Relay-assisted communication : fundamental limits and selection strategies

Abstract

Wireless communication continues to make a profound impact upon our daily lives. The oft-touted benefits of high data rates and improved reliability via wireless communication are limited by its inherent drawbacks, including path loss, fading and interference. One promising strategy for overcoming these problems is to deploy nodes in the region between a transmitter and its intended receiver. These intermediate nodes can improve communication for this transmitter-receiver pair by receiving a transmitted message, processing it and relaying the processed output to the receiver. This transmission strategy, known as relay-assisted communication, can be especially beneficial when the transmitter-receiver pair are either separated by a large distance or when a large obstruction blocks the path between them. In a reasonably dense network, several relays may be available to assist a particular transmitter-receiver pair. Deciding which relays should forward the transmitted message is actually quite difficult. For example, the relay with the best physical-layer channel gain to the destination may also be running low on battery power. Another relay may have a good physical-layer channel gain to the destination and a reasonable amount of remaining battery power, but its queue may be full of messages from other transmitters, so it cannot forward a newly arrived message within a given delay constraint. Thus, optimal relay selection entails carefully balancing all system parameters, which is prohibitively complex in current wireless systems. This dissertation provides novel results for dealing with the relay selection problem in two distinct types of wireless systems. First, several selection algorithms are designed for single-antenna wireless networks, including a decentralized random access-based strategy and centralized methods that are based on throughput maximization and downlink user scheduling. Second, selection algorithms based on transmission hop length are designed for multipleantenna wireless networks. The presented strategies for both single-antenna and multiple-antenna relaying are highly intuitive, as they allow for concise descriptions, making them amenable to practical implementation. Also, the presented strategies illustrate the importance of application-specific design, since each of them yields good performance by focusing on a small set of system parameters. For example, observed latency is of paramount importance for wireless networks that support a significant level of video traffic.