| 考虑多维动态特征交互的高速公路实时事故风险建模 |
| |
| 引用本文： | 袁振洲,胡嫣然,杨洋. 考虑多维动态特征交互的高速公路实时事故风险建模[J]. 交通运输系统工程与信息, 2022, 22(3): 215-223. DOI: 10.16097/j.cnki.1009-6744.2022.03.024 |
| |
| 作者姓名： | 袁振洲 胡嫣然 杨洋 |
| |
| 作者单位： | 1. 北京交通大学，交通运输学院，北京100044；2. 北京航空航天大学，a. 交通科学与工程学院，b. 车路协同与安全控制北京市重点实验室，北京 100191 |
| |
| 基金项目： | 中国博士后科学基金；北京市自然科学基金；山东省高速公路技术和安全评估省级重点实验室开放基金 |
| |
| 摘 要： | 为探究天气和道路等特征，以及交通流、天气、道路及时间等多维动态特征之间的交互作用对实时事故风险预测模型精度的影响，本文基于京哈高速公路北京段的事故数据，以及匹配的交通传感器数据、天气数据和道路特征等，构建4个数据集，分别为只包含交通流变量，包含交通流变量、天气及时间特征变量，包含交通流变量、道路及时间特征变量，包含交通流变量、天气、道路及时间特征变量。从考虑多维动态特征的交互效应出发，基于深度交叉网络，提出一种新的实时事故风险预测模型。结果显示，本文所构建的深度交叉网络模型比其他几种实时事故风险预测方法显示出更高的精度。模型的AUC值(Area Under Curve)可达0.8562，在0.2的概率阈值下，可以正确分类84.26%的非事故数据和77.55%事故数据。结论表明，本文采用的多维动态特征交互样本条件下的深度交叉网络模型能够有效地预测高速公路交通事故，可为我国高速公路安全管理部门提供理论与技术支持。
|
| 关 键 词： | 交通工程 实时事故风险识别 深度交叉网络模型 高速公路 多维特征交互 深度学习 |
| 收稿时间： | 2022-02-16 |
Modeling Towards Freeway Real-time Traffic Crash PredictionConsidering Multi-dimensional Dynamic Feature Interactions |
| |
| YUAN Zhen-zhou,HU Yan-ran,YANG Yang. Modeling Towards Freeway Real-time Traffic Crash PredictionConsidering Multi-dimensional Dynamic Feature Interactions[J]. Journal of Transportation Systems Engineering and Information Technology, 2022, 22(3): 215-223. DOI: 10.16097/j.cnki.1009-6744.2022.03.024 |
| |
| Authors: | YUAN Zhen-zhou HU Yan-ran YANG Yang |
| |
| Affiliation: | 1. School of Traffic and Transportation, Beijing Jiaotong University, Beijing 100044, China;2.a. School ofTransportation Science and Engineering;2.b. Beijing Key Laboratory for Cooperative Vehicle InfrastructureSystems and Safety Control, Beihang University, Beijing 100191, China |
| |
| Abstract: | This paper investigates the impact of weather, road features, and the dynamic mutual interactions amongtraffic flow, weather, road, and time on the accuracy of real-time crash risk prediction. The study developed fourdatasets based on the crash data, traffic sensor data, weather data, and road data collected from the Beijing section ofthe Beijing-Harbin Freeway. The datasets include (1) the simple traffic flow data; (2) the combined traffic flow,weather, and time data; (3) the combined traffic flow, road, and time data; (4) combined traffic flow, weather, road, andtime data. By considering the interactions of multi-dimensional dynamic features, this study proposes a real-time crashrisk prediction model based on the Deep & Cross Network (DCN). The results demonstrate that the DCN modelachieves higher accuracy than other methods in real-time crash risk prediction. The Area Under Curve (AUC) of themodel is 0.8562 and the proposed model is able to correctly classify 84.26% of non-crash data and 77.55% of crashdata with the probability threshold of 0.2. The DCN model used in this study can effectively predict the occurrence offreeway crashes and collisions in time, under the condition of multi-dimensional dynamic feature interactions. Theproposed method has great potential to support the freeway safety management departments of China in boththeoretical and technical aspects. |
| |
| Keywords: | traffic engineering real-time traffic crash recognition deep & cross network freeway；multidimensionalfeature interaction deep learning |
|
| 点击此处可从《交通运输系统工程与信息》浏览原始摘要信息 |
|
点击此处可从《交通运输系统工程与信息》下载全文 |
