All projects
RoboticsLab Projects · 2025
Vision-Guided KUKA Robotic Cell
Designed and implemented a vision-guided robotic system for pick-and-place on a moving conveyor, integrating camera calibration and coordinate frame transformations for accurate part localization.
Role
Robotics Programmer
Duration
Lab Projects · 2025
Team
Team of 2
/Problem
Develop a closed-loop vision-guided pick-and-place system on a moving conveyor, capable of detecting, classifying, and grasping parts with sub-centimeter accuracy.
/Approach
- 01Calibrated camera intrinsics and the camera-to-robot transform using checkerboard patterns and known fiducials.
- 02Implemented blob detection and color/shape classification to distinguish between valid and defective parts.
- 03Computed grasp poses in the conveyor frame and transformed them into the KUKA base frame in real time.
- 04Programmed KUKA KRL motion sequences with conveyor tracking and safe approach/retract waypoints.
/Results
- Achieved reliable picking on a moving conveyor within ±5 mm placement tolerance.
- Achieved >95% classification accuracy across part variants.
- Reduced cycle time by ~30% compared to a static-pick baseline.