A Stochastic Performance Model For Dense Vehicular Ad-Hoc Networks

Abstract

The usual approach to assess performance in dense urban networks is to use mobility simulators and network simulators. This implies strong simplifications and abstractions on the physical layer, while providing fine-grained resolution on the geometry and packet properties. In this paper, we extend the usual approach by including more detailed physical-layer models. This gives a higher-fidelity baseline against which we can compare simplifications. Then, we proceed to reduce the geometric vehicle distribution to a stochastic interferer process. Finally, we show the differences in simulation fidelity at different stages of abstraction. We base our analysis on IEEE 802.11p and demonstrate that the typical simplifications, while well adapted to static networks, fail to capture essential properties of a highly mobile channel.

The usual approach to assess performance in dense urban networks is to use mobility simulators and network simulators. This implies strong simplifications and abstractions on the physical layer, while providing fine-grained resolution on the geometry and packet properties. In this paper, we extend the usual approach by including more detailed physical-layer models. This gives a higher-fidelity baseline against which we can compare simplifications. Then, we proceed to reduce the geometric vehicle distribution to a stochastic interferer process. Finally, we show the differences in simulation fidelity at different stages of abstraction. We base our analysis on IEEE 802.11p and demonstrate that the typical simplifications, while well adapted to static networks, fail to capture essential properties of a highly mobile channel.