Road Design Optimisation for Autonomous Vehicles

Abstract

The rapid progress of Autonomous Vehicle (AV) technology has generated considerable interest in the potential impact on road design. The paper presents the results of research about how the integration of AVs will influence the design and efficiency of road networks. The evaluation utilized a combination of PTV VISSIM software and the Surrogate Safety Assessment Model (SSAM) to construct a network model consisting of a primary arterial road and connecting roads within Dublin City, Ireland. Several scenarios were considered, each exploring various design elements and AV market penetration levels. The study analysed key performance indicators such as throughput volumes, average speed, travel time, vehicle and junction delays, queue length, level of service (LOS), and traffic conflicts to compare the different scenarios. The results revealed that a fully autonomous vehicle fleet led to enhanced throughput volumes and improved junction performance. Surprisingly, a dedicated AV lane did not significantly enhance overall network performance. Instead, a roundabout design emerged as a promising solution, showcasing remarkable improvements in various indicators while also mitigating conflicts compared with other scenarios. Narrower lanes improved travel time and vehicle delay but raised safety concerns due to higher conflict levels. Similarly, as expected, a higher design speed demonstrated better travel speeds. The study underscores the complexity of AV integration and the importance of a balanced approach to optimize safety and efficiency.