A Study for Developing On-line Traffic Control System Model for Urban Expressway

本研究旨在建立城市快速路在线交通控制系统模型。本研究由以下四个研究课题组成。1)第一个课题是开发LP （Linear Programming，线性规划）型流入控制模型，该模型尽可能多地利用实时获得的交通数据。流入模型明确考虑了交通需求和行驶速度随时间的变化。该模型还包括了未来城市高速公路网扩展的驾驶员路线选择模型。2)建立了城市快速路控制方案的“中观”交通仿真模型。该模型具有宏观模拟模型和微观模拟模型的优点。假设每辆车都按照宏观程序更新的交通状况行驶。这一假设使得模型可以明确地考虑每辆车的目的地和路线。通过实测交通状况与估算交通状况的比较，验证了模型的有效性。3)为了提高交通仿真模型和流入控制模型的实用性，有必要开发动态OD交通量估算方法。本文提出了一种基于交通计数和车辆自动识别（AVI）数据的动态OD交通量估算方法。该方法来源于交通量观测值与估计值误差的最小二乘法。4)将流入控制模型、中观交通仿真模型和动态OD交通量估算方法相结合，建立了城市快速路在线交通控制系统模型。通过对某城市高速公路网的数值算例验证了系统模型的有效性和实用性。

This study aims at developing an on-line traffic control system model for urban expressway. This study is composed of the following four research topics.1) The first topic is to develop the LP (Linear Programming) type inflow control model that uses the traffic data obtained in the real-time manner as much as possible. The inflow model explicitly considers time-dependent changes in traffic demand and travel speed. Also the model includes the driver's route choice model for the future extension of urban expressway network.2) This study also proposed a "meso" traffic simulation model by which control schemes for urban expressway can be evaluated. The model has strong points both of macroscopic and microscopic simulation models. Each vehicle is assumed to travel according to traffic condition updated through macroscopic procedure. This assumption leads to the characteristic that the destination and route of each vehicle can be explicitly considered in the model. The validity of model was confirmed through the comparison between observed and estimated traffic conditions.3) In order to enhance the practicability of both the traffic simulation model and the inflow control model, it is necessary to develop the method to estimate dynamic OD traffic volume. This study proposed the method for estimating dynamic OD traffic volume from traffic counts and AVI (Automated Vehicle Identification) data. The method originates from the least square method for the error between observed and estimated traffic volumes.4) A system model for on-line traffic control for urban expressway was developed by integrating the inflow control model, the meso traffic simulation model and the method for estimating dynamic OD traffic volumes. The validity and practicability of the system model was confirmed by some cases of numerical examples for an urban expressway network.