基于SA-RF的公路隧道交通流数据修复模型研究  

Research on Highway Tunnel Traffic Flow DataRestoration Model Based on SA-RF

作　　者：付立家[1] 陈丽阳 尚康 FU Lijia;CHEN Liyang;SHANG Kang(China Merchants Chongqing Communications Technology Research&Design Institute Co.,Ltd.,Chongqing 400067;College of Transportation,Chongqing Jiaotong University,Chongqing 400074)

机构地区：[1]招商局重庆交通科研设计院有限公司,重庆400067 [2]重庆交通大学交通运输学院,重庆400074

出　　处：《公路交通技术》2023年第6期137-144,182,共9页Technology of Highway and Transport

基　　金：交通运输行业重点科技项目(2021-MS4-102)。

摘　　要：为保障公路隧道交通流数据完整性,同时为公路隧道运营决策提供稳定数据支撑,采用SA-RF模型预测修补公路隧道交通流数据缺失数据,根据交通流数据缺失模式,分别建立单变量和多变量修复模式,并将两者相结合构建了SA-RF综合修复模型,以预测修补公路隧道交通流数据,同时对比分析了不同修复模型在3%、5%、10%和15%缺失率下的修复效果。实例证明,SA-RF综合修复模型在不同缺失率下均可高精度修复交通流缺失数据,且修复精度均高于同等缺失率的RF、LSTM、均值插补修复方法。To ensure the integrity of traffic flow data in highway tunnels and provide stable data support for highway tunnel operation decisions,the SA-RF model is used to predict and repair missing traffic flow data in highway tunnels,univariate and multivariate repair models are established based on the missing traffic flow data patterns.and a SA-RF comprehensive repair model is constructed to predict and repair traffic flow data in highway tunnels based on the combination of these two models.The repair effects of different repair models under 3%,5%,10%,and 15%defect rates are compared and analyzed.It is proved that the SA-RF comprehensive repair model can accurately repair traffic flow missing data under different missing rates,and its repair accuracy is higher than that of RF,LSTM,and mean interpolation repair methods with the same missing rate.

关 键 词：公路隧道 交通流缺补 综合修复模型 SA-RF LSTM 均值插补 

分 类 号：U491.112[交通运输工程—交通运输规划与管理]