
Linear diversity precoding design criterion for finite block‐fading parallel MIMO channels
Author(s) -
Sykora Jan,
Knize Milan
Publication year - 2011
Publication title -
wireless communications and mobile computing
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.42
H-Index - 64
eISSN - 1530-8677
pISSN - 1530-8669
DOI - 10.1002/wcm.943
Subject(s) - precoding , fading , codebook , computer science , interleaving , mimo , algorithm , encoder , gaussian , independent and identically distributed random variables , block code , channel (broadcasting) , topology (electrical circuits) , mathematical optimization , mathematics , decoding methods , telecommunications , random variable , statistics , physics , quantum mechanics , combinatorics , operating system
A novel optimal two stage coding for finite set of parallel flat‐fading MIMO channels with single common information source with specific constant rate requirement is derived. The optimality of suggested coding is achieved in terms of the capacity versus outage performance. The well‐known optimal coding rule relies on Gaussian codewords spanned over the whole available finite set of parallel channels. We prove that the equivalent preprocessing to the ideal interleaving is to re‐code independent parallel channels codewords by a linear inner precoder from a special class of unitary precoders complying with the optimality criterion derived in the paper. Performing such linear mixture of codewords sharing common Gaussian block‐wise codebook, the same capacity versus the outage is guaranteed without any interleaving over parallel channels. We utilize a virtual multiple access (VMA) channel approach to derive the optimality criterion. Selected precoders with various space‐time or time‐only domain span were tested against this criterion and we provide the optimality results on variety of the channel parameters. We showed that the temporal processing is the most important one to achieve the optimality of the precoder. A full space‐time precoding does not perform better than one which is temporal‐only. Copyright © 2010 John Wiley & Sons, Ltd.