Code deisgn for SISO and MIMO block-fading channels.

摘要

We study, analyze, and design communication systems for data transmission over block-fading channels. The block-fading channel is a model for communication under slowly-varying fading; where a codeword spans a few independent fading blocks. Code design strategies for block-fading channels are quite different from those for classical additive white Gaussian noise (AWGN) channels or fully interleaved fading channels.;From the expression for pairwise error probability (PEP) bound under maximum likelihood (ML) decoding, we can extract two major parameters for code design on block-fading channels; the diversity order and the coding gain. At high signal to noise ratios, the diversity order determines the slope of the codeword error probability curve, while the coding gain shifts the curve horizontally. Therefore, the diversity order is the determining factor in code design. The optimal diversity order achievable by coding scheme is upper bounded by the Singleton bound, which establishes the fundamental tradeoff between coding rate and diversity order. The family of codes which can achieve the optimal diversity order are referred as blockwise maximum distance separable (MDS) codes. The general approach for code construction on block-fading channels is to design MDS codes with large coding gain.;For single-input single-output (SISO) communication, we propose a blockwise convolutionally encoded bit-interleaved coded modulation (BC-BICM) scheme, which achieves optimal diversity order. In addition, the coding gain can be improved either by choosing a carefully designed signal labeling scheme for the BICM with iterative decoding or by using convolutional codes with longer constraint lengths. We also investigate quasi-cyclic low-density parity-check (QC-LDPC) codes for block-fading channels. With careful design, the proposed QC-LDPC codes exhibit the same good performance as their corresponding random root-LDPC codes. Moreover, the structure of the proposed QC-LDPC codes makes them suitable for efficient encoding.;For the multiple-input multiple-output (MIMO) situation, the system design should take advantage of additional space diversity provided by multiple antennas besides the time diversity. We first study a turbo coded BICM scheme with iterative detection. The signal processing unit at the receiver employs sphere detector to achieve good performance with reduced computation complexity. To obtain the full diversity of the channel, we propose a coded space-time scheme based on modulation diversity, where the channel coding exploits the time diversity and space-time coding provides space diversity. It is shown that the proposed system achieves high throughput by transmitting at full spatial multiplexing.