Cross-layer optimization and cooperative communications in wireless networks.

摘要

This dissertation studies cross-layer optimization and cooperative communications in wireless networks. It consists of three main thrusts.;In the first thrust, a cross-layer design approach is taken to study rate control in multi-hop wireless networks. Due to the lossy nature of wireless links, the data rate received successfully at the destination node is typically lower than the transmission rate at the source node. In light of this observation, each flow is treated as a "leaky-pipe" flow and the notion of "effective utility" is introduced. Then the rate control is explored through effective network utility maximization (ENUM). The "thinning" feature of data flows is explicitly taken into account and distributed hop-by-hop rate control algorithms are developed accordingly.;The second thrust focuses on efficient data transport in wireless sensor networks, where neighboring sensor nodes are organized into coalitions and cooperative communications can be carried out. In particular, three network models are considered: (1) many-to-one sensor networks where data from one coalition are transmitted to the sink directly, (2) multi-hop sensor networks where data are relayed by intermediate nodes to reach the sink, and (3) sensor networks with data Mules which move around to cover the area and collect data from static sensor nodes. Energy efficiency, balancing, and probabilistic coverage and connectivity are investigated in this study.;The third thrust is on the comparison between MIMO and SISO configurations in terms of optimal neighbor selection in multi-channel sensor networks. For each antenna configuration, two neighbor selection schemes are studied. In the scheme of first forward neighbor selection (FFNS), neighbor probing stops as soon as a neighbor with forward progress is discovered. In the scheme of optimal stopping neighbor selection (OSNS), optimal stopping theory is employed to decide whether the neighbor probing should stop at each stage. It can be shown that OSNS always achieves better progress rate than FFNS. The configuration of one MIMO radio achieves better progress rate when the node density is low. However, when the node density is large, the configuration of multiple SISO radios actually achieves better progress rate.