Applications

REVA University Uses AI Vision and RoboDK for Collaborative Robot Automated Gluing

A REVA University team built a collaborative robot gluing system using YOLOv8 vision detection and RoboDK programming, eliminating markers and costly scanners while achieving real-time trajectory adaptation and consistent adhesive dispensing.

Published · Updated · Source: RoboDK · 工业机器人 · AI translation
Illustration
According to the RoboDK blog, a research team at REVA University in Bengaluru, India, has developed a collaborative robot automated gluing system, with results presented at an IEEE conference on Digital Signal Processing. The system uses an Intel RealSense camera to capture images and a YOLOv8 model trained on custom-labeled data to detect parts, outputting the part's full 360° orientation as a 4×4 transformation matrix. Using the RoboDK Python API, the team applies this transformation to the reference frame and glue curve in RoboDK, generating a gluing path that automatically adjusts to wherever the part is placed. RoboDK post-processors convert the path into joint commands for the UR3 collaborative robot, which are streamed in real time over Ethernet, while a pneumatic syringe with a 0.3 mm nozzle dispenses the adhesive. The team says the approach does not rely on external markers or costly laser scanners and achieved robust part alignment, trajectory adaptation, and synchronized real-time gluing with minimal error rates and deviation. Next, the researchers plan to scale the system to multi-part assemblies, integrate deep learning for complex part geometries, and extend the application to sealing, welding, and coating processes. This information comes from the original source and was organized by AI.
Read the original source

Related news