The Development of Macroscopic Hybrid Model on Freeway Ramp Metering Control

• Main Authors: Cheng-Hsiu Tsou, 鄒政修
• Other Authors: 許添本
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/43065222820032153214

碩士 === 國立臺灣大學 === 土木工程學研究所 === 100 === ABSTRACT Urban and interurban freeways had been originally built to provide unlimited mobility to road users. However, the rise of new cities and metropolis need new interchanges to connect the mainline, combined with the on-going dramatic expansion of car-ownership, has inevitably led to the daily appearance of recurrent and non-recurrent freeway congestions of thousands of kilometers in length around the world. Ramp metering is one of the most efficient ways in freeway management systems; it imposes short delays at on-ramps and leads to substantial savings for each individual road user for its function of mitigating congestion area from the angle of time and space along the freeway. The ramp metering algorithm in Taiwan now is classified as local ramp metering which has a built-in traffic responsive function. The shortcoming of local ramp metering lies in its control strategy: only considering the sole ramp and the nearest upstream mainline as control target. The more stringent control strategy will not be activated until the upstream detector perceives that the congestion had exceeded the predetermined metering threshold. The isolated control characteristics and lack of real-time control ability has made local ramp metering harder and harder to deal with long stretch congestion on modern freeway. The dissertation aims at developing a coordinated ramp metering control strategy by using mathematical programming method to create optimization model, which is consistent with the traffic responsive logic and uses detector data to solve optimal ramp metering rate at every control time step. The objective is to improve current control strategy from the status of low reaction time and isolated characteristic. The dissertation also concerns the development of hybrid macroscopic freeway traffic model that is computationally efficient and suitable for use in real-time traffic monitoring and control applications. A primary contribution comes from the combination of traffic flow model, modified cell transmission model (MCTM) and macroscopic incremental node model (INM). The two models are welded into one by using the same vehicle propelling logic and the same computation method on flow transmission. The research achievement includes: 1. Collecting and sorting out detector data to conduct congestion analysis, so as to confirm the coordinated ramp metering scope on mainline. Detector data is also applied to calibrate parameters; it is assigned to cells by complying with a proper considered importing principle. 2. The newly developed hybrid traffic model is proved to have reasonable traffic descriptive ability by complying with a validation principle and MAPE evaluation indicator. 3. Conduct optimization models which differs from their objective functions: with or without penalty term as two scenarios to solve coordinated metering rate separately and analysis the metering results. 4. Consider future practical implementation, a solving architecture is proposed to link academic research and practical applications. 5. VISSIM simulator is used to construct simulation networks and serves as a control strategy operating platform that exports control performance in every trial. The final control performance shows that the coordinated ramp metering control strategy developed in the dissertation can ameliorate main line long stretch congestion, which meets the initial research objective.