VANET applications are often providing street traffic information to vehicles and drivers, regarding, for instance, traffic conditions and parking space availability. This information influences in turn the driving behavior in real-world settings. Mobility models used in current VANET simulations are mostly ignoring this feedback entirely. In cases the feedback is included, it is mainly based on ad-hoc approaches with lack of generality. With this paper, we contribute to the investigation of such feedback loops within VANETs by describing the levels at which feedback loops can be introduced, i.e., on strategic, tactical, and operational levels of mobility. We further describe how feedback loops can be introduced in arbitrary mobility models and in particular in elementary mobility models. We exemplify our approach by introducing two types of feedback loops for the Manhattan Mobility model, the Random Trip model, and the Constrained Random Trip model. One feedback loop represents points of interest attracting vehicles, such as free parking spaces attracting vehicles searching for parking. The other feedback loop focuses on repelling vehicles, such as a traffic jam. We discuss the impacts of the feedback in terms of the mobility metrics: vehicle density per area, number of direction changes, and intensity of direction changes. Furthermore, we discuss the effects in terms of information availability and delays of transmission in an opportunistic vehicular network.

Peer Reviewed