Scholarlink

<p>The study developed microsimulation (VISSIM) models and utilized the Surrogate Safety Assessment Model (SSAM) to assess traffic safety in the case of a three-lane unsignalized roundabout under mixed traffic conditions. The roundabout network was constructed in VISSIM using drone images, and traffic characteristics extracted from the aerial videos were used as input in the simulation model. Subsequently, vehicular trajectories exported from VISSIM were analyzed in SSAM to identify conflict points (setting TTC = 1.5s and PET = 1.5s). A total of 1194 conflicts were identified, of which 77.5% were rear-end conflicts, 21.18% were lane change conflicts, and 1.25% were crossing conflicts. A Generalized Linear Model (GLM) with Negative Binomial (NB) distribution and a log link function was used to predict simulated conflict frequency based on traffic characteristics. The model showed high predictive accuracy (R<sup>2</sup> = 0.89). The study yielded a simple, easy-to-interpret, and user-friendly formulation to find the total hourly conflicts utilizing traffic microsimulation.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Safety Appraisal of Multi-Lane Roundabout under Mixed Traffic Using Microsimulation

  • 期刊论文

  • 发表日期：

    2026年5月14日

查看全文
  • Abhijnan Maji,
  • Deepanshu Soni,
  • Indrajit Ghosh

摘要

The study developed microsimulation (VISSIM) models and utilized the Surrogate Safety Assessment Model (SSAM) to assess traffic safety in the case of a three-lane unsignalized roundabout under mixed traffic conditions. The roundabout network was constructed in VISSIM using drone images, and traffic characteristics extracted from the aerial videos were used as input in the simulation model. Subsequently, vehicular trajectories exported from VISSIM were analyzed in SSAM to identify conflict points (setting TTC = 1.5s and PET = 1.5s). A total of 1194 conflicts were identified, of which 77.5% were rear-end conflicts, 21.18% were lane change conflicts, and 1.25% were crossing conflicts. A Generalized Linear Model (GLM) with Negative Binomial (NB) distribution and a log link function was used to predict simulated conflict frequency based on traffic characteristics. The model showed high predictive accuracy (R2 = 0.89). The study yielded a simple, easy-to-interpret, and user-friendly formulation to find the total hourly conflicts utilizing traffic microsimulation.