基于LS-SVM的快速路入口匝道预测控制  

Nonlinear Predictive Control for Urban Expressway Based on LS-SVM

作　　者：曹成涛[1] 

机构地区：[1]广东交通职业技术学院智能交通工程技术应用中心,广州510650

出　　处：《武汉理工大学学报（交通科学与工程版）》2015年第5期910-914,共5页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国家自然科学基金项目(批准号:50878088);广东省高等学校学科建设专项(批准号:2013WYXM0133);广东省高等学校优秀青年教师培养计划专项(批准号:Yq2013180)资助

摘　　要：由于城市快速路入口匝道交通流具有非线性和不确定性等特点,传统的匝道控制方法难以取得满意的控制效果.文中提出了一种基于LS-SVM的快速路入口匝道非线性预测控制方法.建立了快速路入口匝道宏观交通流模型,设计了LS-SVM快速路入口匝道控制器,结合非线性反馈理论对模型反馈校正,消除了由于时变和干扰等引起的不确定性,采用仿真实例对控制方法的有效性进行了验证.仿真结果表明,提出控制方法学习速度快,控制路段的交通密度和交通流量等状态变量均可有效的跟踪到期望状态,对改善快速路的交通流运行状况有明显效果.For the nonlinearity and uncertainty of urban expressway traffic flow ,it is difficult to obtain the satisfactory performance using general expressway ramp control methods .A new nonlinear predic‐tive method based on the LS‐SVM model is proposed for urban expressway ramp control in this paper . First ,a macroscopic traffic flow model of expressway ramp is built ;and then ,combining nonlinear feedback theory ,a controller of expressway ramp based on LS‐SVM is designed ,which eliminates the uncertainty caused by time‐varying and interference .Finally ,the effectiveness of this new method is validated by a simulation example .Simulation results show that the proposed method has a high learn‐ing speed and improve traffic flow conditions significantly ,which eliminates the effect of the external uncertain parameters on the system and control the state variables such as traffic density and traffic flow to track the desired states efficiently .

关 键 词：城市交通 预测控制 最小二乘支持向量机 快速路 匝道控制 

分 类 号：U491[交通运输工程—交通运输规划与管理]