English | 中文
ROSpider is our ROS2-based hexapod robot platform built for intelligent locomotion, AI vision, SLAM, autonomous navigation, voice interaction, and multi-robot experimentation on Jetson.
ROSpider is our open-source hexapod robot platform built on ROS2 and designed for developers, educators, makers, and robotics learners who want more than a basic walking demo. It brings together motion control, perception, mapping, navigation, voice interaction, and application-level ROS workflows in one integrated workspace, so you can move from hardware bringup to advanced autonomous behaviors with far less setup effort.
We built ROSpider to solve a common problem in legged robotics: many projects can demonstrate motion, and many others can demonstrate perception, but very few connect the full chain from low-level servo control to practical ROS applications on real hardware. ROSpider closes that gap by providing a ready-to-extend software stack for a real six-legged robot platform.
Whether you are teaching robotics, prototyping interactive AI applications, or exploring how locomotion and perception work together on Jetson-based hardware, ROSpider provides a complete and approachable foundation.
ROSpider is built around a six-legged robot architecture with 18 bus servos for the legs and an additional camera pan joint, giving the platform both expressive body motion and active visual tracking capability.
Full-body motion control: The driver packages provide servo control, kinematics, built-in poses, action-set execution, gait control, odometry publishing, and controller interfaces for the hexapod chassis.
Rich onboard hardware integration: The bringup and peripherals packages launch robot control, joystick and keyboard control, RGB lighting, IMU, camera, LiDAR, and other onboard devices together, providing a practical starting point for real robot deployment instead of an isolated algorithm demo.
Jetson-oriented ROS2 design: The workspace is organized around ROS2 Humble on Jetson-class devices, with launch files and environment scripts prepared for the official ROSpider system image.
ROSpider is designed as a complete ROS2 application platform rather than a single-purpose repository.
AI vision applications: app, example, large_models, and large_models_examples include object tracking, line following, hand gesture interaction, intelligent kicking, color and AprilTag demos, MediaPipe examples, YOLO-based examples, and large-model application demos.
Self-balancing and reactive behaviors: The application layer includes IMU-driven self-balancing, LiDAR behaviors, line following, object tracking, and other perception-to-motion examples that connect sensing directly to robot movement.
SLAM and navigation: slam provides ROS2 mapping workflows such as SLAM Toolbox and RTAB-Map related launch files. navigation provides localization, Nav2-based navigation, map loading, RViz launch files, and RTAB-Map navigation support.
Offline voice interaction: The xf_mic_asr_offline and xf_mic_asr_offline_msgs packages enable offline speech interaction and voice-controlled robot behaviors.
Competition and multi-scene examples: competition and example provide scenario-oriented demos such as crossing bridges, narrow-slit traversal, pick and place, intelligent transport, navigation transport, and other classroom or contest-style tasks.
ROSpider is not just a product platform. It is also a development and teaching platform.
Comprehensive examples: example includes a wide range of ROS2 scripts and launch files covering body control, gait control, OpenCV projects, MediaPipe demos, navigation transport, object classification, color sorting, and other creative interaction examples.
Modular ROS2 package layout: Core functions are split into bringup, driver, peripherals, interfaces, navigation, SLAM, applications, competition examples, large-model examples, simulations, and voice interaction packages, making it easier to understand, customize, and extend the system.
Practical engineering workflow: The included launch files and official image environment are designed around a working ROS2 robot workflow, helping you start from a ready-to-run robot platform instead of rebuilding infrastructure from scratch.
- Official Website: https://www.hiwonder.com/
- Product Page: https://www.hiwonder.com/products/rospider
- Official Documentation: https://docs.hiwonder.com/projects/ROSpider/en/jetson-orin-nano-version/
- Technical Support: support@hiwonder.com
- Ubuntu 22.04
- ROS2 Humble
- Python 3
- Jetson-based controller board
- ROSpider hardware with LiDAR, RGB/depth camera, IMU, and serial bus servos
- Clone the repository as a ROS2 workspace:
git clone https://github.com/hiwonder/ROSPider.git ~/ros2_ws
cd ~/ros2_ws- Install ROS dependencies:
rosdep install --from-paths src --ignore-src -r -y- Build the workspace:
colcon build --event-handlers console_direct+ --cmake-args -DCMAKE_BUILD_TYPE=Release --symlink-install
source install/local_setup.bash- If you are using the official system image, load the preset ROSpider environment:
source ~/.robotrcThis script loads ROS2 Humble, the ROSpider workspace, and the device environment used by the official image.
Bring up the base robot stack:
ros2 launch bringup bringup.launch.pyBring up the full app stack:
ros2 launch app start_app.launch.pyStart SLAM:
ros2 launch slam slam.launch.py slam_method:=slam_toolboxStart map-based navigation:
ros2 launch navigation navigation.launch.py map:=map_01Start sample applications:
ros2 launch app object_tracking_node.launch.py
ros2 launch app hand_gesture.launch.py
ros2 launch app self_balancing_node.launch.pyStart peripheral visualization:
ros2 launch peripherals lidar_view.launch.py
ros2 launch peripherals depth_camera.launch.pyROSpider/
└── src/
├── app/ # Object tracking, line following, self-balancing, gestures
├── bringup/ # Base robot bringup and startup checks
├── competition/ # Competition and scenario-oriented robot tasks
├── driver/ # Controller, kinematics, servo, SDK, and hardware drivers
├── example/ # Body control, OpenCV, MediaPipe, transport, and tutorial demos
├── interfaces/ # ROS2 services and interfaces used by applications
├── large_models/ # Large-model runtime assets and related code
├── large_models_examples/ # Large-model application examples
├── navigation/ # Nav2, localization, map loading, and RTAB-Map navigation
├── peripherals/ # Camera, LiDAR, IMU, joystick, keyboard, and sensor support
├── simulations/ # Simulation-related files
├── slam/ # SLAM Toolbox, RTAB-Map, RViz, and mapping launch files
├── xf_mic_asr_offline/ # Offline speech recognition and voice control
└── xf_mic_asr_offline_msgs/# Messages for offline speech recognition
- GitHub Issues: Report bugs and request features
- Email Support: support@hiwonder.com
- Documentation: Comprehensive guides and tutorials
This project is open-source and available for educational and research purposes.
Hiwonder - Empowering Innovation in Robotics Education