V2X e Computer Vision per la protezione degli utenti deboli della strada: la proposta di Xenia Progetti

Abstract

This work illustrates Xenia Progetti’s innovative approach to Smart Mobility,
focusing on reducing road accident risks for vulnerable users. Despite the evolution of traditional ADAS systems, limitations related to the line-of-sight of on-board sensors remain a significant critical issue.
The proposed solution is based on the convergence of two technological pillars: advanced computer vision: the use of Instance Segmentation algorithms allows for the identification of every single road actor
at a pixel level, enabling the system to predict trajectories and identify potential collision zones in real-time; V2X (Vehicleto-Everything) communication: through the C-ITS standard and the use of Road Side Units (RSUs), the infrastructure becomes an active participant in monitoring, transmitting alert messages (CAM/DENM) to vehicles and pedestrians. The study also describes the use of Eclipse SUMO for creating Machine Learning models capable of simulating intermodal traffic scenarios and validating the system's predictive capabilities. The integration of these technologies transforms the urban environment into a cooperative ecosystem, where timely data sharing and artificial intelligence collaborate to ensure safe and proactive mobility.