Comparative evaluation of ramp control methods: A case study on D-100 freeway, Istanbul

Freeway networks are designed to provide a flexible flow of vehicles and uninterrupted routes especially in the metropolitan areas where the traffic volumes are considerably high. However; when the traffic volumes reach the highway capacity, speed differentials occur leading to phenomenon of congestion which either can be recurrent in peak hours or nonrecurrent in case of having an incident on the highway. Congestions are considered a case of matter as the freeway network performance is adversely affected in terms of reducing the speed, increasing the travel time, and increasing the vehicle emissions due to increase the fuel consumption. Also the distraction during congestion period, can increase the probability of accidents occurrence, where all drivers try to pass the congested segment of the freeway. Different solutions were considered for solving or reducing the effect of congestion on freeways. The structrual solutions for example by adding additional lanes in order to increase the system capacity are sufficient up to a certain time, later the congestion starts to form again. The reason is that the traffic volume in another word the demand is driven by increasing the capacity, rather than other enviromental and econmic considerserations that do not allow to increase the system capacity. Thus, since 1960s the trend have been changed to control the link traffic and optimize its capacity by using coordinated and integrated surveillance and control strategies. Several control approaches and strategies are introduced including ramp metering, variable speed limits, and route guidance systems. For the recent decades ramp metering strategies are noticed to be an effective approach for controlling and managing the freeway traffic with the forms of local ramp metering such as ALINEA, PI-ALINEA, and Model Predictive Control, and coordinated ramp metering such as FLOW, HERO, and SWARM. In this thesis the feedback control based algorithms ALINEA and PI-ALINEA are evaluated by uitlizing them under different scenarios on a section of the D-100 freeway, Istanbul. The evaluation process considers different measures of effectiveness such as the travel time reduction, speed increase, volume increase, and occupancy reduction. The algorithms are applied in a micrsoscopic environment using VISSIM integrated with MATLAB via COM interface. The microscopic model is calibrated using a real data provided by the minucipality of Istanbul and the control algorithm is conducted online through MATLAB via COM interface. According to the considered measures of effectiveness such as the increase in the speed, the occupancy reduction, and the improvement in the acceleration rates, PI-ALINEA outperformed ALINEA which agrees with the literature.