文章编号： 1672-5328（2023）04-0099-10

汤若天1, 2，邱红桐1，卢健1，封春房1，郝明阳3
(1. 公安部交通管理科学研究所，江苏无锡214151；2. 无锡华通智能交通技术开发有限公司，江苏无锡 214125；3.北京工业大学交通工程北京市重点实验室，北京100124)

摘要： 在利用人均延误进行信号控制交叉口优化时，研究者为了简化模型通常使用不同车型的车辆 平均载客量和车均延误来估计人均延误。为揭示这一简化方法忽视的人均延误的影响因素，通过对 传统Webster 模型进行改进，从理论上证明了人均延误受到不同车辆载客量和车辆到达顺序的显著 影响。进而提出一种融合改进的Webster 模型和高斯混合模型的人均延误估计方法。利用VISSIM仿 真软件在不同交通量、车辆到达顺序、车辆类型比例和车辆载客量分布组合的场景下测试所提出的 估计方法。结果表明，信号控制交叉口人均延误受不同载客量车辆的到达顺序、车辆类型比例的影 响，证明了所提出的估计方法通过反映这些影响因素能够对处于非饱和交通流条件下的人均延误及 其极值进行更加准确的估计。

关键词： 交通控制；人均延误；影响因素分析；改进Webster模型；VISSIM仿真；信号控制交叉口

中图分类号： U491.1+12

文献标识码：A

Influence Factors of Per Capita Delay at Signal-Controlled Intersections

TANG Ruotian1, 2, QIU Hongtong1, LU Jian1, FENG Chunfang1, HAO Mingyang3
(1. Traffic Management Research Institute of the Ministry of Public Security, Wuxi Jiangsu 214151, China; 2. Wuxi Huatong Intelligent Transportation Technology Development Co. Ltd., Wuxi Jiangsu 214125, China; 3. Beijing Key Laboratory of Traffic Engineering, Beijing University of Technology, Beijing 100124, China)

Abstract： When optimizing signal-controlled intersections based on per capita delay, researchers often simplify models to estimate per capita delay by using average passenger capacity and average delay per vehicle for different vehicle types. This simplified method ignores certain influential factors of per capita delay. Based on the improved traditional Webster model, this paper presents theoretical proof to demonstrate that per capita delay is significantly influenced by different passenger capacities and arrival sequences of vehicles. An estimation method for per capita delay that combines the improved Webster model with Gaussian mixture model is proposed. The proposed estimation method was tested using VISSIM simulation software under various scenarios, including different traffic volume distributions, arrival sequences of vehicles, proportions of vehicle types, and vehicle passenger capacity distributions. The results show that per capita delay at signal-controlled intersections is affected by the arrival sequence of vehicles with varying passenger capacities and the vehicle type proportions. With these influential factors reflected, the proposed estimation method provides more accurate estimates of per capita delay and the extreme values under non-saturated traffic flow conditions.

Keywords： traffic control; per capita delay; analysis of influence factors; improved Webster model; VISSIM simulation; signal-controlled intersections