Marc-Antoine Beaudoin, P.Eng.

PhD student
IA Lab, McGill University
Centre for Intelligent Machines

McConnell Engineering Building, Room 503
3480 University, Montréal, QC H3A 2K6
ma.beaudoin [at] mail.mcgill.ca

I research the data-driven control of dynamical systems, with a particular focus on electrical vehicles. I work at the intersection of engineering product development, systems & control, and machine learning.

Machine learning is a powerful tool that can be used to enhance engineering products. It enables the automation of complex tasks that were once out of reach. It also increases the performance of already automated engineering systems by leveraging the capacity to learn from data collection. However, it can be challenging to predict the behavior of machine learning enhanced dynamical systems. My research thus focuses on the predictability aspect of machine learning control.

Publications

Google Scholar

  • Fundamental limitations to no-jerk gearshifts of multi-speed transmission architectures in electric vehicles
    Marc-Antoine Beaudoin, Benoit Boulet
    Mechanism and Machine Theory 160, 104290, 2021. (publisher) (arXiv)

  • Analytical lump model for the nonlinear dynamic response of bolted flanges in aero-engine casings
    Marc-Antoine Beaudoin, Kamran Behdinan
    Mechanical Systems and Signal Processing 115, 14-28, 2019. (publisher)

  • Nonlinear Dynamic Analysis of Bolted Flanges in Aero-Engine Casings
    Marc-Antoine Beaudoin
    MASc Thesis, University of Toronto, 2018. (university)

Education

  • PhD, McGill University, Electrical and Computer Engineering (2019-present).
    Thesis supervisor: Prof. Benoit Boulet.
    My thesis focuses on machine learning control of dynamical systems — electrical vehicles, more particularly.

  • MASc, University of Toronto, Mechanical Engineering (2018).
    Thesis supervisor: Prof. Kamran Behdinan.
    My thesis was the development of a new model for the nonlinear dynamical analysis of aero-engines. I showed that a simple lump model based on first principles captures the structural nonlinearities required for accurate predictions, and computes fast. The model is therefore well suited for an early product development phase.

  • B.ing, Université de Sherbrooke, Mechanical Engineering (2015).
    Notable project: in Project Beyond we built an ultra-efficient electric vehicle. We placed first at the Shell Eco Marathon Americas 2016; we also took home the Best Design Award.

Industry

  • Nordresa [Now part of Dana Inc.] / Powertrain engineer & Project manager (2018-19)
    Developed an electric delivery truck from scratch

  • Bombardier Transportation/ Structural engineering internship (2013 & 2014)
    Designed and analyzed parts for passenger trains

  • Heroux-Devtek/ Design engineering internship (2012)
    Designed parts for aircraft landing gears