Developing Truck Corridor Crash Severity Index

Abstract

According to NHTSA, more than 400,000 truck accidents occurred in 2009 and approximately 7,800 of those were fatal crashes. Compared with extensive studies conducted on freeway truck safety, the research on arterial streets is considerably disproportionate. Making the connections between truck traffic generators, arterial streets are key links in door-to-door deliveries. There is an urgent need to study truck safety on arterial streets because of the strong growth of truck traffic. Truck-related crashes are expected to be reduced through careful planning of the location, design, and operation of driveways, median openings, street connections, and street sections. Through the collection of extensive data on selected arterial corridors that are heavily used by trucks, contributing factors to truck crash frequency and severity were identified with a negative binomial model and multinomial logit model. Corridor truck miles traveled, annual average daily traffic, signal density, shoulder width, and pavement serviceability index and its standard deviation are significant factors for crash frequency prediction. The multinomial logit model identified 12 causal factors for crash severity, such as posted speed limit, lane width, number of lanes, pavement condition index, and undivided roadway portion. Subsequently, a crash severity index for truck arterial corridors was developed. The findings from the study not only will benefit state and local agencies in planning, design, and management of a safer truck arterial corridor, but will also help carriers to optimize their routes from a safety perspective.

Analysis of Safety Effects of Traffic, Geometric, and Access Parameters on Truck Arterial Corridors

“…On the basis of the analysis it has become obvious that improvements in the design of commercial driveways could help to improve the safety of truckpreferred arterial corridors. AADT and pavement serviceability index and its standard deviation were no longer statistically significant variables after introduction of the access-related variables (6).…”

“…In order to undertake the investigation of truck crashes from a corridor perspective based on arterial roads, truck corridor selection was confined to principal arterials and minor arterials. Truck corridors were identified on the basis of criteria established in a previous study (6). The number of corridors was changed from 100 to 74 because the current study considered the corridors with signalized intersections.…”

“…The study began with an extensive literature review and focused on collecting data relating to access parameters. To investigate the consequences of incorporating access-related variables in a previous model, two negative binomial models were compared, and the implications of the statistically significant variables were discussed on the basis of the study context (6).…”

“…Note: Underlined variables are not significant in the model at 5% significance level. a Results from a previous study(6). b Final model includes only statistically significant variables.…”

Uncovering Deep Structure of Determinants in Large Truck Fatal Crashes

“…Some of these examples consist of the classification tree model ( 19 ), hierarchical Bayesian random intercept model ( 20 ), skewed logistic model ( 21 ), and a risk analysis model dependent on collision diagram ( 22 ). Even though most of these studies concentrated on highway crashes, Qin et al analyzed the factors contributing to the severity and frequency of truck-related crashes at freeway diverging section with the use of NB and MNL models ( 23 ). They found different genetic elements associated with median/shoulder width, deceleration areas, numbers of lanes, curvature, grade, speed limit, truck percentage, and annual average daily traffic (AADT) to be key factors in large truck-related crashes at freeway sections.…”