Mobility Patterns in Macrocellular Wireless Networks

Abstract

Realistic user models are indispensable for performance evaluation of mobility prediction algorithms. The macrocellular wireless service user population comprises users with diverse mobility characteristics. This paper investigates mobility patterns in macrocellular wireless networks, based on empirical data gathered from several users. Based on the observed statistics a user classification based on mobility is achieved. Further the paper characterizes the distribution of cell residence time (CRT), which is the length of time that a user spends in a cell, before moving into the service area of another cell. Studies reported in literature concentrate on cell residence time distribution of mobile terminals in-session (dedicated channel allocated) and ignore their out-of-session (idle mode) characteristics, which critically influence several network management tasks. Investigation shows that a user's out-of-session CRT distribution can be accurately modeled using heavy-tailed arithmetic distribution with infinite mean and variance, contrary to the assumptions made in the literature