With the annual increase in urban vehicle numbers, the design capacity of existing intersections in many large cities is no longer sufficient to meet the needs of modern urban traffic. Proper intersection design, traffic organization, and management are essential to improving road capacity and ensuring traffic safety. During peak hours, the oversaturation of traffic at key intersections in urban networks, particularly traffic conflicts between left-turning vehicles and other crossing flows, is a significant factor that can lead to network paralysis. Additionally, remote monitoring with cameras still has limitations. Among various non-traditional intersection configurations, the continuous flow intersection (CFI), such as the displaced left-turn design, has attracted considerable attention. In this study, a deep learning approach was utilized, employing the YOLOv5 object detection algorithm to detect vehicles in video footage. After detection, the Deep SORT algorithm, also based on deep learning, was used to track the vehicles, allowing for the extraction of vehicle trajectory data. A specific intersection in Xi'an was chosen as the study site, where detection trials were conducted for different types of vehicles, including private cars, large trucks, and buses. The vehicle detection accuracy reached 98.4%, with a tracking loss rate below 1% and a misidentification rate of approximately 5% in traffic volume statistics, achieving the expected outcomes. The intersection’s current conditions, including geographic location, lane traffic volume, signal timing, saturation, service level, queue length, and delays, were analyzed to evaluate its suitability for implementing a displaced left-turn design. Based on the conditions for displaced left-turn implementation and signal timing schemes, a design was proposed for the intersection, with the optimal length of the left-turn waiting area determined through enumeration. Using VISSIM simulation software, a comparative analysis was conducted between the existing intersection setup and the redesigned displaced left-turn configuration. The simulation results indicate that the displaced left-turn intersection can effectively improve service levels, reduce delays, and shorten queue lengths, demonstrating the effectiveness of the proposed displaced left-turn solution.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Research on displaced left-turn intersection system based on neural network and VISSIM simulation


    Contributors:
    Feng, Zhengang (editor) / Mikusova, Miroslava (editor) / LI, Wei (author) / Xue, Panpan (author) / Cai, Yichen (author)

    Conference:

    International Conference on Smart Transportation and City Engineering (STCE 2024) ; 2024 ; Chongqing, China


    Published in:

    Proc. SPIE ; 13575


    Publication date :

    2025-04-28





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    VISSIM-Based Simulation of the Left-Turn Waiting Zone at Signalized Intersection

    Chen, Hui / Zhang, Ning / Qian, Zhendong | IEEE | 2008


    The Applicability of Displaced Left-Turn Intersection

    Pan, Binghong / Ying, Jinfeng | ASCE | 2020



    Performance Analysis of Parallel Flow Intersection and Displaced Left-Turn Intersection Designs

    Dhatrak, Amit / Edara, Praveen / Bared, Joe G. | Transportation Research Record | 2010


    Optimization of signalized intersection based on VISSIM-SSAM simulation

    Yang, Jie / Guo, Tangyi / Chen, Yaxin | SPIE | 2022