Comparative analysis of microscopic models of road traffic data

Traffic tools simulations are based on different models whose purpose is to describe the behavior and dynamics of traffic flow on network. There are three modelling approaches which describe the road traffic flow: macroscopic, mesoscopic and microscopic. The microscopic models are used to describe the interactions between vehicles and their environment. They model the individual behaviour of each vehicle considering, in particular, interactions between two successive vehicles. Microscopic models aim to reproduce the behaviour of each pair vehicle-driver. In this paper, two car following models have been implemented: linear model [1] [2] and intelligent driver model (IDM) [3]. However, their shortcomings emanate from the difficulty to appreciate the consistency of behaviour of all traffic flow. Emphasizing on microscopic simulation, this paper is a comparative analysis of these microscopic models by identification of properties and limitations of each model. In order to describe road traffic at a granular level, the present study focuses on the kinematic variables. The variables used are essentially: spacing headway, relative speed, position, velocity and acceleration. They are designed for each vehicle and are closely related to its dynamics. In fact, microscopic models should be able to reproduce a real traffic situation, that´s why an efficient calibration on real data is necessary to clarify the relevance of results. Therefore, the calibration of these microscopic models would be on real data.