDAR Functions

• High-precision TOF LiDAR with encoder and IMU (gyroscope) fusion data for high-precision mapping and navigation
• Supports multiple mapping algorithms and Archive Mapping
• Supports single-point and multi-point navigation; can be operated via an APP
• Relocation navigation technology reduces positioning drift, improving navigation stability and reliability
• Mapping and navigation modes shown: Gmapping LiDAR mapping, Cartographer LiDAR mapping, slam_toolbox LiDAR mapping, IMU LiDAR fusion filtering, APP mapping navigation
• Example behaviors shown: LiDAR obstacle avoidance, LiDAR following, LiDAR guardian, road network planning

Depth Camera Functions

• 3D structured light depth camera generating depth images and point cloud data
• Depth distance and volume calculation; constructs high-precision 3D color maps when combined with radar data
• Example applications shown: RTAB-Map 3D vision mapping and navigation, wood block volume measurement, edge detection, depth camera distance measurement

YOLOv11 Model Detection

• Supports image segmentation, pose estimation, image classification, and oriented object detection

AI Visual Recognition / Interaction

• Supports frameworks such as OpenCV and MediaPipe
• Recognition examples shown: human feature recognition, gesture recognition, finger tip trajectory recognition, QR code recognition, 3D detection, 3D face detection, color recognition, AR vision
• Interaction examples shown: gesture control, MediaPipe posture following, machine code control, visual line tracking, color tracking, face tracking, KCF object following, deep learning object tracking

Autonomous Driving (Sandbox) Notes

Autonomous driving sandbox testing is shown as supported on: RDK X5, Orin Nano, and Orin NX. Raspberry Pi boards are shown as not supporting this function. Functions demonstrated include road sign detection, lane keeping, autonomous parking, and steering decision.

Applications

• SLAM mapping and navigation
• Road network planning, route planning, and multipoint navigation
• Scene understanding, visual following, deep distance Q&A, and autonomous cruise demonstrations
• Multi-robot synchronous motion control and formation control

Tutorials

ROSMASTER-M3 Tutorials

For configuration help before purchase (versions, controller selection, and accessories), contact https://rcdrone.top/ or email support@rcdrone.top.

Details

Meet ROSMASTER M3: a ROS2-ready robot car platform built for multimodal AI and SLAM navigation on popular edge controllers.

Multimodal interaction, 3D perception, and omnidirectional mobility come together in a single integrated platform.

Dify workflow support and multiple mapping options help move from demos to deployable robotics applications.

Choose the right kit level by comparing perception sensors, controller compatibility, and chassis performance.

Optional single/dual TOF LiDAR and programmable RGB lighting expand navigation and presentation use cases.

Run text, voice, and vision models together for richer semantic understanding and interactive robotics.

A practical vision stack supports tracking, recognition, and interactive Q&A for real-world scenarios.

SLAM workflows cover mapping, point-to-point navigation, and task-oriented exploration.

Higher-level planning combines perception and mapping to execute step-by-step tasks more reliably.

Use the selection guide to match your controller and sensor needs across Standard, Superior, and Ultimate options.

Sensor fusion and ROS tooling support mapping, obstacle avoidance, and depth-based measurement.

Vision features include detection, tracking, gesture recognition, and multi-robot formation control.

Autonomous driving behaviors include lane keeping, sign recognition, parking routines, and steering decisions.

ROS2 Humble development pairs with RViz simulation and flexible remote-control options for testing and demos.

An exploded view highlights modular add-ons like the depth camera, LiDAR, optional display, and onboard lighting.

The ROS robot control board bundle includes a 12V 6000mAh Li-ion battery pack and supports an optional 7-inch HD touchscreen for interactive control.

The ROSMASTER M3 course syllabus lays out the video lesson modules and learning roadmap for ROS2 AI robot projects.

The ROSMASTER M3 package includes organized tutorial and code folders covering chassis control, LiDAR setup, and AI model development topics.

The ROSMASTER M3 learning resources outline AI large model tutorials, ROS2 basic course videos, and practical materials to guide setup and development.

Yahboom provides ROSMASTER M3 3D model files and after-sales technical support to help with DIY modeling and setup.

ROSMASTER M3 platform options cover Ackermann steering, RGBD/USB camera choices, a 0.91-inch OLED display, and multiple control board selections.

ROSMASTER M3 uses a mecanum-wheel chassis with 80 mm wheels and lists options like an AI voice module, multiple controller boards, and a 12.6V 6000mAh battery.

ROSmaster M3 uses a mecanum-wheel chassis with multiple camera and control board options, plus a 12.6V 6000mAh battery pack for mobile builds.

ROSMaster M3 PRO combines a mecanum-wheel chassis with a 6‑DOF robotic arm and supports LiDAR, depth camera, and Raspberry Pi or Jetson control boards.

The ROSMASTER M3 specification sheet includes dimension drawings and key details like ROS2 support and Python programming.

The ROSMASTER M3 kit includes the robot chassis along with core electronics, sensors, and essential cables and accessories for assembly.

The ROSMASTER M3 accessory lineup includes LiDAR and depth camera modules, a 7-inch screen with brackets, mounts, and different main control board bundles.