A terrain-capable, vision-guided hexapod robot that carries payloads, tracks objects, and navigates autonomously — built with Raspberry Pi 5, Coral USB, STM32, and ROS2.
- 🤖 3DOF Hexapod – 18-DOF servo system (35KG torque), 5kg payload capacity
- 🧭 Autonomous Locomotion – ROS2-based gait generation, terrain-ready
- 🛠️ Modular Frame – Custom-designed in SolidWorks, 3D-printed and torque-optimized
- 🧠 AI Vision – YOLOv5n object detection accelerated with Coral USB
- 🧩 Multi-Processor System – STM32/ESP32 for real-time servo control, Pi for high-level ROS2 tasks
- 🌐 Sensor Fusion – Optional IMU + Lidar integration for navigation
- Languages: Python 3.9+, C/C++
- Middleware: ROS2 Humble, micro-ROS, UART
- AI: YOLOv5n (TFLite quantized), Coral USB Accelerator, OpenCV
- Microcontrollers: STM32 (Nucleo), ESP32
- Hardware: Raspberry Pi 5, 35KG Serial Bus Servos, IMU, USB Camera
- Mechanical Design: SolidWorks (frame, servo mounts)
- Power: 3S LiPo Battery, UBEC, XT60 harness
- Tools: VS Code, PlatformIO, Git, RViz2
SpiderBot/
├── ros2_ws/ # ROS2 workspace
│ └── src/
│ └── hexapod_control/
│ ├── launch/
│ ├── hexapod_control/
│ │ ├── gait_node.py
│ │ ├── imu_node.py
│ │ └── serial_bridge.py
├── vision/ # Coral USB AI detection
│ ├── detect.py
│ └── tflite_model/
├── firmware/ # STM32 gait control code
│ └── main.c
├── solidworks/ # CAD source files
├── utils/ # UART & helper tools
├── README.md
└── requirements.txt
- Flash STM32 with
main.c(via STM32CubeIDE or PlatformIO) - Build and source ROS2 workspace:
cd ros2_ws && colcon build
source install/setup.bash
ros2 launch hexapod_control bringup.launch.py- Run object detection:
python3 vision/detect.pyThis project is licensed under the MIT License.