Urban Traffic Lane Saturation Prediction with Least Square Support Vector Regression based on Genetic Algorithm
作者单位：College of Automation Science and Technology South China University of Technology Guangzhou Institute of Standardization
会议名称：《第30届中国控制与决策会议》
会议日期：2018年
学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 082303[工学-交通运输规划与管理] 0835[工学-软件工程] 082302[工学-交通信息工程及控制] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 0823[工学-交通运输工程]
基 金：supported by YangFan Innovative&Entepreneurial Research Team Project of Guangdong Province(2016YT03G125) the Fundamental Research Funds for the Central Universities(2017BQ098) Natural Science Foundation of Guangdong Province(2017A030313385) Science and Technology planning Project of Guangdong Province(2017A040405023 2017A040403064 2017B090901040 2017B030306017 2017B090914001) Science and Technology Program of Guangzhou(201707010152)
关 键 词：Urban Traffic Lane Saturation Prediction Least Square Support Vector Regression Genetic Algorithm
摘 要：In this paper, urban traffic lane saturation prediction model is built with Least Square Support Vector Regression（LSSVR） based on Genetic Algorithm（GA）. The traffic data is collected and the relationships among different road crosses are analyzed. The traffic data is preprocessed by correcting some unusual data with data filtering method. To construct the lane saturation prediction model, Principal component analysis（PCA） is adopted to select the appropriate input parameters. LSSVR method is then used to establish the lane saturation prediction model. The model parameters are optimized by using GA method. Finally, the real measured data is used to test the prediction model and the results show that the proposed model can forecast the lane saturation level effectively.