Department of Electrical and Computer Engineering
Ph.D. Thesis Abstract
COOPERATVE STRATEGIES FOR WIRELESS RELAY NETWORKS
The goal of this thesis is to understand the mechanisms and potential benefits of relaying and node cooperation in wireless networks. We analyze specific cooperative schemes, present practical cooperative protocols for large networks, and derive capacity results for systems with limited transmitter cooperation.
Motivated by sensor applications, the first part of the thesis considers cooperation in large, energy-constrained networks that have to deliver multicast data. We use the insights offered by network information theory to propose an accumulative broadcast strategy that allows nodes to collect energy of unreliably received overheard signals. As a message is forwarded through the network, nodes will have multiple opportunities to reliably decode the message by collecting energy during each retransmission. We analyze two problems concerned with energy-efficient multicast/broadcast. First, we formulate the minimum-energy accumulative broadcast problem. We show that the problem is NP-complete and propose an energy-efficient heuristic algorithm. We then address the maximum lifetime multicast problem and present an MLAB algorithm that finds the optimum solution. The resulting transmit power levels ensure that the lifetimes of the active relays are the same, causing them to fail simultaneously.
The proposed broadcast scheme employs a decode-and-forward (DF) relay strategy. In general, however, the optimal relay strategy is unknown. In the second part of the thesis, we evaluate the performance of amplify-and-forward (AF) strategies for energy-constrained networks. For a single source-destination pair, we characterize the optimum AF bandwidth and present the optimum power allocation among the AF relays which can be viewed as a form of maximum ratio combining. Motivated by large bandwidth resources we further consider orthogonal signaling at the nodes. While the result for the optimum bandwidth still holds, the relay power solution in this case has the form of water-filling. In contrast, in a network with unconstrained bandwidth, the DF strategy
operates in the wideband regime and requires a different choice of relays. Thus, in a large scale network, the choice of a cooperative strategy goes beyond determining a coding scheme at a node; it also determines the operating bandwidth and the best distribution of the relay power.
In the third part of the thesis, we introduce limited transmitter cooperation to the interference channel with two independent sources and two receivers. Transmitter cooperation enabled by side-channel links with finite capacities allows for a partial message exchange between encoders. After cooperation, each encoder will know a common message partially describing the original messages, and its own private message. We first determine the capacity region of the compound multiple-access channel (MAC) in which both common and private messages are decoded at both receivers. We then relax the decoding constraint and consider the interference channel with common information in which each private message is decoded only by one receiver. We determine the stronginterference conditions under which the capacity region of this channel is found to coincide with the compound MAC capacity region. Finally, we determine the strong interference conditions for the interference channel with unidirectional cooperation in which messages sent at one encoder are known to the other encoder, but not vice versa.
Ph.D. Dissertation Director: Professor Roy D. Yates
Download the Thesis: PDF