New Approach to Intelligent Pedestrian Detection and Signaling on Crosswalks

Abstract

Traffic signaling systems play a crucial role in improving driver attention and reducing road speed. Nevertheless, most available solutions face challenges such as limited commercial availability, high infrastructure costs, lack of intelligence, and incomplete coverage for all road users. To address these obstacles and bolster road safety, this manuscript introduces an innovative intelligent crosswalk featuring speed bumps with integrated light signaling, facilitating precise pedestrian detection through artificial intelligence. The design methodology incorporates resins, aggregates, and reinforcing fibers, cold-injected into an aluminum mold. Notably, the system operates autonomously on solar power, ensuring sustainability and robust protection against environmental elements. To validate the crosswalk, quantitative indicators of road safety improvement are compared against a conventional crosswalk and a prior system based on fuzzy logic. A comprehensive ROC analysis of the implemented machine learning techniques revealed an accuracy rate of 99.11% in pedestrian detection, representing a substantial leap forward in road safety. In addition, a study assessing the system’s impact on user behavior found a 46.5% improvement in pedestrian trajectory, along with speed reductions observed for both pedestrians (10.24%) and drivers (32.83% during the day, and 70.6% at night). The study was further completed with an analysis of the opinion of users who perceived a significant improvement in safety and compliance with regulations with the intelligent crosswalk, highlighting the potential of the system to significantly contribute to the enhancement of road safety.