Unorganized Traffic Simulation in Urban Street

Abstract

A tra c phenomenon is one of the most complex phenomena to simulate because it is di cult to model autonomous behaviour of drivers and their interactions. This research study presents a model to simulate unorganized tra c which is dif- ferent from other lane-based or following-based approaches. Organized tra c with drivers heeding to well de ned tra c rules is less dynamic and erratic than mod- elling unorganized tra c, wherein the behaviour of the whole system depends on the interplay of di erent human characteristics. The overall tra c pattern can be viewed as emergent behaviour of such individual interactions. This research applies multi-agent simulation for unorganized tra c. The study introduce the concept called aggressive level and tra c parameters of individual drivers such as preferred speed, maximum acceleration and safety distance are decided accord- ing to aggressive level. In three sub models called communication, car following and switch route has been introduced in design and the motion of vehicles has been modelled using kinetic equation of motion. The impact of adverse weather condition has been also integrated to proposed model. Under adverse weather conditions the impacts of heavy rain and light rain have been mainly focused. A multi-agent simulation has been developed with 3D graphics using OPENGL. In particular, the study shows statistics about average speed, number of overtakes with di erent proportions of aggressive drivers and under di erent weather states.