RoboCup Jr: Arduino Q AI Cognitive Targets
On-device AI and autonomous robotics
RoboCup is one of the clearest expressions of the FermiLabs Learning by Doing method: a real ruleset creates hard constraints, every design choice has consequences, and each competition cycle becomes evidence for the next redesign.
For the 2025–26 Rescue Maze evolution, the team redesigned the robot around a more compact PETG chassis, complex-geometry airless wheels and a custom PCB shield intended to simplify wiring while improving motor-control integration.
The portfolio recordsArduinohardware support and logo usage for this evolution. The software stack was rewritten for the Arduino Q platform and for the February 2026 Rescue Maze rules, including recognition of cognitive targets in confined spaces.
The result is not presented as a single finished robot but as a systems-engineering programme joining mechanics, electronics, embedded software, sensing, AI, navigation and repeatable validation.
