Project Marvin is an advanced robotic system built using ROS 2 Humble, visualizing real-time robotic manipulation and pose detection in Rviz2 and Gazebo Fortress. It captures human pose landmarks using a camera, processes these landmarks to determine specific joint angles (e.g., shoulder and elbow angles), and maps these angles to the joints of a robotic manipulator arm, specifically designed to mimic human arm movements. The project integrates Mediapipe for pose detection and OpenCV for image processing, operating on an Ubuntu 22.04 system.

• Custom Interfaces (PoseLandmark.msg): Defines a ROS 2 message format for publishing pose landmarks detected from the video feed.
• Pose Detection Node (poseDetection.py): Captures video, detects human pose landmarks using MediaPipe, and publishes them as PoseLandmark messages.
• Manipulator Arm Control Nodes: Includes nodes like shoulderFlexion.py and shoulderAdduction.py that subscribe to PoseLandmark messages, calculate joint angles, and publish JointState messages to control the manipulator arm.
• Robot Description (marvin.urdf): Provides the URDF model for the Marvin manipulator arm, which is visualized in RViz and can be simulated in Gazebo.
• RViz and Gazebo Launch Configuration (launch/full.launch.py): Launches the system including pose detection nodes, manipulator arm control nodes, RViz for visualization, and Gazebo for simulation.

• Ubuntu 22.04 LTS
• ROS 2 Humble
• Python 3.8 or later
• OpenCV
• MediaPipe
• Matplotlib

Ensure ROS 2 and the necessary dependencies are installed on your system.

1. Clone this repository to your local machine:

   git clone https://github.com/jaehho/Marvin.git

2. Navigate to the project directory and install the required Python libraries:

   pip3 install opencv-python
   pip3 install mediapipe
   pip3 install matplotlib

3. Build the ROS 2 workspace:

   cd ros2_ws/
   colcon build

1. Source the Workspace: Source the workspace to utilize the package executables

   source install/setup.bash

2. Launch the System: Use the provided launch file to start all nodes and RViz:

   ros2 launch marvin full.launch.py

This will start the pose detection node, manipulator arm control nodes, and launch RViz and Gazebo for visualization and simulation, respectively.

Contributions to Project Marvin are welcome. Please follow the standard GitHub pull request process to submit your contributions.

This project is licensed under the MIT License - see the LICENSE file for details.