The Mini Autonomous Vehicle Sandbox Robot is an advanced ROS Deep Learning Autonomous Driving Training Platform, engineered for comprehensive education and research in robotics and AI. Featuring a unique hybrid Mecanum and Ackermann drive system, it offers unparalleled versatility for simulating various automotive steering models and advanced motion control. Powered by an NVIDIA Jetson Nano controller (with optional Raspberry Pi 5 or Jetson Xavier NX / Orin NX upgrades), this platform is pre-installed with Ubuntu 18.04/20.04 and ROS Melodic/Noetic, providing a robust environment for developing and testing autonomous systems.
Equipped with a 2D LiDAR for SLAM and path planning, a depth camera for visual detection and lane tracking, and an IMU, the robot delivers rich sensor data for real-world AI applications. It supports the ROS Navigation Stack, GMapping, Cartographer, and AMCL for real-time SLAM and autonomous navigation. AI functions include deep learning object recognition, lane detection, traffic-sign recognition, and neural network-based path planning, all integrated with TensorFlow, PyTorch, and OpenCV AI modules.
Designed for scalability and ease of use, this compact sandbox robot is ideal for university AI curricula, deep-learning based autonomous driving simulations, SLAM and vision algorithm development, and ROS path planning research. Its aluminum alloy frame ensures durability, while comprehensive connectivity options (USB, I2C, UART, Ethernet, Wi-Fi, Bluetooth, optional 4G) and a 7-inch touch screen enhance its functionality and user experience. Programmed with Python/C++ for ROS and supporting AI training frameworks like TensorFlow Lite and PyTorch Jetson SDK, it is a perfect tool for both beginners and advanced researchers.
Key Features
- Hybrid Mecanum + Ackermann Drive Mechanism: Accurately simulates different automotive steering models for learning vehicle control and trajectory tracking.
- AI Vision and Deep Learning Integration: Jetson Nano enables on-board training for object and lane recognition using real-time image streams.
- ROS SLAM and Navigation Pre-Installed: Provides ready-to-use GMapping and Cartographer SLAM modules for mapping and path optimization tasks.
- Sandbox-Ready Teaching Platform: Compact autonomous vehicle base ideal for indoor navigation, AI education, and autonomous driving demonstrations.
- Scalable AI Framework Support: Fully compatible with TensorFlow, PyTorch, and OpenCV for custom AI module deployment and experimentation.
- Educational and Research Applications: Designed for AI autonomous driving curriculums, graduate-level robotics labs, and ROS integration research.
Technical Specifications
| Feature | Detail |
|---|---|
| Product Name | ROS Deep Learning Autonomous Driving Training Platform |
| Model | Mini Autonomous Vehicle Sandbox Robot |
| Main Controller | NVIDIA Jetson Nano / Raspberry Pi 5 (optional Jetson Xavier NX / Orin NX) |
| Operating System | Ubuntu 18.04 / 20.04 | ROS Melodic / Noetic | Pre-installed navigation and perception packages |
| Drive Architecture | Hybrid Mecanum & Ackermann structure for multi-mode motion simulation |
| Control System | DC encoder motors with PID feedback via ROS topics for velocity and steering control |
| Sensors | 2D LiDAR (for SLAM and path planning) + Depth Camera (for visual detection and lane tracking) + IMU |
| Navigation System | ROS Navigation Stack / GMapping / Cartographer / AMCL for real-time SLAM and autonomous navigation |
| AI Functions | Deep learning object recognition, lane detection, traffic-sign recognition, and path planning via neural networks |
| Software Stack | TensorFlow / PyTorch / OpenCV AI modules integrated with ROS node communication |
| Connectivity | USB / I2C / UART / Ethernet / Wi-Fi / Bluetooth (optional 4G network module) |
| Display & Control | 7-inch touch screen (optional keyboard / manual control mode) |
| Power System | 12 V Li-ion battery pack with smart power management module |
| Chassis Material | Aluminum alloy frame with precision servo mounts for camera and sensor modules |
| Dimensions | Approx. 450 × 350 × 250 mm |
| Max Speed | ≈ 1.5 m/s |
| Payload | ≈ 10 kg |
| Programming | Python / C++ for ROS | AI training frameworks: TensorFlow Lite, PyTorch Jetson SDK |
| Applications | University and AI curriculum training, Deep-learning based autonomous driving simulation, SLAM and vision algorithms development, ROS path planning and sensor fusion testing |
| Certifications | CE / RoHS / Educational Use Compliant |