| Mobility clearly is an integral issue in mobile wireless networks. However, little research has focused on realistic mobility. Unrealistic mobility models, such as the popular Random Waypoint or Random Walk models, may result in incorrect conclusions regarding the performance evaluation of mobile wireless network protocols.; This thesis introduces a mobility model and simulator that attempt to generate realistic mobility for mobile wireless network simulation in an urban setting. The mobility models are based on data collected in a number of surveys. The result is a 3-layer hierarchical model. The highest layer is an activity model that determines the high level activity (e.g., working). The second layer is a task model that determines the specific task within an activity (e.g., meeting with three people). The third layer is an agent model that determines how mobile nodes move from one location to another (e.g., how a node navigates down a crowded hallway).; A mobility simulator implements the mobility model and is based on the technique of discrete event simulation with rescheduling. This simulator is the first realistic simulator for mobile wireless networks. It is hoped that this simulator will greatly increase the fidelity of mobile wireless network performance evaluation. Further, it is hoped that this work is usher in a new investigation of urban mobile wireless networks. |
