Understanding user mobility is important in designing location-aware systems and wireless networks, and for simulations of mobile devices in a wireless network. I will start this presentation by providing a better understanding of user mobility. For example, we analyzed periodicities in real wireless network traces and found that 24% of users move in random patterns while the other users exhibit periodic patterns.
I will then describe our experiences in extracting user mobility characteristics from wireless network traces, and developing a mobility model based on these characteristics. I will present a method to extract the physical path of users from the sequence of access points recorded in network traces. Using this method, we were able to analyze mobility characteristics; we discovered that the distributions of speed and pause time each follow a log-normal distribution and that the direction of movements closely reflects the direction of roads and walkways. I will finally describe a mobility model that focuses on user movements among popular regions. I will conclude the talk with a discussion of remaining challenges in mobility modeling.
Colloquia Series page.