A new mobilty model based on maps

Abstract

In this paper we present a mobility model for simulations of wireless access scenarios. This new model has been developed to reduce the shortcomings of previous models in capturing higher-order statistical dependencies in user behavior. The proposed model strikes a balance between the abstraction level of some purely stochastic path generation models that lack certain important properties of the real world and completely deterministic models that fail to produce answers to inherently stochastic problems such as call admission control, hand-off/handover or prefetching of data. The proposed model uses freely definable building maps and radio cell coverage information and generates sensible paths between randomly selected way-points. The paths are generated by a combined diffusion/steepest gradient algorithm. Various models for user speeds can be incorporated. The resulting patterns of user mobility and temporal radio coverage are well applicable in driving simulations with the purpose of addressing key questions in mobile wireless scenarios.