Modeling Traffic Flow on Urban Highways: the Application of Cellular Automata and Nested Logit

Abstract

The current study attempts to model the drivers’ behavior in a basic freeway in order to use it in future simulations and predictions as well as to find out its effective variables and analyze the sensitivity of model to them, using a cellular automata model. Cellular automata are considered as one of the applied methods in traffic flow simulation. These models consist of a set of regular and finite grid of cells each of which can be in one of a finite number of defined states. In order to find the rules, a nested logit model is constructed. The parameters of the nested logit model are then estimated and analyzed. In order to be able to estimate the parameters, real data are needed. Therefore, the data which are extracted from a selected highway in Iran are used for the parameter estimation of the nested logit model. Results of the analysis confirmed the capability of nested logit model in simulation of traffic flow under irregular driving conditions, in which lane changing occurs frequently. They also showed that crash risk has no effect on reduction of longitudinal speed of vehicles, while the maximum safe speed plays a main role in driving behavior, showing that the focused drivers choose their next position in accordance with the maximum safe speed. The current paper’s center of attention is its methodology. This method can be used to improve the utilization of the existing systems (without lane addition) through better traffic management and operations strategies, and improve the geometric design of roads and highways, and consequently improve mobility, safety, and air quality, which are of great importance.