High-Precision Imitation Learning for Real-Time Robotic Control

In recent years, an increase in industrial robots in manufacturing has emerged. However, there are still possible safety issues and difficulty in specifying tasks for the robots to perform. The objective of this research project is to make a path planning system that uses demonstrations of how to perform a task to learn how to perform the task using techniques from the field of machine learning. These demonstrations will also show the robot how to move in the workspace safely and without entering collision with items in its surroundings. This system aims to be integrated into Mecademic’s Meca500, which will make the robot more user-friendly, safer and more accessible to people unfamiliar with industrial robotics.

Intern: 
Alexandre Coulombe
Faculty Supervisor: 
Hsiu-Chin Lin
Province: 
Quebec
University: 
Partner: 
Sector: 
Partner University: 
Program: