Using computer vision to teach robots how to see, navigate, and act autonomously.
Overview
I've been a paid researcher for LIMRL for around a year now. In that time I've grown tremendously while tackling some of the hardest problems I've ever seen.
At LIMRL we both compete internationally in humanoid and vehicular robotics, as well as collect research data for the mining industry Sudbury is home to. I've built vision systems in OpenCV, ROS2 nodes for complex behaviours; including object grabbing and caring for humanoid robots, as well as sign navigation and obstacle avoidance for autonomous vehicles.
In summer 2026 I had the great pleasure of competing on the world stage for LIMRL in the Fira Roboworld Cup on the Autonomous Car team. Countless days and nights were spent preparing the team and our robots for the week-long event. In the end, our robust vision system, and fine tuned navigation system won first place in Race, Urban Driving, and the Technical Cooperative Challenge. A paper on our strategy will be published soon.
My involvement at LIMRL extends far beyond software. During fall 2026 and into the winter I was assigned to design new hardware for next year's competition. Alongside this, I'm assisting in a months-long research project to collect information on the effects of weather and terrain on electric vehicles operating in a mining environment. The study primarily focuses on battery-life, and the viability of EVs as workhorses in the mining industry.
What I Learned
- Built real-time computer vision pipelines for autonomous navigation, object detection, camera processing, and movement decision-making
- Learned to write performant C++ and Python for situation when power is limited frame rate is critical
- Gained crucial experience with collaborative development on large projects with many contributors
- Gained experience as a researcher working alongside my peers on projects funded by real world companies
- Learned to focus efforts to achieve the best results while development time is constrained
Gallery
Robot at competition
Hardware assembly
Post competition 2026 LIMRL dinner party