The miniaturization of electronic devices, starting from personal digital assistants down to imperceptible sensor systems, combined with cheap and effective communication evolved in the last years into the new idea of integrating these trends into a network of things. This trend, the third wave in computing, is called Ubiquitous Computing.

In current research a lot of different groups work on a multitude of sub areas of Ubiquitous Computing. One of these research groups is the Center of Excellence 627 at the Universität Stuttgart that works on Spatial World Models for Mobile Context-Aware Applications. The main idea of this group is to find solutions for organizing the enormous amount of information from (static and mobile) electronic devices and the communication in between them in an open and flexible infrastructure, called Nexus. This infrastructure will be able to deliver a consistent view on a digital model of the real world. To show the applicability of Nexus, in this paper we want to propose an application from traffic telematics as an example for mobile context-aware systems. There we can show on the one hand the benefit of Nexus for new applications. On the other hand our application serves as a prototype for the Nexus concepts.

The main idea of the application is to use the digital world models of Nexus for a dynamic traffic assignment on a microscopic level. In the classical chain of traffic modeling (demand, choice of mode/time and choice of route as an input for traffic flow modeling) we will first focus on the choice of route and time. The influence of external parameters on the route choice behavior is still a present question in traffic modeling. Therefore in our application we will use the new trends in informatics for traffic engineering problems. Especially the (admittedly visionary) assumption of having an extensive view on traffic data in the so called world models of Nexus will be used to investigate new methods for choice behavior modeling. Expanding our existing approaches for modeling the traffic flow, we will secondly make use of the historic view on the digital world model for calibrating the traffic flow simulation. Typically this would be done by centrally analyzing the data that is relevant in the simulation in one application and comparing it to the real situation. A decentralized concept is to use VANETs. In cooperation with co-projects in Nexus we will investigate the portability of the central situation recognition to decentralized VANETs.

This paper shows the principle ideas and a first concept of implementing a dynamic traffic assignment in a ubiquitous system. Further it shows the potential of the interdisciplinary cooperation. The research will be done in cooperation with computer and telecommunication scientists founded by the German Research Foundation.