User-Centric Mobility Models for Opportunistic Networking

In this chapter we survey the most recent proposals for modelling user mobility in mobile pervasive networks, and specifically in opportunistic networks. We identify two main families of models that have been proposed. The first modelling approach is based on the observation that people tend to visit specific places in the physical space, which therefore exert special attraction on them. The mechanics of user movements are defined based on these attractions. The second approach is based on the fact that people are social beings, and therefore they move because they want to interact and meet with each other. Movements are thus defined based on the social relationships established by users among themselves. Both modelling approaches show good match with popular traces available in the literature. However, we note that each approach misses the other's point: people actually move both because they are attracted by other people, and because they spend time in preferred physical places. Therefore, we describe a new mobility model (Home-cell Community-based Mobility Model, HCMM) that takes both properties into account, i.e., social relationships and attraction of physical places. HCMM matches well-known statistical features of real human mobility traces. Furthermore, it provides intuitive and easy-to-use knobs to control overall system statistical properties generated by users' movements (e.g., the average time spent by users inside or outside preferred places).