An important challenge in the design of wireless and mobile systems is that two key resources---communication bandwidth and energy---are significantly more limited than in a tethered network environment. In addition, the time-varying characteristics of wireless channels make it hard to consistently obtain the same performance. These restrictions require innovative communication, networking, and design techniques for the efficient utilization of the bandwidth and energy. One of the most rapidly developing areas in wireless networks is wireless ad hoc networks. A wireless ad hoc network is an autonomous system consisting of nodes, which may or may not be mobile, connected with wireless links and without using pre-existing communication infrastructure or central control. Ad hoc networking is expected to play an important role in future wireless mobile networks due to the widespread use of mobile and hand-held devices. Mobile Ad hoc Networks (MANETs) and Wireless Sensor Networks (WSNs) are two prominent classes of these infrastructerless wireless networks. While MANETs exhibit dynamic topology changes due to free node mobility, WSNs have unreplinishible energy limitations. Hence, topology control, Quality of Service (QoS) routing, and power control become challenging issues. We argue that cluster-based techniques coupled with cross-layer design can achieve better performance in this harsh environment. This dissertation supports this claim by introducing strategies for topology control, QoS routing in MANETs, and energy efficient routing in WSNs. First, we develop the Virtual Grid Architecture (VGA), which is a fixed and stable architecture for ad hoc networks that can support efficient routing and network control. We show that although VGA clustering is simple, it is close to optimal. Then, we develop two QoS routing protocols that combine the ideas of cluster-based routing and cross layer design to achieve good performance in terms of delay, bandwidth, and user-perceived quality. These protocols, operating on top of VGA, show improved system call success rate and packet delivery ratio by an order of magnitude compared to general-purpose approaches. Finally, we develop GRASP (Grid-based Routing and Aggregator Selection Protocols), a scheme for WSNs, that combines the ideas of fixed cluster-based routing of VGA together with application-specific data aggregation functions. GRASP is able to enhance the network performance in terms of extending the network lifetime, while incurring acceptable levels of latency in data aggregation. Our studies together show that creating stable and scalable architecture can achieve topology robustness, enhance quality of service, and attain the energy and latency efficiency needed for wireless networks.
展开▼
