Overview
A vision-guided, 6-DoF assistive feeding manipulator designed to aid individuals with upper-limb impairments. The system employs a hybrid 3D localization technique using sensor fusion combining monocular vision with laser sensing, to accurately identify and retrieve food items.
Key Contributions
- ●Designed the custom counterweight-based actuation system.
- ●Developed hybrid inverse kinematics (closed-form + numerical) for smooth trajectory planning.
- ●Implemented the vision pipeline for face and mouth tracking using MediaPipe.
- ●Lead author of the associated paper published at IEEE ICRAI 2024.
Technologies
PythonMediaPipeInverse KinematicsRaspberry PiLaser SensingFDM 3D Printing
Team
- Usama Jahangir (Team Lead)
- Fahad Aamir
- Wajid Ali
- Dr. Mohsin Islam Tiwana (Supervisor)
- Dr. Hamid Jabbar (Co-Supervisor)