A Multiple Representation Entity-Based Approach to Hybrid Traffic Simulation Model

Abstract

Traffic simulation is a widely used tool to analyze dynamic traffic system. In traffic simulation field, Microscopic simulation offers detailed representation of traffic system, but it is difficult to model large traffic network; Macroscopic and mesoscopic simulations can capture the dynamics of large network, but only on a low level of details. These limitations of mono-resolution traffic simulation emerge more clearly, as the development of its application. Multi-resolution traffic simulation, which uses hybrid traffic simulation model, can overcome those limitations by integrating traffic simulation models on different levels into a hybrid model, which combines advantages of all composing models. In this paper, we introduced UNIFY framework into traffic simulation and developed UNITRANS, a Multiple Representation Entity-based (MRE-based) framework, to integrate traffic simulation models on different resolution levels and maintain the consistency among their jointly-executing. A hybrid model was built using UNITRANS and a three-legged intersection was modeled with it. The results showed that UNITRANS was feasible and effective.