Search

Highway work zones are most vulnerable roadway segments for congestion and traffic collisions. Hence, providing accurate and timely prediction of the severity of traffic collisions at work zones is vital to reduce the response time for emergency units (e.g., medical aid), accordingly improve traffic safety and reduce congestion. In predicting the severity of traffic collisions, previous studies used different statistical and machine learning models with accuracy as the main evaluating factor. However, the performance of these models was generally not good, especially on fatal and injury crashes. Also, looking into the prediction accuracy only is misleading. This paper aims to propose a novel deep learning-based approach with a customized f1-loss function to predict the severity of traffic crashes. Underlying this objective is to compare the results of deep learning models with machine learning model considering two performance indicators, namely precision, and recall. The data used in the analysis include a sample of traffic crashes that occurred at work zones in Louisiana from 2014 to 2018. This dataset includes valuable information (features) related to road, vehicle, and human factors affecting the occurrence and severity of those crashes. The proposed methodology is based on transforming these features/variables into images. Image transformation is conducted using a nonlinear dimensionality reduction technique t-SNE and convex hull algorithm. A CNN based deep learning algorithm with a customized loss function was used to directly optimize the model for precision and recall. The results showed improved performance in predicting the crash severity of fatal and injury crashes using the deep learning approach, which can help to improve traffic safety as well as traffic congestion at work zones and possibly other roadways segments., (Copyright © 2021 Elsevier Ltd. All rights reserved.)

Inefficient operation of traffic in work zone areas not only leads to an increase in travel time delays, queue length, and fuel consumption but also increases the number of forced merges and roadway accidents. This study evaluated the safety performance of work zones with a conventional lane merge (CLM) configuration in Louisiana. Analysis of variance (ANOVA) was used to compare the crash rates for accidents involving fatalities, injuries, and property damage only (PDO) in each of the following 4 areas: (1) advance warning area, (2) transition area, (3) work area, and (4) termination area. The analysis showed that the advance warning area had higher fatality, injury, and PDO crash rates when compared to the transition area, work area, and termination area. This finding confirmed the need to make improvements in the advance warning area where merging maneuvers take place. Therefore, a new lane merge configuration, called joint lane merge (JLM), was proposed and its safety performance was examined and compared to the conventional lane merge configuration using a microscopic simulation model (VISSIM), which was calibrated with real-world data from an existing work zone on I-55 and used to simulate a total of 25 different scenarios with different levels of demand and traffic composition. Safety performance was evaluated using 2 surrogate measures: uncomfortable decelerations and speed variance. Statistical analysis was conducted to determine whether the differences in safety performance between both configurations were significant. The safety analysis indicated that JLM outperformed CLM in most cases with low to moderate flow rates and that the percentage of trucks did not have a significant impact on the safety performance of either configuration. Though the safety analysis did not clearly indicate which lane merge configuration is safer for the overall work zone area, it was able to identify the possibly associated safety changes within the work zone area under different traffic conditions., (Copyright © 2012 Taylor & Francis Group, LLC)