Distributed Space-Time Block Codes Achieving Optimal Diversity Function with Linear Receiver

Abstract

<p>The design and analyses of Space-Time Block Codes (STBC) for both single antenna and two-antenna distributed relay channels are considered in this thesis. Due to the fact that the equivalent channel gains for two-phase relay channels are the product of two channel coefficients, many current STBC designs for Multiple Input Multiple Output (MIMO) channels cannot be implemented to distributed relay channels efficiently. The direct application of Orthogonal Space-Time Block Coding (OSTBC) for MIMO systems to distributed cooperative relay networks makes the equivalent channel matrix for maximum likelihood (ML) detection lose its orthogonality. Hence, a new design that makes the channel matrix be \textit{orthogonally distributed} (OD) for a suboptimal symbol-by-symbol detector (SBSD) is proposed in this thesis. With ODSTBC, an asymptotic symbol error probability (SEP) formula with SBSD is derived, showing the optimal diversity gain function for single antenna distributed relay channels $\frac{\ln^N\rho}{\rho^N}$ is achieved. In addition, two ODSTBC designs for the distributed relay networks are presented, which interestingly renders that SBSD is equivalent to the ML detector. The ODSTBC enjoys both optimal diversity function and low detection complexity. However, the symbol rate of ODSTBC is relatively low in order to maintain the orthogonal conditions. To address this problem, another Alamouti Based Toeplitz Space-Time Block Code (ABTSTBC) for two-antenna distributed relay channels is proposed. Both the code structure and the equivalent channel matrix has a block Toeplitz structure, whose blocks are the addition and product of two Alamouti matrices, respectively. With the linear SBSD, the optimal diversity function $\frac{\ln^N\rho}{\rho^{2N}}$ is achieved. At the same time, the advantages of low computational complexity and high symbol rate are maintained. Numerical results verify the diversity analyses and indicate competitive error performance to currently available distributed STBC designs with much lower complexity.</p>