Freeway traffic flow modeling and control with emphasis on congested off-ramp areas

Abstract

Most metropolitan cities of developed countries have now an extensive network of urban and peri-urban freeways, which aims to provide virtually unlimited and fast mobility to road users around the metropolitan area. However, the increase of traffic demand, especially during the peak hours, and the occurrence of traffic incidents, leads to daily appearance of recurrent and non-recurrent freeway congestion which results in significant increase of travel times, increased fuel consumption, environmental pollution as well as reduced safety. The problem of freeway congestion in urban and peri-urban freeways cannot always be faced by expanding the existing infrastructure, for economic and environmental reasons; instead, efficient traffic control measures may be employed to mitigate the problem. However, the development of effective real-time traffic control measures implies the availability of suitable mathematical traffic flow models which may be used for the development and testing of the proposed control strategies. This thesis investigates the particular, but quite frequent, case of (recurrent) freeway congestion due to saturated off-ramps. This kind of congestion is difficult to deal with, and for this reason this frequent traffic flow degradation is rarely addressed in the traffic control literature. Moreover, within the traffic flow modeling literature there are, so far, no studies undertaking validation and comparison of different traffic flow models regarding the reproduction of traffic conditions in such areas. The aim of this research is to investigate traffic flow modeling and traffic control issues for congested freeway off-ramp areas. In particular, within this thesis the most popular discrete time-space macroscopic traffic flow models, namely the CTM and the METANET models, were validated and compared regarding the representation of traffic conditions at congested freeway off-ramp areas. The models were calibrated and validated using real traffic data from Attiki Odos freeway in Athens, and by employing various optimization methods. Apart from the modeling approach, various innovative real-time traffic control measures were developed for congested freeway off-ramp areas. In particular, two different cases were examined and suitable traffic control strategies were proposed for every case. In the first case, a hypothetical network was simulated, and various route diversion strategies were developed that aim to reroute the drivers through alternative routes, towards the same destination, preventing the off-ramp queue spillover and the creation of mainstream congestion. In the second case, a real traffic network was examined where recurrent freeway congestion is created due to congestion on the surface street network which propagates to the freeway mainstream through a saturated off-ramp. The network was simulated by use of microscopic simulation and a real-time merging traffic control algorithm was proposed that aims to maximize the surface street network throughput and at the same time to prevent the off-ramp queue from spilling back into the freeway mainstream. The simulation results, for both investigated cases, showed that the proposed traffic control measures can improve the prevailing traffic conditions, preventing the formation of mainstream congestion. Thus, they are both very promising for a field implementation.