VVM
say “pick up the red ball” and a robot arm finds it and picks it up. my first ROS2 project.
jun – jul 2024 · me · simulation · retired
Voice Video Manipulator. You say what you want, a camera finds it, an arm grabs it. All in Gazebo, with an Open Manipulator X.
The pipeline: speech recognition turns your words into a target (“red ball”), object detection finds it in the camera frame, some geometry turns the 2D pixel into a 3D position, and inverse kinematics moves the arm there.
This was the project where I learned ROS2 — nodes, topics, the whole way of thinking. Before this I wrote robots as one big loop. After this, everything I built was modular. It shows up in every project that came after.
made with
- ROS2
- Gazebo
- RViz
- Python
- OpenCV
- speech recognition
- inverse kinematics