Assessing the Impact of Realistic Simulation Environment on Vehicular Communications

Over the last few decades, simulations have emerged as the primary tool in vehicular ad hoc network (VANET) research. The unique characteristics of Vehicle-to-Vehicle (V2V) communications require more detailed and flexible simulation approaches. For example, incorporating V2V channel models which take into account site-specific and different types of obstructions, vehicular movement patterns and their impact on the system-level performance metrics. However, most of the simulation tools use simplified techniques at the physical (PHY) layer to achieve an optimum mix of accuracy and computational complexity. In this paper, we utilized a generalized simulation framework that combines a suite of tools for assessing the impact of a realistic simulation environment on VANET performance. The objectives are twofold and closely related. Firstly, to obtain a more realistic representation of vehicular channel characteristics using building data and vehicular mobility traces at a citywide scale. Secondly, to augment the network simulator with the results collected from representative V2V channel model to further improve on the accuracy of simulation-based studies. The simulation results show significant differences in terms of performance metrics, e.g., Packet Loss Ratio (PLR), throughput and reliable communication range between the combined simulation framework and the standard propagation models implemented in the network simulator.