Keywords: traffic flow, cellular automata, stochastic velocity model, urban city traffic.

This paper describes the traffic flow simulation in an urban city using a cellular automata simulation. The traffic flow simulations based on the cellular automata have been presented by a number of researchers. The first traffic simulation model which is named the rule-184 CA model has been presented by Wolfram[1], The vehicles move at constant velocity and do not accelerate even when the distance between vehicles is very large. Therefore, the rule-184 CA model cannot simulate the actual traffic flows well. For overcoming this difficulty, the Nagel-Schreckenberg (NaSch) model[2] and the vehicle following model have been presented. In the models, the vehicle velocity is changed in a random way or according to the distance between the vehicles. When, however, it is applied to urban city traffic flow, the behaviour rules of the vehicles become very complicated due to the existence of the traffic signals, the intersections and the branch lines. For overcoming the problem, the authors have presented the stochastic velocity model (SV model)[3]. In the SV model, the movement of the vehicle is controlled with the random variable. Only the vehicles driving at the maximum velocity can move by one cell at each time step and the movement of the vehicles driving at a velocity smaller than the maximum velocity is performed according to the random number. Since the moving range of each vehicle is restricted to one cell, the local rule for the vehicle movement can be simplified.

In this paper, the SV model is applied to the traffic flow simulation in Nagoya City, Japan. The city road network is defined with the cells of length and width and each vehicle is expressed with two cells. In the stochastic velocity model, the vehicle velocity is defined with the stochastic variable and the vehicle movement is up to one cell at each time step. The vehicle behavior is controled according to the local rules and the distance between the vehicles. The local rules are classified into the behavior, velocity and the movement local rules. The behavior local rule is composed of the rules for a through-driving vehicle, a left-turning vehicle and a right-turning vehicle. The velocity local rules are classified into relative and absolute local rules. The former changes the vehicle velocity according to the distance between the vehicles and the latter decelerates the vehicle safely to avoid obstacles such as traffic signals, stationary vehicles, and so on.

The simulation model is applied to the traffic flow simulation along the Nagakute-Hongo way in east Nagoya City. In the object domain to be considered, there are eight intersections, seven two-lane roads and one four-lane road. The traffic quantity has been estimated from 6:30 AM to 9:30 PM on November 19, 2001. Before performing the simulation, the vehicle characteristics such as the safety distance between vehicles and the vehicle acceleration rate should be defined in advance. They are estimated from the data of the actual vehicles.

The inflow traffic amount from each road is specified according to the actual estimated data and the outflow traffic amount is estimated at eleven estimation points. The outflow traffic quantity determined by the simulation is compared with the real traffic data[4]. The simulation result well agrees with the real data at each estimation point. Finally, the validity of the present simulation model is confirmed.

1

S. Wolfram, "Cellular Automata and Complexity", Adison-Wesley Publishing Company, 1 edition, 1994.

2

K. Nagel and M. Schreckenberg, "Cellular automaton model for freeway traffic", Journal of Physics I France, Vol. 2, pp. 2221-2229, 1992. doi:10.1051/jp1:1992277

3

T. Tamaki, S. Yasue and E. Kita, "Traffic Flow Simulation using a Stochastic Velocity Model and Cellular Automata", in Proceedings of the Seventh International Conference on the Application of Artificial Intelligence to Civil and Structural Engineering, B.H.V. Topping, (Editor), Civil-Comp Press, Stirling, United Kingdom, paper 66, 2003. doi:10.4203/ccp.78.66

4

Transport Lab, Hongo area benchmark dataset. http://www.i-transportlab.jp/bmdata/HongoBM, 2001.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £135 +P&P)