Traffic Safety Analysis for Urban Highway Ramps and Lane-Change Bans Using Accident Data and Video-Based Surrogate Safety Measures.

摘要

The purpose of this study was to evaluate the traffic safety of urban highway segments near exit and entrance ramps in the Montréal metropolitan area. The city's tight urban environment has resulted in the construction of sub-standard highway ramp merging sections (e.g. short merging lengths, inadequate visibility, influence zone overlap, etc.). In order to mitigate safety problems associated with these inadequately designed features, a special lane-change ban treatment (technical designation LCGV1) was implemented several years ago at various ramps. This study used accident data and video-based surrogate safety measures to evaluate the safety effectiveness of the treatment.;The cross-sectional accident analysis controlled for factors such as lane configuration, merge length, traffic flow and speed, area of influence overlap (inter-ramp distance), lane and shoulder widths, horizontal and vertical curves, and covered the presence of the treatment across 10 years of accident data at multiple sites along Montréal's busiest highways. The time-to-collision conflict measure obtained from automated video-based vehicle trajectory extrapolation was analyzed and used to identify microscopic behaviour patterns and conflicting interactions.;The study generally concludes that, across all sites, the presence of the treatment has led to no appreciable change in accident rate and that other contributing factors have played a greater role in observed accident rate, time-to-collision distribution, and lane changes. However, the study also indicates that there was significant variation between contributing factors across all analysis sites, leading to the conclusion that adopting a general policy of treating an entire urban region is a futile exercise. In addition, it was observed that the treatment has had a slight accident migration effect. These conclusions lead to the recommendation that the treatment should be applied on a case-by-case basis only, and otherwise that the default case (no treatment) should remain in effect so as not to hinder the normal navigation and operation of highway drivers.