Ad hoc routing protocols like DSR, AODV, and OLSR have been studied under many simulation tools in many literatures recent years. The evaluation results in these literatures are mostly obtained under the wireless channel model with constant coverage and attenuation, and usually studied under particular scenarios. This lead to a consequence that the simulation result isn't consistent with the actual situation and the modified algorithm can't be well applied to the real environment. We evaluate the AODV protocol more accurately and comprehensively based on the following two conditions: firstly, we apply the NAKAGAMI propagation model which can perform proximately to the real wireless channel. Secondly we define different topology densities for the simulation. The results are given in forms of good put, delivery rate and end-to-end delay, which showed great statistical proximity to the real scenario. Our results confirmed that the simulation model used in our simulation is more accurate than others, and can provide important reference for evaluating other routing protocols or application traffics in wireless multi-hop ad hoc networks.
