An automated surrogate safety analysis at protected highway ramps using cross-sectional and before-after video data

This study presents a method for surrogate safety analysis to investigate the safety of limited-access highway facilities. The proposed methodology is based on automated trajectory collection and behavioural analysis from surrogate safety measures (in particular, time-to-collision). The methodology is applied to a sample of urban highway sections at on-ramps and off-ramps to study the effectiveness of a lane-change ban treatment in Montreal, Canada. To the authors' knowledge, this is the largest automated video-based surrogate safety analysis of real sites. The applicability of the methodology is explored using (i) a cross-sectional comparison and (ii) a before-after comparison. Video data is collected using the highway traffic surveillance system and a mobile video camera unit. Various methods of aggregating the data, spatially and temporally, are explored. Although the treatment does not have a statistically significant impact on the time-to-collision distributions, it is found empirically that lane changing interactions are less predominant than rear-end interactions at these highway ramps, lane changes across the protected side of the treatment (infractions) occur in great numbers regardless of the implementation of the treatment, and that the start of the treatment produces an artificial critical point in the highway stream causing increased lane-change interactions at this point.