Relève étoile Louis-Berlinguet 
September 2019



Pierre-Yves Lajoie

MSc student in computer engineering
Polytechnique Montréal


Award-winning publicationModeling Perceptual Aliasing in SLAM via Discrete-Continuous Graphical Models

Published in: PIEEE Robotics and Automation Letters
 

Abstract

Like humans, robots need to know where they are in order to find their way to their destination. And, like humans, robots can make mistakes, get lost and fail at their mission. These types of errors are caused by a phenomenon known as perceptual aliasing, in which a robot perceives two different areas as visually similar. The innovative approach developed by Pierre-Yves Lajoie makes it possible to model and counteract the effects of perceptual aliasing as part of the localization process to avoid errors and enable the robots to successfully complete their mission.