Diagnosis of Encroachment-Related Work-Zone Crashes by Applying Pattern Recognition

Abstract

Work-zone safety is one of the critical goals of transportation agencies. Vehicles are required to change travel paths and lanes over a short length of a road section at work zones. Distracted drivers, unable to see advanced warning signals and pavement markings delineating the work-zone travel paths, could increase the likelihood of a crash. Recent statistics showed that fatal collisions in work zones had increased by 46% in 2019 compared with 2011. The frequency of roadway departures at work zones, the higher risk of fatalities, and little insight into encroachment types at work zones underscored the need for a thorough study. This study aimed to examine the vehicle encroachment conditions associated with work-zone locations and focused on 4 years (2016 to 2019) of crash data from the Texas Department of Transportation by applying a unique data-mining method known as cluster correspondence analysis. This method identified four clusters in both “no-injury” and “fatal and injury” crash data. Major factors contributing to vehicle encroachment were identified. Three dominating clusters were median-related crashes on two-lane divided high-volume roadways; single-vehicle overturning collisions on two-way divided roadways with unprotected median; and overturning crashes on two-lane undivided roadways in controlled traffic. The findings of this study will be useful for safety engineers to contribute to reducing encroachment-related work-zone crashes.

Bayesian Network for Motorcycle Crash Severity Analysis

“…Conditional probabilities establish a relationship between the inputs and the result. Decisions can also be incorporated into the influence diagram so that the decision maker can comprehend how each alternative influences the probability of a result (27,28). Using the Netica software, Figure 1 illustrates an influence diagram in which Y is an uncertain outcome with two possible outcomes y1 and y2 (with probabilities of 17.5% and 82.5%, respectively), X is an uncertain factor with two possible states x1 and x2 (with probabilities of 25% and 75%, respectively), and Z is a decision with two alternatives z1 and z2.…”

Highway Construction Safety Analysis Using Large Language Models

“…Incidents also account for huge costs in the construction industry [4]. As the deteriorating condition of highways is a pressing concern, with over 44% of highway systems in the United States exhibiting a poor condition, an increase in projects related to maintenance and rehabilitation is expected in the next years [5]. With these alarming figures, there remains a need for comprehensive accident analyses in the field to help mitigate safety hazards in the highway industry.…”