An area-scalable human-based mobility model

Mobile networks have been growing rapidly, where it is of paramount importance to evaluate the performance of any protocol designed for such networks before deployment. Mobility is one of the main mobile network issues that directly affects performance metrics. Since mobile nodes are often carried by their owners, nodes movement should be modeled according to humans movement. In this paper, we extend our previously published human mobility model based on real collected data. The real human movement is traced, cleaned up and analyzed. To the best of our knowledge, our collected data set is the only one that contains the detail of both walking and driving nodes. The analysis shows several unique characteristics some of which are absent in former human mobility models. The new model is proposed based on a graph that is extracted from geographical constraints. In our model, the spatiotemporal variation of movement is also taken into account. To evaluate the precision of our model, we make a comparison with our own collected data and also the data of former works, the results of which show close matching. Furthermore, exhaustive simulation is done using ns-2 network simulator. Network parameters are also evaluated using different mobility models.