Wireless Ad Hoc Networks: Bridging the Gap between Theory and Practice

Wireless ad hoc networks are highly appealing for many reasons. They can be rapidly deployed, reconfigured and tailored to specific applications. They are also highly robust due to the distributed nature. These features have enabled a plethora of studies on wireless ad hoc networking in recent years.To facilitate a comprehensive understanding of the performance limit of wireless ad hoc networks, a great deal of effort has been made in investigating the network capacity from the information-theoretical perspective. Despite various traffic scaling laws revealed in those studies, delay analysis, which is a fundamental issue in data networking, remains untouched. In addition, a central controller is usually required to perform the perfect scheduling to approach the proposed scaling laws. Few insights can be shed to the practical protocol design for wireless ad hoc networks, where all the network protocols must operate in a distributed manner.This project aims to bridge the gap between theory and practice in wireless ad hoc networks. In particular, to capture the essence of traffic and the distributed nature of ad hoc networking, a statistical network model will be established for wireless ad hoc networks. Both the traffic statistics and the fundamental characteristics of wireless channels will be incorporated into the proposed network model. The major tasks of the project include:Analyzing the network throughput and multi-hop delay performance based on the statistical network model. The effect of key factors, such as traffic pattern, node distribution, channel statistics and transmission power, on the network performance will be investigated;Examining the traffic scaling laws of wireless ad hoc networks and formulating the optimal rate allocation problem with the constraints of network requirements such as QoS and fairness performance;Developing distributed MAC and routing algorithms based on the above theoretical analysis and evaluating the performance of the proposed algorithms in practical environments. Extensive performance comparisons with the protocols applied in current wireless ad hoc networks will be conducted to illustrate the performance gains.