IEEE 802.11p Performance Evaluation: Simulations vs. Real Experiments

The employment of IEEE 802.11p-compliant devices in the automotive environment is crucial to traffic safety and control applications. Thus, performance evaluation of the standard before its widespread adoption in real systems is of paramount importance. Nonetheless, due to the high cost and reasonably low availability of off-the-shelf devices, most of the work involving IEEE 802.11p vehicular networks is still based on simulations. In order to investigate the distance to real-world equipment, this work investigates the results of experimentation with commercial devices, in V2I (Vehicle-to-Infrastructure) and V2V (Vehicle-to-Vehicle) scenarios. The performance obtained by real on-board and road-side units is then compared to those presented by widely used NS-3 simulator. Three key metrics are evaluated: the maximum range, packet delivery rate (PDR), and packet inter-reception time (PIR). The influence of different modulations defined in IEEE 802.11p and of different mobility levels are analyzed. For both simulation and experimentation, the results agree on the PDR degradation at intense vehicle speeds and higher PHY data rates. Nevertheless, the results in terms of maximum range, moderate speed impact, and the weak correlation between the PIR and PDR show that NS-3 simulation models still need to evolve.