Research
Papers
Peer-reviewed publications and patents in computer vision, robotics, and applied AI.
Vision method and system for coating processes and systems
WIPO (PCT)
This patent uses computer vision and projected light patterns to measure the position and shape of metal strips during coating. It enables non-contact monitoring for better process control and coating quality.
Vision-based target localization and online error correction for high-precision robotic drilling
Robotica
This research explores how computer vision can give industrial robots greater spatial awareness. A stereo-vision system tracks circular targets, measures the 6D pose of the robot and workpiece, and provides real-time feedback for motion correction. This enables more accurate tracking, robot guidance, and autonomous operation in precision manufacturing.
Automated vision-based measurement and inspection techniques for robotic drilling of aerospace composites
University of Manitoba
My PhD research explored how computer vision and AI can make industrial robots more accurate and autonomous. I developed vision-guided methods to correct positioning errors in single- and multi-robot drilling systems, along with a deep-learning system that detects damage and cracks in aerospace composites. Together, these technologies reduce manual setup, improve quality control, and move industrial manufacturing closer to autonomous production.
Calibration of multi-robot cooperative systems using deep neural networks
Journal of Intelligent & Robotic Systems
When industrial robots work together, their positioning errors can combine and reduce the accuracy of the entire system. In this research, we used neural networks to learn and correct these errors in a cooperative two-robot setup. The result shows how AI can make multi-robot manufacturing more precise, reliable, and adaptable.
Method for automated 3d part localization and adjustment of robot end-effectors
WIPO (PCT)
This patent uses 3D computer vision and point clouds to locate parts and automatically guide a robot to the correct position. It enables flexible industrial automation without CAD models or complex offline programming.
A 3D deep learning model for rapid prediction of structural dynamics of workpieces during machining
Procedia CIRP
Machining changes a workpiece’s geometry and vibration behaviour throughout production. In this research, we combined finite-element simulations with a 3D neural network to predict these changes in milliseconds. This could help machines detect vibration risks and adjust their operating parameters in real time, enabling smarter and more autonomous manufacturing.
A machine vision framework for autonomous inspection of drilled holes in CFRP panels
International Conference on Control, Automation and Robotics (ICCAR)
This research presents a machine-vision system for automatically inspecting drilled holes in carbon-fibre composite panels. It uses image processing to locate holes and measure defects such as delamination, reducing the need for slow and subjective manual inspection. The approach supports faster, more consistent quality control in aerospace and automotive manufacturing.