Mobility models have traditionally been tailored to specific application domains such as human, military, or ad hoc transportation scenarios. This tailored approach often renders a mobility model useless when the application domain changes. Furthermore, the failure to adapt the mobility model to accurately match the new domain naturally leads to wrong conclusions about the performance of protocols and applications running atop. In this paper, we propose a mobility modeling framework based on the observation that the mobility characteristics of most mobility-based applications can be captured in terms of a few fundamental factors: (1) Targets; (2) Obstacles; (3) Dynamic Events; (4) Navigation; (5) Steering behaviors; , and (6) Dynamic Behaviors. We have designed and implemented a Universal Mobility Modeling Framework (UMMF), which enables the instantiation of a mobility model from a wide universe of possibilities defined by the aforementioned factors. We describe the mapping from applic...