Random walk modeling of mobility in wireless networks

The evaluation of quality of service (QoS) parameters in wireless cellular networks is different from that in fixed networks due to user mobility and the user confinement to cells within the coverage area. Simulation models and simplified analytical models approximating the source mobility in a cell or a multiplicity of cells have largely served as the means to study the performance of mobile cellular systems. Although these efforts may have been sufficient for the design and architecture of the first and early second generation mobile networks, more exact analytical models are needed to guide the design and assist in evaluation of architectural alternatives and selection of parameters for the next generation of mobile computing and communications networks. In this paper, we introduce more versatile models based on random walk which can capture the effect of source mobility. We develop mobility models by a two-dimensional random walk and the underlying Markov chain property of this random walk to derive the CDF for dwell time. We present preliminary results for the cases of rectangle and circular cell shapes. Our approach in this paper provides modeling mainly for distance-based criterion of boundary crossing. However, the models can be extended to take into account the radio link propagation effects