Architecture and Development of Agent-Based Unified Simulation Environment for ITS Services

Abstract

Agent-based traffic simulation has become increasingly interesting and important for developing new intelligent transport systems (ITS) services. Till date, a variety of studies and developments that combine simulators and evaluate ITS services on the combined simulators have been conducted. In this paper, we propose a simulation environment called as Agent-based unified simulation environment (USE) for ITS services. To confirm the effect of ITS services on the society via simulation, it is generally necessary to implement a model of the service on a simulator. The Agent-based USE provides an easy-to-build simulation environment for ITS-related services. Particularly, by connecting simulators with ITS services, the Agent-based USE determines behaviors to be changed on the simulators using the data of the services such as recommendation results generated by the services and communicates the decisions to the simulators; the Agent-based USE then obtains the data representing the current situation on the simulators and sends it to the services as feedback so as to enable the services to generate the next recommendation. Additionally, using the Agent-based USE, it is possible to construct a co-simulation environment where simulation is performed by synchronizing some types of simulators and services and by sharing simulation information. In this paper, we introduce the overview and architecture of the Agent-based USE for traffic simulation and discuss its usefulness.