Overview
Timeframe: Sep 2025 - Dec 2025
Institution: University of Toronto - MIE444 (Mechanical Engineering Design Project)
Team Size: 4 members (Sam Bahrami, Kelvin Cao, Clive Fellows, Mo Taban)
Role: Mechanical Design and System Integration Engineer
Built a compact rover that navigates a maze with onboard sensing and compute, keeping the chassis rigid, serviceable, and easy to reconfigure during weekly test cycles.
The Problem
The platform needed tight packaging for motors, battery, compute, and sensors while holding alignment for reliable localization. It also had to come apart often for debugging without disturbing critical geometry.
Constraints
- Size and mass caps for the maze course
- Limited shop access and semester timeline
- Rigid, repeatable mounts for LIDAR, IMU, and encoders
- Clean cable routing and quick teardown for troubleshooting
Solution
I led mechanical layout and integration:
- Modular chassis with isolated trays for drive, compute, and sensors
- Rigid 3D-printed brackets and tight tolerances to keep sensor alignment
- Wheel, tire, and ratio selection for controllable speed and low slip
- Service-friendly hardware and wiring to support rapid swaps
Testing
- Bench checks for drivetrain smoothness and encoder accuracy
- Maze runs to tune wheel materials and vibration isolation
- Iterative sensor placement tweaks to cut noise and drift
Impact
- Delivered a reliable rover platform that survived repeated maze runs
- Improved localization through rigid, well-aligned sensor mounting
- Kept iteration fast with modular assemblies and tool-light service
Key Learnings
- Mechanical precision directly improves software performance.
- Modularity is the fastest way to debug a student robot.
- Collaboration with controls teammates surfaces mechanical improvements early.