Traffic Safety Assessment with Integrated Communication System of Connected and Automated Vehicles at Signalized Intersections

Abstract

Connected and automated vehicles (CAVs) are expected to improve traffic safety effectively at signalized intersections. Considerable studies have been conducted to investigate the benefits of CAVs in improving traffic mobility and efficiency. However, in most previous research, vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communication have been considered separately rather than concurrently, to study the characteristics of CAVs, resulting in the potential of CAVs not being fully exploited and inconsistency with reality. In this paper, an integrated communication system of CAVs (ICSC), which incorporates V2V and V2I communication, is proposed, to assess traffic safety at signalized intersections. In this study, the intelligent driver model (IDM) is used to approximate V2V communication between a subject CAV and preceding CAVs. A reinforcement learning algorithm is adopted to model V2I communication between a CAV and a traffic light. The traffic safety effect of ICSC, V2V-only, and V2I-only scenarios is evaluated for different market penetration rates (MPRs). The results show that the ICSC scenario significantly reduces traffic conflicts and outperforms V2V-only, V2I-only, and benchmark scenarios when the MPR is equal to or higher than 50% with different surrogate safety measures (SSMs), such as time-exposed deceleration (TED) to avoid crashing, time exposed time-to-collision (TET), and use of a spacing gap (SGAP). Moreover, the mobility effect of the ICSC scenario is studied, and appears to increase average speed and reduce delay time. Finally, the results suggest that the ICSC can improve traffic safety and mobility concurrently and exploit the potentials of CAVs at signalized intersections.