Effects of Autonomous Vehicles on the Signalized Intersections’ Performance

Abstract

Traffic congestion is one of the main problems facing major urban cities around the world. Traffic delays at traffic signals are causing significant loss of productive time. The introduction of Autonomous Vehicles (AVs) has opened the door for possible advancements in the area of transportation engineering. One of the major challenges facing the implementation of such technology is undoubtedly the interaction between AVs and Human Driven Vehicles (HDVs). In this thesis, the impacts of AVs as well as HDVs on traffic signals were examined in a mixed system. The evaluation considers different traffic congestion levels, different distribution of the total traffic on the intersection approaches, different turning percentages on each approach, and different percentages of AVs in the traffic flow. Hence the impact of AVs will be studied at 0%, 20%, 40%, 60%, 80% and 100% market penetration rate (i.e. the percentage of AVs out of the total traffic). The total traffic on intersections was considered at values of 5000, 6000, and 7000 veh/hr. The distribution of traffic on the approaches considers equal distribution on all the approaches or the dominant traffic to be on two opposing directions. The left turn percentage on the main approach is considered as 15%, 30%, and 45%. The traffic signal optimization software Synchro is used to optimize signal timing for each traffic condition when considering 0% AVs. The well-known microsimulation package Vissim is used to simulate different traffic scenarios under different AVs penetration rates. It is concluded that AVs could reduce intersection delay proportionally with increase in market penetration rates. The maximum average delay reduction was 60%, which occurred at 7000 vph and 30% left turn at a penetration rate of 100% AVs. The magnitude of the delay improvement depends mainly on the traffic condition and the percentage of AVs used. The benefits from AVs are lowered significantly at oversaturated conditions and where left turn movement is beyond the left turn lane capacity. The combination between high left turn movement, imbalance distribution of traffic and high total traffic volume resulted in the lowest improvement rate with values of about 1% only in some cases.