Smart road safety for cyclists development & assessment of C-ITS emergency braking

Introduction: Urban intersections with restricted visibility are among the most hazardous conditions for interactions between cyclists and vehicles, as delayed mutual perception and limited maneuvering margins significantly increase the risk of severe collisions. Method: This study addresses these challenges by developing an integrated simulation framework that combines a digital model of real intersections with a micro-simulation model calibrated using cyclist and driver field-collected trajectories. The SUMo simulation framework includes Python APIs that reproduce reciprocal visual perception, heterogeneous reaction times, and realistic braking behavior, enabling the generation of realistic vehicle trajectories. A compound risk indicator based on post-encroachment time and speed differential is calculated to jointly assess the likelihood of a collision and the severity of injuries. Within this simulation environment, we introduced a proactive, cyclist-oriented Cooperative Intelligent Transportation System (C-ITS) Braking Message System (BMS) and compared collision risk with that from only human braking and with that from state-of-the-art vehicles equipped with Automatic Emergency Braking (AEB). Results: Across 20,000 simulated interactions under different visibility conditions, the BMS reduces the compound collision risk by approximately 90% compared with obstructed-visibility scenarios without Advanced Driver Assistance Systems (ADAS), achieving post-encroachment times comparable to those observed under clear-visibility conditions. Failure analysis reveals that a residual 14% of the most critical cases occur when BMS warnings are issued at short distances and at high cyclist speeds, with the cyclist arriving at the conflict point first. Conclusions and practical applications: These findings demonstrate the importance of C-ITS solutions coupled with Vehicle‑to‑Pedestrian (V2P) communication in conditions of limited visibility and provide operational thresholds for future more adaptive deployment.