Exploring the safety level of a signalized roundabout with crossing BRT: an observational pilot, in Israel

: Bus rapid transit (BRT) systems promote public transport use in big cities, but their implementation requires substantial changes in urban infrastructure, with impacts on road user behaviours and safety. In Israel, recently, a BRT system called “Matronit” was introduced in the Haifa metropolitan area. The extension of the BRT network included a new setting - a signalized multi-lane roundabout with a bi-directional BRT running through its center and traffic lights for the circular traffic when crossing the BRT route. Such a setting was not familiar to Israeli drivers and thus the initial operation of the rounda-bout was accompanied by an observational pilot. The study focused on safety-related behaviours of road users, e.g. drivers’ compliance to the red-light; vehicle behaviours in the roundabout entrance areas; giving right-of-way to pedestrians. A before-after comparison was not possible, thus, the study aimed to characterize the safety level of the new setting by estimating the rates of risky behaviours observed and comparing them with indicators reported in the literature for similar traffic settings. The data were extracted from video-records of the traffic control center, in four areas of the roundabout, two near the traffic lights and at two entrance areas. The results showed that red-light violations were committed by 14% of the vehicles in one traffic-light area, with no cases in the other. On average, one-two red-light violations can be expected per hour, the frequency comparable with signalized inter-sections. Furthermore, 86% of the cases occurred during the first two seconds of the red-light, with no “near-collisions” between the vehicle and the BRT. In both entrance areas, when traffic was present inside the roundabout, 10%-20% of entering vehicles did not slow down, and this behaviour was more frequent under the red-lights. However, in most cases, the distance between the interacting vehicles was sufficient, while conflicts (with braking) were rare, in 1% of cases. Pedestrian appearance on the crosswalks was rare, but many of them (38%) were ignored by entering vehicles, while such behaviour is common for multi-lane roundabouts. In summary, the pilot demonstrated that the safety level of the roundabout was comparable to other settings, but risky driver behaviours were present both near the traffic lights and in the entrance areas. The findings seem to be site-specific and further research is needed to better fit infrastructure design solutions to various traffic volumes at roundabouts with crossing BRT.