Lumen Grid: Competitive Multi-Robot Parking Game

Published in EECS 373 Intro. to Embedded System Design – University of Michigan, 2025

Poster

Group Members: Haobo Fang, Guanyu Xu, Xiang Jiang, Varun Agrawal.

Project Status: Completed. For more implementation details, please refer to our github repo.

Abstract: Lumen Grid is a competitive multi-robot parking game that showcases end-to-end embedded integration and real-time multi-agent coordination. Four Zumo robots race on a 4×4 ft LED-marked field where parking spots are randomly generated each round and indicated by dynamic LED patterns. A PixyCam overhead vision pipeline localizes each robot using unique color tags and updates spot occupancy in real time. Players control robots via a handheld STM32-based gesture controller that maps IMU roll/pitch to speed and steering, includes an FSR-triggered emergency stop, and provides haptic feedback alongside an OLED status display. On-board STM32 control uses yaw sensing and PID-based steering to execute smooth parking maneuvers, while Bluetooth links deliver commands and velocity feedback across subsystems.