Faithin Omaha eBook Collection

Download or read book Modeling Lane-Change Risk in Urban Expressway Off-Ramp Area Based on Naturalistic Driving Data written by Lanfang Zhang and published by . This book was released on 2020 with total page 15 pages. Available in PDF, EPUB and Kindle. Book excerpt: Off-ramp areas are considered the critical sections of urban expressways where the exiting vehicles and straight-through vehicles merge. Therefore, lane-change behaviors frequently occur at the upstream of the urban expressway off-ramp, which lead to high chance of traffic crashes. This study looks at the risk of lane-change behaviors in the multilane urban expressway off-ramp areas. First, lane-change process information of exit vehicles in urban expressway off-ramp area was extracted from the Shanghai Naturalistic Driving Study (SH-NDS) database. Second, for each lane-change movements of exit vehicles, a risk evaluation indicator (risk perception, RP) was adopted to quantify the lane-change risk. Based on the RP, the study proposed a four-rank risk classification criterion using K-means clustering to define the risk rank of each lane-change movement. Finally, a lane-change risk rank classification model was developed for traffic in the off-ramp areas of multilane expressways using four distinctive influencing factors. Four influencing factors, namely, traffic congestion level, demand lane change times, lane-change direction, and relative distance between vehicle and exit, were used to describe the traffic flow characteristics and exiting lane-change route for the modeling purpose. The risk model was developed using two support vector machine models, which were based on the partial binary tree structure and the directed acyclic graph structure, respectively. The results showed that the overall accuracy of the partial binary tree structure classifier was 65.71 % and the average AUC value was 0.9004, both of which shows a better performance of the partial binary tree structure classifier, compared with the directed acyclic graph structure classifier.