Agent - Based Traffic Simulation
using Microscopic Modelling

Abstract

An agent is simply an object or class entity with its own attributes and methods. Agents simulate realistic behavior through efficient model abstraction while minimizing software complexity i.e. to synthesize life like systems to exhibit behavior of natural living systems . A microscopic model is able to simulate traffic in urban areas in real time for use in driving simulators. The vehicles considered in the simulation , namely user-driven vehicle at the center of the simulation model and other vehicles interact with its surroundings . Simulation is done in control zone and traffic light simulation is included only in this reduced control area. In order to understand vehicle dynamics , a wide range of microscopic models are used like car following model , lane changing model etc. which provides a realistic modeling of driver and vehicle behavior , and requires high computation resource. The primitive models developed here are the basic intersection model of traffic lights. The color changes indicating the permission for one lane to process the instruction one at a time whereas the other lane needs to wait. It is observed that for random traffic along the lanes the waiting time remains regular along both directions. Traffic light alternates between green and red phase. By varying the time of green and red phases, the number of vehicles is controlled and observation are carried out in 6 lanes. This greatly depends upon the traffic density or throughput of the vehicles. Traffic throughput is observed through the waiting time associated with each vehicle. Unlike highway environments , in urban nature consists of different other constraints governed by VANET routing protocols to carry and forward mechanism to deal with challenges.