A Distributed Processing Architecture Using Driver Behavior and a Physical Model for Large-Scale Traffic Simulation

Abstract

Traffic simulation is important technology for assessing the influence of traffic on the environment. It is also necessary to model driver behavior to evaluate the influence of new ITS services. In this paper, we describe a layered model which adds a physical car dynamics layer to the three-layer model of driver behavior and proposes a new distributed processing architecture to reduce the computational complexity using the space locality and time granularity of each layer.