基于多维数据的高速公路交通预测算法  

Expressway Traffic Prediction Algorithm Based on Multidimensional Data

作　　者：徐云 XU Yun(CASIC Guangxin Intelligent Technology Co.,Ltd.,Hangzhou 310000,China)

机构地区：[1]航天科工广信智能技术有限公司,杭州310000

出　　处：《价值工程》2024年第14期115-117,共3页Value Engineering

摘　　要：目前高速公路交通流的算法较多,但很多算法要基于车辆的GPS信息,或者单一的卡口信息,结果往往滞后或者不准确,我们基于浙江省高速公路温州段上的传感器(卡口与雷达)数据进行梳理,形成一套数据相互补充、相互验证的交通流短时预测算法,提高了交通状态预测,已用于实战。本文首先确定了数据的来源与格式,其次介绍了数据之间的相互补全,最后从短时交通状态演化特性的剖析到深度学习的交通状态预测,最终得到了温州高速路网的交通状态预测结果。Currently,there are many algorithms for highway traffic flow,but many algorithms rely on vehicle GPS information or a single checkpoint information,resulting in often lagging or inaccurate results.Based on sensor(checkpoint and radar)data on the Wenzhou section of the Zhejiang Province highway,we have sorted out a set of traffic flow short-term prediction algorithms that complement and verify each other's data,improving traffic state prediction and have been used in practical combat.This article first determines the source and format of the data,then introduces the complementarity between the data,and finally,from analyzing the short-term traffic state evolution characteristics to deep learning traffic state prediction,the traffic state prediction results of Wenzhou expressway network are finally obtained.

关 键 词：交通状态预测 卡口流量 雷达流量 混合流量 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]