N-Body: Social Based Mobility Model for Wireless Ad Hoc Network Research

Accurate reproduction of real human movement patterns is necessary in simulations of mobile ad hoc networks in order to obtain meaningful performance results. Human activities are often socially organized, resulting in a certain level of tendency of forming groups. There exists a few mobility models that are taking this tendency into account; however, all these models require a certain level of understanding of the underlying social structure of the target scenario, which limits their application scope. In this paper we propose an N-body mobility model that tackles such social aspects from a different perspective. We extract the social information from real human movement traces, and reproduce them in the mobility model. We show that the N-body model is capable of capturing and synthesizing the group-forming tendency that matches to those observed from sample traces. Simulation results also show that N-Body models exhibit a similar pairwise heterogeneity in ad-hoc network performance as the sample traces do.