城市内密集交通流行程时间预测数学建模仿真
Mathematical Modeling and Simulation of Travel Time Prediction of Dense Traffic Flow in the city
宋娜娜 1葛杨 2程海涛3
作者信息
- 1. 重庆移通学院,重庆 401520
- 2. 重庆工商大学派斯学院软件工程学院,重庆 401520
- 3. 青岛理工大学,山东 青岛 266000
- 折叠
摘要
城市交通网络包含大量的道路交叉口和车辆,且受工作日和休息日多因素的影响,使得交通流量具有不确定性,增加了预测的难度.为此,提出城市密集交通流行程时间预测数学建模研究.取样城市车辆历史数据,明确数据插值点,插值处理车辆运行数据.利用深度学习数据归一化,计算平均行程时间、行程时间方差以及可靠度指标,提取城市密集交通流特征.基于卡尔曼滤波将预测问题转化成空间状态计算问题,实现密集交通流行程时间预测.通过实验证明,所建模型能够准确预测城市密集交通流行程时间,平均绝对百分误差均在 2.3%以下,能帮助驾驶人合理规划出行.
Abstract
Currently,urban traffic networks contain a large number of road intersections and vehicles.Due to the influence of multiple factors such as workdays and rest days,traffic flow is uncertain,which increases the difficulty of prediction.Therefore,this article proposed a mathematical modeling study on predicting trip time in dense urban traf-fic.Firstly,we sampled historical data of urban vehicles and then identified data interpolation points to interpolate ve-hicle operation data.Secondly,we used deep learning data normalization to calculate average travel time,travel time variance,and reliability indicators,thus extracting the features of dense urban traffic flow.Finally,based on the Kal-man filter,we transformed the prediction problem into a problem about spatial state calculation,thereby achieving the prediction of trip time in dense traffic.Experiment results prove that the model can accurately predict the travel time in dense urban traffic and help drivers plan their trips reasonably.Meanwhile,the average absolute percentage error is less than 2.3%.
关键词
城市密集交通/行程时间/时间预测/插值处理Key words
Urban density/Trip time/Time prediction/Interpolation processing引用本文复制引用
基金项目
重庆市高等教育教学改革研究一般项目(193318)
出版年
2024
NETL
NSTL
万方数据