Rethinking Wireless Network Simulation
David Meacock · 2006
Wireless network simulation, though a popular field, suffers from a distinct lack of credibility in parts due to the presence of many simplifying assumptions in current research. Of these, perhaps the most concerning is that radio propagation is perfect and circular, and that obstacles have no effect on the propagation of signal over an area. Additionally, there are no widely accepted practices for the validation and verification of wireless network simulations with which a researcher may construct a credible simulation experiment. We survey the wireless coverage of a university campus and confirm from this collected data that walls and objects do indeed impact on signal propagation. With this in mind, we construct several models to best represent wireless prop-agation. The accuracy of these models is then compared to real life observations across a series of trials, testing both indoor and outdoor RF signal propagation. Upon analysing experimental data, we are able to extract several equations representative of RF signal propagation in the presence of walls and other ob-jects. Implementing these equations inside a network simulator and comparing the results to both reality and an ideal distance-based loss model, we conclude with the recommendation that all future wireless network simulations take into account the environment in which the simulations run, as well as the properties of the protocols used.