The doctoral dissertations of the former Helsinki University of Technology (TKK) and Aalto University Schools of Technology (CHEM, ELEC, ENG, SCI) published in electronic format are available in the electronic publications archive of Aalto University - Aaltodoc.
|
|
|
Studies of Traffic Situations Using Cellular Automata
Arto Hämäläinen
Dissertation for the degree of Doctor of Science in Technology to be presented with due permission of the Department of Engineering Physics and Mathematics for public examination and debate in Auditorium E at Helsinki University of Technology (Espoo, Finland) on the 9th of October, 2006, at 13 o'clock.
Dissertation in PDF format (ISBN 951-22-8369-7) [5575 KB]
Dissertation is also available in print (ISBN 951-22-8368-9)
Abstract
The growing volumes of vehicular traffic necessitate new solutions to traffic problems. When changes to the road network are planned and traffic control systems are set up, it is usually not possible to test in advance the resulting effects in the real world. In such situations simulation models can be of great help.
Traffic simulation models can roughly be divided into macroscopic and microscopic ones. When macroscopic models examine the dependencies between traffic flow, traffic volume, and average velocities, microscopic models investigate the movements of individual vehicles. Microscopic models for vehicular traffic include car-following and particle hopping models — the latter are usually implemented as cellular automata (CA). Despite the simplicity of the CA models, the most important features of traffic flow can be modelled with them.
From the physical point of view, vehicular traffic can be seen as a nonequilibrium many-particle system with the flow of traffic fluctuating between different dynamical phases.
This work deals with simulation of vehicular traffic in urban and freeway environments. The behaviour of simple CA models is studied in connection with providing indicators — queue lengths at traffic signals, average velocities on different routes, and flow rates — describing traffic situations in a road network. The knowledge of traffic situations is important for traffic management and control, which aim in maintaining road traffic fluent, comfortable, and free of accidents and excess pollution effects. — The results also suggest how well the measuring procedure functions in generating the above-mentioned indicator values.
Road traffic is studied in different simulation test environments. The urban cases comprise two- and five-intersection areas with multiple lanes and traffic signal controlling. The freeway area corresponds to a stretch of a two-lane freeway with several on- and off-ramps. The vehicle arrivals are generated on a random basis, or they are gathered from measurements on real locations. The signal status changes use fixed period lengths. The performance of the CA models is examined against the functioning of a more refined simulation model that uses car-following approach. The simulation periods are typically from two to five hours in real time.
The results show that CA simulation usually produces traffic situation indicator values well when compared to those of the reference model, especially when averages over several runs are calculated. Separate runs may generate dissimilar outcome, which can create problems when used in on-line simulations. The results can be improved by adjusting the measuring technique (the detection of vehicles inside the network, for example).
The CA traffic models are also used to study the statistics of queue formation at some obstacle on a one-lane road. The probability for the effective queue length area seems to obey exponential decrease with increasing length values. The queue lengths grow rapidly with higher flow values.
Keywords: cellular automata, vehicle traffic simulation, traffic situations, vehicle queue formation, average vehicle velocities, intersections
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
© 2006 Helsinki University of Technology
Last update 2011-05-26