LEE- Ulaştırma Mühendisliği Lisansüstü Programı

Bu topluluk için Kalıcı Uri

http://hdl.handle.net/11527/19476

Gözat

Yazar "Ademoğlu, Muhammed" ile LEE- Ulaştırma Mühendisliği Lisansüstü Programı'a göz atma
  • Öge
    A game-theoretical approach for analyzing effects of combined control on freeway traffic: Case of integrated ramp metering and variable speed limiting
    (Graduate School, 2024-11-27) Ademoğlu, Muhammed ; Çelikoğlu, Hilmi Berk ; 501201426 ; Transportation Engineering
    Freeways are designed to meet high transportation demands at high service levels. According to the Highway Capacity Manual (HCM), freeways are fully access-controlled, divided roads with at least two lanes in each direction. Throughout their service life, freeways encounter both recurrent and non-recurrent congestion. Recurrent congestion, as the name suggests, occurs regularly and has predictable causes. Non-recurrent congestion, on the other hand, does not occur repeatedly in the same place or time and is not predictable. An example of non-recurrent congestion is when lanes become unusable due to an accident on the road. Recurrent congestion, on the other hand, can be exemplified by traffic congestion caused by increased demand during peak hours. Modern transportation and traffic engineering approaches aim to enhance capacity not through new investments, but by adopting Intelligent Transportation Systems (ITS), traffic management, and traffic control methods to ensure more efficient and effective use of existing capacity. Under the ITS framework, various control methods have been developed to combat highway congestion. These methods include ramp metering (RM), variable speed limits (VSL), and route guidance (RG). These control strategies aim to optimize and improve the utilization of road capacity. Ramp metering aims to control traffic flows merging with the main stream via on-ramps, by delaying these merging flows in a controlled manner, thereby maintaining the density of the main stream below a certain threshold. This approach seeks to keep average speeds high and achieve a more stable traffic flow. Another ITS system, the variable speed limit application, aims to enhance flow homogeneity by reducing the speeds of vehicles in the main stream, thus delaying or preventing congestion. In some studies, variable speed limit has been integrated with ramp metering to combine the advantages of both methods, targeting more efficient use of capacity. In this thesis, a framework for modeling, control, and analysis has been developed to investigate the effects of autonomous vehicles on traffic flow dynamics and freeway capacity. Compared to similar studies in the literature, the unique aspect of this work lies in its game theory-based comparison for evaluating traffic flow performance. Game theory is a mathematical framework used to analyze strategic interactions between individuals or groups, where each player aims to optimize their own outcomes by considering the potential actions of others. This analysis evaluates possible strategies and their outcomes to predict the most rational decisions players might make. Game theory is widely used in fields such as economics, politics, sociology, psychology, and even biology to understand and model decision-making processes. In this context, an analysis based on the game defined within the thesis has been conducted. The proposed framework was tested using microscopic simulation, based on a model developed using real network and demand data. This study examined a 5-kilometer section of the D-100 freeway in Istanbul, extending from Zincirlikuyu to Halıcıoğlu. Ramp metering and variable speed limit algorithms, commonly used in the literature, were integrated into the simulation model. In the study, the Asservissement Linéaire d'Entrée Autoroutièr (ALINEA) algorithm was used for ramp metering. ALINEA is a local ramp metering algorithm with a closed-loop structure. This algorithm was integrated into a microscopic simulation software via the Component Object Model (COM) interface of the MATrix LABoratory (MATLAB) environment and applied to three ramp metering sections. For the variable speed limit control, a commonly used algorithm from the literature was also integrated into the simulation model. Additionally, an H∞ controller-based algorithm, which applies both ramp metering and variable speed limit control in an integrated manner, was selected as the third control scenario. Along with the uncontrolled scenario, a total of four primary scenarios were evaluated in the study. In each control scenario, the percentages of autonomous and human-driven vehicles were gradually adjusted, and the effects of these percentages on traffic flow under mixed traffic conditions were analyzed. For each control method, 11 different scenarios were created by gradually increasing the percentage of autonomous vehicles, and the results were compared to those of the uncontrolled scenarios. The findings were analyzed using various performance metrics, and, finally, a new performance comparison method based on game theory was proposed within the scope of the thesis. The findings obtained from this study are organized under two main headings. The first heading focuses on traffic flow performance-based analyses. In this analysis, the scenario without control, ramp metering, variable speed limit, and the integrated control scenario based on the H∞ method were evaluated based on changes in the percentage of human-driven and autonomous vehicles in the traffic flow. The evaluation was conducted using metrics such as total vehicle throughput, total travel time, and the mass of pollutants emitted by the vehicles. For both the controlled and uncontrolled scenarios mentioned above, the critical percentages of autonomous vehicles were determined. The effectiveness of these control scenarios in mixed traffic conditions was assessed by comparing the controlled scenarios to the uncontrolled scenario. The other focus of the findings is a game theory-based analysis. In this analysis, the players were first defined, and the game was structured around two players. These players are the vehicles traveling in the main traffic flow and the vehicles entering the highway from on-ramps. In the sub-scenarios created, the four different control scenarios mentioned above were considered. In this two-player game, it was assumed that one of the players consisted entirely of human-driven vehicles. For the other player, the proportion of autonomous vehicles was gradually increased from 0% to 100%, allowing for a detailed investigation of the impact of autonomous vehicles in mixed traffic. In this context, the effect of autonomous vehicles on traffic flow was analyzed based on delay per vehicle.