Duration: 18:35

196 views

3

Nowadays, smart mobility applications could benefit from environment perception, enabled by evolving sensor technology and processing capabilities available for the traffic entities. On the application level, in many cases, information about detected objects is required instead of the raw sensor data.

Developing and evaluating the impacts of such applications can be done in co-simulation frameworks, which combine modeling of different domains such as application, communication, and traffic. Eclipse MOSAIC is a suitable solution for this task, combining the traffic simulation of Eclipse SUMO with other simulators, such as the integrated Application Simulator, or OMNeT++ and ns-3 for modeling communication.

However, a model for perceiving surrounding traffic entities, such as vehicles, traffic signals, and traffic signs, is only available to a limited extend. In this paper we introduce an object-level perception module to the MOSAIC Application simulator. It takes advantage of state-of-the art spatial indexing methods to get rapid access to traffic objects, especially moving objects, within a defined field of view.

We furthermore evaluate the computational performance of the indexing techniques as well as the integration with the traffic simulator SUMO using TraCI and Libsumo. With the aid of this model, novel connected applications which analyze or share surrounding objects, e.g. for an improved traffic state estimation, can now be evaluated with Eclipse MOSAIC.