Early Warning Model for Traffic Congestion on Freeway

博士 === 逢甲大學 === 土木水利工程與建設規劃博士學位學程 === 107 === In the previous studies we found that ramp metering can effectively control the number of cars from getting into the main lane, in order to reduce the congestion of the main lane, but the method is beforehand calculating the historical data of instrument...

Full description

Bibliographic Details
Main Authors: HUANG,CHI-CHANG, 黃啟倡
Other Authors: Lin,LIANG-TAY
Format: Others
Language:zh-TW
Published: 2019
Online Access:http://ndltd.ncl.edu.tw/handle/cb2kzw
Description
Summary:博士 === 逢甲大學 === 土木水利工程與建設規劃博士學位學程 === 107 === In the previous studies we found that ramp metering can effectively control the number of cars from getting into the main lane, in order to reduce the congestion of the main lane, but the method is beforehand calculating the historical data of instrument control rate ramp instrument to release the data to ramp instrument controller in advance; or on the other hand, only proceed the control when congestions occur at the main lane. The former cannot immediately control rapid traffic flow according to traffic changes, while the latter will cause a continuously expanding congestion, needing more time to alleviate the congestion. This study aims to find the capacity and critical density of the studied section according to the previous literature, and develop early warning of traffic congestion threshold that gives enough buffer time in prior for responding and issuing control measures in time. Because the threshold parameter value is determined still with fuzzy perspectives, fuzzy logic control method is adopted to confirm the threshold value of congestion early warning. The fuzzy logic control (traffic flow mode) proposed in this study can effectively increase the alarm coverage rate, while the modified mode can increase the alarm coverage rate to more than 80% and the result verified by examples proved it can reach more than 70%. Although the alarm accuracy is low, but please consider to apply this mode to the non-continuous period. Because density k is 60(vehicle/km/hour) and fuzzy logic control (traffic flow mode) has a high warning accuracy, it is suggested that it can be used for congestion warning during a single holiday or even long holidays. In this study, TSIS is used to construct a simulated road network, and actual data are used to verify it. When congestion warning occurs, within minutes in advance, each ramp in the upstream section is controlled and successfully reduced the traffic volume by about 20%, which can effectively reduce the congestion and creating it.