[home] [Personal Program] [Help]

---

Session: Session 10: Defect Prediction and Mitigation in AFP
Session starts: Thursday 16 April, 13:10
Presentation starts: 13:30
Room: Main

---

Stig McArthur (University of Strathclyde)
Catherine Yokan (University of Strathclyde)
Jorn Mehnen (University of Strathclyde)

Abstract:
Real-time adaptive trajectory control for Automated Fibre Placement (AFP) systems using industrial middleware to mitigate common placement defects during deposition is demonstrated as a proof-of-concept. By integrating the Robot Operating System (ROS) with the Interfacing Toolkit for Robotic Arms (ITRA), our system achieves 12ms real-time control of a 12-degree-of-freedom industrial robot, enabling on-the-fly corrections to the deposition trajectories executed on the robot based on laser profilometry feedback and controlled using standard computing hardware. The system utilised an automated defect detection system capable of detecting common placement-induced defects and applies in-process corrections to the pre-programmed trajectories to mitigate or remove the defects. The architecture is designed with future capabilities in mind, enabling on-the-fly trajectory generation that would dramatically reduce pre-production programming requirements. Real-time deposition trials on 2-dimensional layup paths validate the approach, demonstrating how middleware-enabled flexibility bridges the gap between programmed deposition plans and manufacturing reality. This work establishes a practical framework for intelligent, self-correcting AFP systems that respond to material deposition irregularities in real-time rather than after-the-fact.