A Closed Loop Robotic Welding System Using Point Clouds and Drawing Data

This project addresses a flexible autonomous robotic welding system. Despite the advancements in robotic systems, there are still significant challenges in developing welding robots for industrial applications. Challenges such as accurately determining the weld seam coordinates, ensuring the electrode head remains at the correct distance from the weld seam, and mitigating high environmental noise have constrained the deployment of these systems to a limited range of activities, necessitating operator intervention during setup. In this project, an accurate approximation of the weld seam is achieved by integrating data from various sources, including point clouds from stationary cameras, point clouds from eye in hand camera, and drawing data. To maintain the correct electrode-to-weld distance during welding, adjustments are made within a closed-loop control system designed to regulate the welding arc’s impedance. The proposed framework is implemented within a simulation environment and subjected to evaluation.

Faculty Supervisor:

Mehran Mehrandezh

Student:

Partner:

Yazd University

Discipline:

Engineering

Sector:

Artificial Intelligence; Manufacturing and Construction; Advanced Manufacturing

University:

University of Regina

Program: