AI-Driven Robotic System for Predictive Maintenance: Urban Road Defect Detection in Smart Cities

Abstract

This paper presents the development and evaluation of an integrated AI & robotic solution for predictive maintenance in smart cities, focusing on urban road defect detection. The proposed system integrates a mobile robot equipped with a high-resolution camera and a GPS module to capture video footage and geolocation data of road surfaces. The collected data is processed using cloud-based AI models, with YOLOv9 and Roboflow 3.0 Object Detection (Fast) identified as the most effective for analyzing footage to detect cracks and potholes. Experimental results validate the system's ability to accurately identify defects and generate timely reports for maintenance teams. While the approach proved effective in reducing maintenance costs, improving road safety, and extending infrastructure lifespan, limitations were identified, including GPS inaccuracies within a few meters and the challenges posed by an insufficient dataset. These findings emphasize the potential of robotic systems for enhancing urban infrastructure management while highlighting areas for future improvement.