Project AURA

Led a 5-person team building an autonomous cart capable of carrying 300 lb loads while following a user via reflective tape tracking—designed as a proof-of-concept for hands-free shopping carts or airport luggage transport. Managed cross-functional integration across mechanical, electrical, and software subsystems under aggressive 2-month timeline.

Defined system architecture and interfaces between subsystems (sensor data formats, power requirements, mechanical constraints for sensor mounting). Managed cascading delays from component backordering and integration challenges—descoped advanced features (path planning, high-speed operation) to prioritize core functionality (reliable following, obstacle avoidance) when timeline compressed. Debugged critical steering failure through systematic subsystem isolation: tested electrical signals, verified mechanical assembly for binding, then identified root cause in software—mechanical torque calculations were incorrect, causing stepper motor to skip steps. Solution: reduced step size to prevent gear skipping. Navigated pushback from software team on feature cuts, prioritizing safety (lower max speed) and demo reliability over advanced capabilities.

Designed and implemented power distribution across multiple voltage sources (motors, sensors, compute)—underestimated complexity of voltage regulation and noise isolation, leading to mid-project redesign. Integrated reflective tape tracking sensor and RealSense depth camera for obstacle detection, managing cable routing to avoid mechanical interference. Created electrical documentation and wiring diagrams enabling team troubleshooting and modifications.

Successfully demonstrated autonomous following and obstacle avoidance in controlled conditions. System limitations exposed gap between proof-of-concept and production readiness: tracking lost user when reflective tape moved out of view or when other reflective objects appeared in environment; obstacle avoidance occasionally over-sensitive in crowded spaces.

Integration is where theory meets reality—mechanical calculations that looked correct on paper caused motor skipping in practice. Managing multiple voltage sources proved harder than expected. Descoping isn't failure; it's prioritization under constraints. For future projects: build integration time into schedule, not just component development.