On Cooperative Transmission Techniques For Wireless Relay Networks

Owing to its capability of increasing coverage, throughput, and transmission reli-ability, cooperative relaying has been one of the most widely explored topics in wirelesscommunications over the past few years. The rationale is that multiple relays collab-orate with each other to forward the source information via a virtual multiple-inputmultiple-output (MIMO) formulation. In this sense, cooperative relaying has paved theway for applying MIMO technologies to single-antenna devices, e.g., most handsets incellular mobile networks and nodes in wireless ad-hoc or sensor networks. For cooper-ative relaying, much interest has been centered on how to develop efficient transmissionschemes at the relays to harvest the advantages of traditional MIMO technologies. Thisthesis investigates three cooperative transmission techniques in wireless relay networks,namely, relay selection, distributed space-time coding, and cooperative beamforming.Specifically, the main contributions of this thesis are summarized as follows:1. For relay selection, our focus is on two-way opportunistic relaying (TWOR)systems in amplify-and-forward (AF) strategy.i) We first present a performance analysis of the equal power allocation (EPA)scheme for the TWOR-AF system over independent and non-identically dis-tributed Rayleigh fading channels. Closed-form lower and upper bounds as wellas an approximation of the system outage probability are established to show theasymptotic performance of the EPA scheme, which are all tight at high signal-to-noise ratio (SNR). The diversity gain, coding gain and diversity-multiplexingtradeoff are obtained from the tight approximation. Moreover, using the momentgenerating function approach, a general symbol-error-rate expression in closed-form is derived for the EPA scheme with Μ-ary phase shift keying (Μ-PSK)or Μ-ary quadrature amplitude modulation (Μ-QAM) signals. The theoreticalanalysis is then validated by simulations.ii) To further improve the performance, transmit power optimization is also studiedfor the TWOR-AF system. With instantaneous channel strength known beforetransmitting, we develop an optimum power allocation (OPA) scheme by solvingan optimization problem of SNR balancing. While with only statistical channelknowledge, we propose a sub-OPA scheme that minimizes the high-SNR approx-imation of the system outage probability under more general Nakagami-m fading channels. Simulation results are presented to show that the proposed power allo-cation schemes significantly outperform the EPA scheme in outage probability.2. For distributed space-time coding, our focus is on decode-and-forward relaynetworks which combine bit-interleaved coded modulation (BICM), distributed orthog-onal space-time block coding (DOSTBC) and orthogonal frequency division multiplex-ing (OFDM) in frequency-selective fading channels.i) By exploiting the inter-subcarrier relationship of OFDM, we propose a pilot-efficient noncoherent decoder for such a BICM-DOSTBC-OFDM system, whichrequires neither channel state information (CSI) nor relay state information (RSI)a priori. Moreover, it can perform decoding within one DOSTBC-OFDM block,thus suitable for relatively fast fading environments where both CSI and RSI varyfrom one block to another.ii) In the performance aspects, we are mainly concerned with the diversity orderachieved by the proposed noncoherent decoder. To this end, we study a classof carefully designed transmission schemes, called perfect channel identifiabil-ity (PCI) achieving schemes. Then we establish a worst-case diversity analysisframework to show that the proposed noncoherent decoder is able to achievethe maximum cooperative-frequency diversity offered by the BICM-DOSTBC-OFDM system, provided that the PCI-achieving transmission schemes are em-ployed. Simulation results are presented to show that the proposed noncoherentapproach exhibits near-coherent bit error performance and significantly surpassesthe pilot-aided channel estimation methods.3. For cooperative beamforming (CB), our focus is on AF relay networks in thepresence of multiple eavesdroppers. The goal is to design efficient CB schemes forblocking eavesdropping attempts.i) Under both the total and individual relay power constraints, we first propose aquasi-optimal CB scheme based on secrecy rate maximization (SRM). To han-dle the non-convex SRM problem, we apply a two-level optimization frameworkconsisting of one-dimensional search and semidefinite relaxation (SDR), whichinvolves solving a sequence of semidefinite programs (SDPs). To reduce the com-plexity of SRM, a null-space-based optimal CB scheme is proposed that maxi-mizes the information rate at the destination while completely eliminating the information leakage to all the eavesdroppers. We prove that this design problemcan be exactly solved by SDR with one SDP only.ii) In view of the fact that artificial noise (AN) has potential in jamming the eaves-droppers effectively, we also design an optimal joint CB-AN scheme. The basicidea is to use the previous two-level optimization framework to solve the AN-aided SRM problem. Unlike the CB-only optimization, we prove that SDR hereis tight due to the use of AN. We further robustify the optimal joint CB-ANscheme for imperfect eavesdroppers’ CSI by extending the two-level optimiza-tion framework to solve a more complicated worst-case (WC)-SRM problem.Simulation results are presented to show the efficacy of the proposed schemes insecrecy rate improvement.