Tools

A collection of utility packages for the extender robot control framework, providing message definitions, motion capture, media publishing, and data replay capabilities.

Packages

dynamic_parameters_identification

Package: dynamic_parameters_identification

Dynamic parameters identification pipeline (trajectories generation and parameters identification). Works only for manipulators.

Features:

• Generates non synchronized sinusoids trajectories for each joints.
• Record and replay data in order to identify parameters via Weighted Least Square estimation.
• Saves the parameters in a *.npy file in order to re-use them.

apriltag_detector

Package: apriltag_detector

Real-time AprilTag detection and 3D pose estimation from camera feeds. Uses the apriltag library to detect Tag36h11 family tags and compute their position and orientation.

Features:

• AprilTag detection from image streams
• 3D pose estimation using camera intrinsics
• Configurable tag sizes and detection parameters
• Integration with USB camera for easy deployment
• Publishes detected tag poses as AprilTagPoseArray messages

Topics:

• Subscribes: /image_raw, /camera_info
• Publishes: on a topic defined by a parameter.

extender_msgs

Package: extender_msgs

ROS2 message definitions for the extender framework. Defines custom message types for teleoperation commands, joint control, and vision-based applications.

Messages:

• TeleopCommand: Teleoperation commands with velocity and mode selection
• JointPositionCommand: Joint position commands with named joint targeting
• AprilTagPose: Single AprilTag detection with pose information
• AprilTagPoseArray: Collection of detected AprilTag poses

mediapipe_mocap

Package: mediapipe_mocap

Motion capture using Google MediaPipe. Detects and tracks landmarks from RGB camera streams with real-time visualization capabilities.

Components:

• hand_landmarks_node: Detects hand landmarks from an image using mediapipe
• oak_hand_landmarks_node: Detects metric hand landmarks with an OAK camera
• HandLandmarksViewer: Shared Python visualization library used by landmark nodes

offline_media_publisher

Package: offline_media_publisher

Simulated camera stream publisher for testing and development. Publishes images or video files as camera topics without requiring physical cameras.

Components:

• video_publisher: Publishes video files as camera topics with configurable playback

replay_joint_positions

Package: replay_joint_positions

Utility for replaying recorded joint position trajectories. Useful for testing controllers and reproducing robot motions from recorded data.

Dependencies

Common Dependencies

• ROS2 (Humble or later)
• Python 3

Package-Specific Dependencies

apriltag_detector:

• OpenCV
• apriltag
• Eigen3
• cv_bridge
• sensor_msgs
• usb_cam (optional, for camera input)

mediapipe_mocap:

• MediaPipe
• OpenCV
• NumPy

offline_media_publisher:

• OpenCV

extender_msgs:

• ROS2 message generation tools

Contributing

When adding new tools:

1. Follow ROS2 package structure conventions
2. Include comprehensive documentation
3. Add appropriate tests and examples
4. Update this README with package descriptions
5. Ensure compatibility with the extender framework

License

TODO: Add license information