基于移动定位数据和群智感知的城市交通态势预测方法  

A Method for Predicting Urban Traffic Situation Based onMobile Location Data and Group Intelligence Perception

作　　者：禄铠铣 杜慧宇 田润东 李银 张乘蜜 LU Kaixian;DU Huiyu;TIAN Rundong;LI Yin;ZHANG Chengmi(Yunnan Yuntong Digital Link Technology Co.,Ltd.,Kunming 650100,China)

机构地区：[1]云南云通数联科技有限公司,云南昆明650100

出　　处：《贵州大学学报(自然科学版)》2025年第1期91-97,共7页Journal of Guizhou University:Natural Sciences

基　　金：云南省科技厅《云南省数字交通重点实验室》资助项目(202205AG070008)。

摘　　要：当前对于城市交通态势预测多采用注意力机制算法,但该方法易受到交通特征指标二次离差波动量的影响,导致预测精度较低。为此,论文提出基于移动定位数据和群智感知的城市交通态势预测方法。以移动定位数据为基础,选取能够表征城市交通状况的总量特征、相对特征与动态特征,并将特征指标进行按序加权处理,以消除其二次离差波动量,结合交通态势接口数据对路段特征进行分类,并计算特征指标隶属度,由此构建交通态势预测模型,基于此,采用群智感知技术对预测模型的控制参数进行优化,从而求取交通态势值,并将计算结果与交通态势等级划分标准相比较,以明确交通态势等级,进而实现城市交通态势预测。对比试验结果表明,将所提方法应用于城市交通态势预测中,能够获取较高的预测精度。Currently,attention mechanism algorithms are commonly used for urban traffic situation prediction,but this method is susceptible to the influence of the momentum of the secondary dispersion wave of traffic characteristic indicators,resulting in low prediction accuracy.To this end,a method for predicting urban traffic situations based on mobile positioning data and swarm intelligence perception is proposed.Based on mobile positioning data,select total,relative,and dynamic features that can characterize urban traffic conditions,and weight the feature indicators in order to eliminate their secondary deviation fluctuations.Combine the traffic situation interface data to classify road segment features and calculate the membership degree of the feature indicators,thereby constructing a traffic situation prediction model.Based on this,using swarm intelligence perception technology to optimize the control parameters of the prediction model,the traffic situation values are obtained,and the calculation results are compared with the traffic situation level classification standards to clarify the traffic situation level and achieve urban traffic situation prediction.The comparative experimental results show that applying the proposed method to urban traffic situation prediction can achieve high prediction accuracy.

关 键 词：移动定位数据 群智感知技术 城市交通 交通态势预测 

分 类 号：U491.14[交通运输工程—交通运输规划与管理]