Impact of packet error modeling in VANET simulations

Simulation is a very valuable mechanism to study the performance of vehicular ad-hoc networks (VANETs), due to the high cost and effort involved in the deployment of vehicles in urban scenarios. Accordingly, as in real VANETs, simulations should involve several factors related to realistic topologies, movements, obstacles, error modeling, etc. In this paper we concentrate on evaluating the impact of packet error modeling in VANETs, by measuring losses, end-to-end delay and number of hops in realistic scenarios. For this, we considered three different nodes' densities and three channel capacities. Through simulations in a multi-hop vehicular scenario, we show that a basic packet error model can generate very similar results to the obtained by a realistic packet error model, when the configuration of the former is properly set at a low or medium channel capacity. Nevertheless, we found significant differences between the results of both models when the simulations use a high capacity channel (27 Mbps) even with a low data traffic load.