Throughput optimization and energy efficiency in wireless ad hoc and sensor networks.

摘要

Wireless networks have drawn a lot of attentions due to their rapid devolvement and their potential applications in various areas. In addition to major benefits of wireless networks, there are many challenges that remain in enabling this technology to have more utility for society. With omni-directional antennas, signals from transmission nodes will be received by nearby nodes. This will cause interference and thus will dramatically reduce the throughput in the wireless network if routing and scheduling are not carefully designed. Therefore, it is crucial to avoid or reduce interference and increase throughput in wireless networks. Additionally, nodes in wireless networks, especially sensors, are often powered by batteries. This necessitates the development and improvement of energy efficient protocols and battery life. Moreover, a number of other challenging issues exist in wireless networks, such as compatibility issues due to different standards, security issues in wireless networks due to vulnerable and unprotected transmission media - air, and so on.;In this thesis, we will study the asymptotic upper and lower bounds of broadcast and/or multicast with a fixed data rate channel model and Gaussian channel model for multi-hop wireless networks when nodes are randomly deployed in Chapter 3. Our effort was the first to derive the analytical upper and lower bounds under a physical interference model. These studies investigate the fundamental performances of random wireless networks. We then study data collection and data aggregation in sensor networks in Chapter 4 and Chapter 5. We design energy efficient protocols for low-data-rate WSNs, where sensors consume different energy in different radio states (transmitting, receiving, listening, sleeping, and being idle) and consume energy for state-transition. When general energy cost model is used, we study the problem of designing energy-efficient data aggregation algorithm. To the best of our knowledge, no algorithms with theoretical performance guarantees are known for this problem in the literature.