ABSTRACT

The focus of this study is to examine the service quality of unsignalized intersections based on the perception of automobile drivers. This was achieved by developing Automobile Level of Service (ALOS) models using two computational intelligence methods – Functional Linked Artificial Neural Network (FLANN) and Differential Evolution (DE). The required data were collected at 47 unsignalized intersections in India with widely varying driving environments. Traffic simulation models were developed to estimate parameters that could not be directly measured in the field. DE model exhibited higher prediction efficiencies with coefficient of determination (R²) of 0.94 and 0.93 for training and testing datasets, respectively. Applying sensitivity analysis, significant parameters affecting ALOS were arranged in descending rank of their relative influence. Pavement condition is the most significant parameter influencing ALOS of unsignalized intersections. The proposed model will help the transportation administrators in prioritizing the key factors for investment in infrastructural development.

摘要

本研究的重点是基于汽车驾驶员的感知检查无信号交叉口的服务质量。这是通过使用两种计算智能方法——功能链接人工神经网络 (FLANN) 和差分进化 (DE) 开发汽车服务水平 (ALOS) 模型来实现的。所需数据是在印度 47 个无信号交叉口收集的，这些交叉口具有广泛不同的驾驶环境。开发了交通模拟模型来估计无法在现场直接测量的参数。DE 模型表现出更高的预测效率和确定系数 ( R ²) 的 0.94 和 0.93 分别用于训练和测试数据集。应用敏感性分析，影响ALOS的重要参数按其相对影响的降序排列。路面条件是影响无信号交叉口 ALOS 的最重要参数。所提出的模型将帮助交通管理人员优先考虑投资基础设施发展的关键因素。