Runway Cleaning Robot based on Deep Learning

First author — training the deep-learning model for FOD detection and system integration

Develop a cost-effective autonomous runway-cleaning robot that performs real-time detection and collection of foreign object debris (FOD) and pavement cracks, replacing labor-intensive manual removal.

A YOLOv3 one-stage detector (a deep-learning model) deployed on a Jetson TX2 is employed and integrated with localized path planning.