基于神经网络的无信号控制交叉口内车速预测模型对比

doi:10.3969/j.issn.1674-0696.2021.12.06

重庆交通大学学报（自然科学版） ›› 2021, Vol. 40 ›› Issue (12): 33-41.DOI: 10.3969/j.issn.1674-0696.2021.12.06

• 交通+大数据人工智能 • 上一篇

基于神经网络的无信号控制交叉口内车速预测模型对比

马莹莹，张子豪，吴嘉彬

(华南理工大学土木与交通学院, 广东广州510641)

收稿日期:2020-09-18 修回日期:2020-11-30 发布日期:2021-12-27
作者简介:马莹莹(1983—)，女，吉林省吉林市人，副教授，博士，主要研究方向为、智能交通管理与控制。E-mail：mayingying@scut.edu.cn 通信作者：吴嘉彬(1993—)，男，海南海口人，博士研究生，主要研究方向为交通运输规划与管理。E-mail:ctwjb@mail.scut.edu.cn
基金资助:
广东省自然科学基金项目(2018A0313250)

Comparison of Prediction Models of Vehicle Speed at Urban Unsignalized Intersection Based on Neural Network

MA Yingying, ZHANG Zihao, WU Jiabin

(School of Civil Engineering and Transportation, South China University of Technology, Guangzhou 510641, Guangdong, China)

Received:2020-09-18 Revised:2020-11-30 Published:2021-12-27

摘要/Abstract

摘要： 城市道路交叉口是交通冲突和交通事故的集中发生地，尤其是无信号控制交叉口，而车速是影响交叉口安全的重要因素，为了能高效准确地预测车辆通过交叉口的速度值，建立了基于神经网络的无信号控制交叉口内车速预测模型。首先，通过录像的方式对广州市建设路-建设大道交叉口进行拍摄，并利用tracker视频识别软件对拍摄交叉口进行数据提取并分类，共收集了700份有效数据；其次，利用相关性分析，确定了影响车辆通过无信号控制交叉口内平均速度的显著因素，并把它们作为预测模型的输入；最后，分别使用BP神经网络、GA-BP神经网络、深度神经网络(DNN)、自适应模糊推理系统(ANFIS)进行测试，并利用4种评价指标对模型的预测效果进行对比分析。预测结果表明：4种模型都能很好的拟合车辆通过交叉口的平均速度与其影响因素之间复杂的关系，其中DNN预测精度为94.44%，BP神经网络预测精度为89.08%，利用遗传算法(GA)改进的BP神经网络预测精度在训练集上提升了1.55%，测试集上提升了2.17%，ANFIS的预测精度为84.42%。

关键词: 交通运输工程；无信号控制交叉口；车速预测；神经网络；自适应模糊推理系统；视频识别

Abstract: Traffic conflicts and accidents were concentrated at intersections in the urban road, especially at unsignalized intersections. Vehicle speed was an important factor affecting intersection safety. In order to efficiently and accurately predict the speed of vehicles passing through intersections, a prediction model of vehicle speed at urban unsignalized intersection based on neural network was proposed. Firstly, the intersection of Jianshe Road and Jiansheer Avenue in Guangzhou was photographed by video, and the data of the photographed intersection was extracted and classified by using Tracker video recognition software. A total of 700 valid data were collected. Secondly, the significant factors influencing the average speed of vehicles passing through unsignalized intersection were determined by using the correction analysis, and they were used as the input of prediction model. Finally, BP neural network, GA-BP neural network, deep neural network and adaptive neuro-fuzzy inference system were used to test respectively, and four evaluation indexes were used to compare and analyze the prediction effect of the model. The prediction results show that these four models are able to fit the complex relationship between the average speed of vehicles passing through the intersection and its influencing factors. Among them, the prediction accuracy of DNN, BPNN and ANFIS is 94.44%, 89.08% and 84.42% respectively, and the prediction accuracy of BP neural network improved by genetic algorithm (GA) was improved by 1.55% on the training set and 2.17% on the test set.

Key words: traffic and transportation engineering; unsignalized intersection; speed prediction; neural network; adaptive neuro-fuzzy inference system; video recognition

中图分类号:

U491.2+3

马莹莹，张子豪，吴嘉彬. 基于神经网络的无信号控制交叉口内车速预测模型对比[J]. 重庆交通大学学报（自然科学版）, 2021, 40(12): 33-41.

MA Yingying, ZHANG Zihao, WU Jiabin. Comparison of Prediction Models of Vehicle Speed at Urban Unsignalized Intersection Based on Neural Network[J]. Journal of Chongqing Jiaotong University(Natural Science), 2021, 40(12): 33-41.

参考文献

［1］孙立云.城市道路交叉口事故预测模型及算法研究［D］.北京：北京交通大学，2011.
SUN Liyun. Research on Urban Road Intersections-Accident Prediction Model and Algorithm［D］.Beijing: Beijing Jiaotong University, 2011.
［2］施晓芬.城市信号交叉口驾驶员速度控制和视觉搜索行为研究［D］.西安：长安大学,2011.
SHI Xiaofen. Study on Speed Control and Visual Search of Driving Behavior at Urban Signal Intersection［D］. Xian:Changan University,2011.
［3］ NEKOUI M, NI D, PISHRO-NIK H, et al. Development of a VII-enab-led prototype intersection collision warning system［J］. International Journal of Internet
Protocol Technology, 2009, 4(3):173-181.
［4］〖JP3〗ABULEBDEH G, CHEN H. Exploring the potential benefits of intellidrive enabled dynamic speed control in signalized networks［C］∥Transportation
Research Board.［S.l.］:［s.n.］, 2010:1-19.〖JP〗［4］ KAMAL M A S, IEEE M, TAGUCHI S, et al. Efficient driving on multilane roads under a connected vehicle
environment［J］. IEEE Transactions on Intelligent Transportation Systems, 2016, 17 (9):2541-2551.
［6］〖JP3〗KAMAL M A S, TAGUCHI S, YOSHIMURA T. Proceedings of the 18th International IEEE Annual Conference on Intelligent Transportation systems［C］.
Spain:IEEE, 2015.〖JP〗［5］李鹏凯, 杨晓光, 吴伟, 等.车路协同环境下信号交叉口车速引导建模与仿真［J］.交通与计算机,2012,30(3):136-140,156.
LI Pengkai, YANG Xiaoguang, WU Wei,et al. Modeling vehicle speed guidance at signalized intersections under IntelliDriver SM［J］. Computer and
Communications,2012, 30(3): 136-140, 156.
［6］李鹏凯, 吴伟, 杜荣华, 等.车路协同环境下多车协同车速引导建模与仿真［J］.交通信息与安全,2013,31(2):134-139，148.
LI Pengkai, WU Wei, DU Ronghua, et al. Modeling and simulation of multi-vehicle coordinated speed guidance in a vehicle-road collaborative environment［J］. Traffic
Information and Safety, 2013, 31(2):134-139，148.
［7］楼挺. 基于深度学习的道路车速预测模型研究［D］.杭州：浙江工业大学,2019.
LOU Ting.Research on Future Vehicle Speed Prediction Model on Roads Based on Deep Learning［D］. Hangzhou:Zhejiang University of Technology,2019.
［8］王栋, 邓北川, 仇建华, 等.山区高速公路直线段车速预测方法［J］.交通科学与工程,2014(3):81-86.
WANG Dong,DENG Beichuan, QIU Jianhua, et al. The method of the straight section speed prediction on the mountainous freeway［J］. Journal of Transport Science and
Engineering, 2014 (3):81-86.
［11］〖JP3〗KALAMARAS I, DROSOU A, VOTIS K, et al. A multi-objective data mining approach for road traffic prediction［C］∥Ifip International Conference on
Artificial Intelligence Applications & Innovations,2018.〖JP〗［9］张良力, 祝贺, 吴超仲, 等.基于ARMA预测模型的交叉口车辆碰撞风险评估［J］.交通运输系统工程与信
息,2015,15(5):239-245.
ZHANG Liangli, ZHU He, WU Chaozhong, et al. Risk estimation to vehicles collision at intersection based on arma prediction model［J］. Journal of Transportation
Systems Engineering and Information Technology, 2015, 15(5):239-245.
［10］祝贺. 交叉口冲突方向车辆速度时序分析与预测研究［D］. 武汉: 武汉科技大学,2016.
ZHU He.Analysis and Prediction of Vehicles on Conflicting Directions in Intersection Area［D］. Wuhan:Wuhan University of Science and Technology,2016.
［11］徐嘉俊. 基于GA-BP神经网络的居民出行行为预测［D］. 北京:北京交通大学,2019.
XU Jiajun.Prediction of Residents Travel Behavior Based on GA-BP Neural Network［D］. Beijing:Beijing Jiaotong University,2019.
［12］颜秉洋, 唐敏佳, 周长庚, 等. 基于ANFIS混合模型的短时交通流预测［J］. 计算机系统应用, 2019,28(6):247-253.
YAN Bingyang, TANG Minjia, ZHOU Changgeng, et al. Short-Term traffic flow prediction based on ANFIS hybrid model［J］. Computer Systems & Applications, 2019,28
(6):247-253.

基于神经网络的无信号控制交叉口内车速预测模型对比

Comparison of Prediction Models of Vehicle Speed at Urban Unsignalized Intersection Based on Neural Network

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 1

编辑推荐

Metrics