Comparison of propagation and packet error models in vehicular networks performance

Abstract Over the years, we have witnessed how wireless communications and transportation technologies converge. Likewise, the research on the so-called vehicular ad hoc networks (VANETs) makes more and more sense. Due to the high costs and effort of deploying vehicles in real scenarios, network simulation is a popular alternative when studying the performance of vehicular networks. However, the dynamism of such environments makes their simulation a rarely simple task. In fact, many parameters have to be considered by VANET simulation, such as a dynamic topology, omnipresent obstacles, traffic flow, different mobility models, traffic lights, changing vehicular speeds, etc. Unsurprisingly, the way how some events are modeled in a simulation might bias the performance measures of vehicular communications. Accordingly, in this paper, we concentrate on assessing the impact of packet error modeling on VANET simulations. With this aim, we measure different parameters such as losses, end-to-end delay, and number of hops over a realistic urban scenario. We also test three different densities of nodes and three channel capacities. The performance metrics obtained after our simulations suggest that, in the best cases, the basic packet error model may obtain reliable results (i.e., results similar to those from a realistic error model). This evaluation is performed over a multi-hop scenario in which the antenna is configured with high sensitivity values. This simple technique bases on the assumption that interference levels will not exceed the SINR (Signal to Interference and Noise Ratio) threshold when errors begin to appear. This requirement may not be met in all simulations, for example in simulations with very high traffic load.