Factor analysis to choose the input variable of the car-following model for microscopic traffic simulation

With the rapid development of urban traffic, the driver's behavior has played a more and more important role to traffic safety and capacity. For the existed car-following models in microscopic traffic flow simulation, most of models were on the basis of the sensible understanding to traffic phenomena; some of the models have made statistical analysis of the field data, but not come down to the question of the choice of input variables. In this paper, the nonlinear multivariable statistical method of factor analysis is used to extract the useful information from the representative field data to seek the basic structure of the car-following model and several inner variables with higher common factor loading and information capacity for car-following process.