The UGV Beast ROS 2 Open-source Off-Road Tracked AI Robot is a versatile platform featuring dual controllers, computer vision capabilities, and an all-metal body. It's designed for flexibility and expandability, making it ideal for use with Raspberry Pi 4B and Raspberry Pi 5.
Beginner Selection Recommendation
For users looking to utilize a Raspberry Pi as the host controller, we recommend the UGV Rover PT PI5 ROS2 Kit. If an NVIDIA Jetson Orin Nano is preferred, consider the UGV Rover PT Jetson Orin ROS2 Kit.
Both kits offer exceptional reliability, a high-torque 2-axis Pan-Tilt, D500 Lidar, and an OAK-D Lite depth camera. They come with comprehensive basic functionalities and ROS2 tutorials, empowering beginners to quickly grasp and apply robotics knowledge. These robot kits boast superior expandability, allowing for diverse application deployments.
Product Model Name Introduction
Our product model names are structured to convey key features: chassis type, Pan-Tilt option, host controller model, and function type.
| Model Name | Description | Corresponding Module | |
|---|---|---|---|
| Chassis Type | UGV Rover | 6 Wheels 4WD chassis | |
| UGV Beast | Independent suspension off-road tracked chassis | ||
| Pan-Tilt Option | PT | with 2-axis Pan-Tilt | |
| Host Controller | Jetson Orin | Using Jetson Orin Nano as host controller. Note: "Acce" kits do not include the host controller. | |
| PI 5 | Using Raspberry Pi 5 as host controller. Note: "Acce" kits do not include the host controller or heatsink. | ||
| PI 4B | Using Raspberry Pi 4B as host controller. Note: "Acce" kits do not include the host controller or heatsink. | ||
| Acce | Kits with "Acce" do not include the host controller, suitable if you already own a host board. | ||
| Function Type | AI Kit | Basic AI kit with a 5MP ultra-wide-angle camera, includes computer vision functionality and tutorials. | |
| ROS2 Kit | Expands on the AI Kit with a D500 360° Lidar and OAK-D Lite depth camera, offering complete ROS2 functionality and tutorials. | ||
Features at a Glance
The UGV Beast ROS2 Kit is an AI robot designed for exploration and creation. With exceptional expansion potential, based on ROS 2 and equipped with Lidar and a depth camera, it seamlessly bridges your imagination with reality. It's the ideal choice for tech enthusiasts, makers, or programming beginners eager to explore the world of intelligent technology.
Equipped with high-performance Raspberry Pi computers, the UGV Beast meets the challenges of complex strategies and functions, inspiring your creativity. Its dual-controller design combines the high-level AI functions of the host controller with the high-frequency basic operations of the sub-controller, ensuring accurate and smooth performance.
Control is effortless via the UGV Beast Web Application, requiring no software downloads—simply open your browser to begin. You can access basic ROS 2 functions without needing a virtual machine on your PC. Supporting high-frame rate real-time video transmission and multiple AI Computer Vision functions, the UGV Beast is an ideal platform to bring your ideas to life!
Kit Selection
An optional Pan-Tilt Module (PT Version only) enhances expansion potential. Choose from host controller options including Raspberry Pi 4B or Raspberry Pi 5, or select the Acce Version if you already own a Pi. All kits include a camera, mounting accessories, TF card, and cooling fan.
| Model | UGV Beast PT PI4B/PI5 ROS2 Kit | UGV Beast PT PI4B/PI5 ROS2 Kit Acce | UGV Beast PI4B/PI5 ROS2 Kit | UGV Beast PI4B/PI5 ROS2 Kit Acce |
|---|---|---|---|---|
| Dimension | 196.82×231.46×251.78mm | 196.82×231.46×152.25mm | ||
| Weight | 2411 ±5g | 2095 ±5g | ||
| Chassis height | 25mm | |||
| Pan-Tilt DOF | 2 | - | ||
| Pan-Tilt SERVO TORQUE | 30KG.CM | - | ||
| Pan-Tilt SERVO | ST3215 Servo | - | ||
| Host controller | RPi 4B 4GB / RPi 5 4GB | NOT included | RPi 4B 4GB / RPi 5 4GB | NOT included |
| Host System support | Debian Bookwrom | |||
| ROS2 Version | ROS2-HUMBLE-LTS | |||
| Camera FOV | 160° | |||
| Power supply | 3S UPS Module | |||
| Battery support | 18650 lithium battery x 3 (NOT included) | |||
| Demo control methods | Web application / Jupyter Lab interactive programming | |||
| Default Max speed | 0.35m/s | |||
| Number of Driving wheels | 2 | |||
| Suspension material | Stainless steel | |||
| Track width | 40mm | |||
| Minimum turning radius | 0M (In-situ rotation) | |||
Based on Raspberry Pi
The UGV Beast supports Raspberry Pi 5 and Raspberry Pi 4B, offering powerful computing performance to handle complex tasks and unlock more possibilities.
Connecting with Raspberry Pi 4B
Connecting with Raspberry Pi 5
Dual-Controller Design for Efficient Collaboration and Upgraded Performance
The host controller utilizes Raspberry Pi for AI vision and strategy planning, while the sub-controller employs ESP32 for motion control and sensor data processing, ensuring seamless and precise operations.
Raspberry Pi OS + ROS2 Docker
This configuration ensures advanced decision-making performance for the robot and excellent system compatibility. It supports all AI functions of previous AI Kit series products.
360° Flexible Omnidirectional Pan-Tilt
Equipped with a 5MP 160° wide-angle camera, the Pan-Tilt module captures every detail with precision.
Open Source for All ROS 2 Development Resources
All demos for both host and sub-controllers are open source, including the robot description file (URDF model), sensor data processing nodes for the sub-controller, kinematic control algorithms, and various remote control nodes.
Integrates Various ROS 2 Mapping Methods
The UGV Beast supports diverse mapping methods to meet the needs of different scenarios.
Gmapping 2D mapping
Cartographer 2D mapping
RTAB-Map 3D mapping
Multiple Cost-Effective Sensors
The robot incorporates several practical and cost-effective sensors for enhanced functionality.
Auto Exploration and Mapping
Using SLAM Toolbox, the UGV Beast simultaneously performs mapping and navigation in unknown environments, simplifying task execution. This allows the robot to autonomously explore and map unknown areas, ideal for unmanned applications.
Supports Natural Language Interaction
Leveraging Large Language Model (LLM) technology, users can issue commands to the robot using natural language for tasks such as moving, mapping, and navigation.
Provides Web Console Tool
Access basic ROS 2 functions directly through a web browser, eliminating the need for a PC virtual machine. This cross-platform operation works on Android or iOS tablets. Users can simply open a browser to control the robot for movement, mapping, navigation, and more.
ROS2 Node Command Interaction
Users can send control commands to the robot via scripts for operations like movement, current location retrieval, and navigation to specific points, facilitating secondary development.
Gazebo Simulation Debugging
Gazebo robot mode and a complete functionality library are provided for simulation debugging, aiding in system verification and testing during early development stages.
Continuing the Adventure as Night Falls: Supports Tactical Extension
A high-brightness LED light ensures clear images in low-light conditions.
It comes with a 21mm wide rail and a 30KG.CM high-precision & high-torque bus servo for tactical extensions.
Note: Accessories shown in the image above are for reference only and are NOT included.
Standard Aluminum Rail
Includes 2 × 1020 European standard profile rails, allowing for the installation of additional peripherals via boat nuts to meet diverse needs and expand special operation scenarios.
Note: Only the rail, boat nuts, and M4 screws are included; other accessories must be purchased separately.
Supports Driving in Complex Terrain
The tracked mobile robot chassis features independent suspension systems, providing more stable off-road crossing capabilities.
Easy To Control Via Cross-Platform Web Application
No app installation is required. Users can connect and control the robot via mobile phones, tablets, and computers using a browser-based Web App. It supports shortcut key control like WASD and mouse input via a PC with a keyboard.
WebRTC Real-time Video Transmission
Utilizes a Flask lightweight web application based on WebRTC for ultra-low latency real-time transmission. Written in Python, it's easy to extend and works seamlessly with OpenCV.
Recognition, Tracking, and Targeting
Based on OpenCV, it achieves color recognition and automatic targeting. It supports one-key Pan-Tilt control and automatic LED lighting, allowing for expansion with more functions.
Face Detection: Automatic Picture or Video Capturing
Based on OpenCV, it performs face recognition, automatically taking photos or recording videos when a face is detected.
Intelligent Object Recognition
The UGV Beast supports recognition of many common objects using its default model.
Gesture Recognition: AI Interaction with Body Language
Combining OpenCV and MediaPipe, it enables gesture control of the Pan-Tilt and LED.
Gesture control for photo taking
LED ON/OFF and backlight control
More MediaPipe Demos for Easily Creating Complex Video Processing Tasks
MediaPipe, an open-source framework by Google, facilitates building cross-platform multimedia processing pipelines. Its pre-built components and high-performance capabilities allow the robot to respond to and process complex multimedia inputs, such as real-time video analytics.
Face Recognition
Attitude Detection
40PIN GPIO Extended Header
The robot only uses the UART interface of the Raspberry Pi GPIO for communication. The outer 40PIN header of the driver board is available for expanding more peripherals and functions.
Obtains Real-time Information Feedback
Real-time monitoring of the robot's operating status is available.
Web Page Command Line Tool: Multiple Functions for Easier Expansion
Quick to set up and easy to expand. Effortlessly customize and add new functions without modifying front-end code.
ESP-NOW Wireless Communication Between Robots
Based on the ESP-NOW communication protocol, multiple robots can communicate without requiring IP or MAC addresses, enabling multi-device collaboration with 100-microsecond low-latency communication.
Gamepad Control for Better Operation Experience
Includes a wireless gamepad for more flexible robot control. Connect the USB receiver to your PC to remotely control the robot via the Internet. An open-source demo is provided for customizing your interaction methods.
Supports 4G/5G Expansion
Supports installing a 4G/5G module for application scenarios without WiFi access.
* You may need to use Tunneling Services like Ngrok, Cpolar, or LocalTunnel to expose the robot's local network service (Flask application) to the Internet for remote control.
Realizes Remote Control Across the Internet
- Our web application demos are based on WebRTC for real-time video transmission.
- WebRTC (Web Real-Time Communications) enables web applications to establish peer-to-peer connections and stream audio/video or exchange data directly between browsers.
- We provide comprehensive Ngrok tutorials to help you quickly set up and control the robot remotely over the internet.
* We provide Ngrok usage tutorials only; Ngrok accounts or servers are not provided. You can follow our tutorial to set up your own Ngrok service or choose other tunneling services as needed.
Supports Installing Smartphone Holder
If you have a spare smartphone, you can mount it on the robot using the included holder. Use the phone to create a hotspot for the robot, achieving cost-effective remote control over the internet.
* Comes with a smartphone holder with a 1/4″ screw in the package.
Cross-Platform Interactive Tutorial: Develop While You Learn
Access Jupyter Lab via mobile phones and tablets to read tutorials and edit code directly on the web page, simplifying the development process.
Rich Tutorial Resources
We provide complete tutorials and demos to help users quickly get started with learning and secondary development.