Concatenated space-time coding for large antenna arrays

Wireless communications system performance under fading conditions can be greatly improved when multiple antennas are used at the transmitter and/or at the receiver. Information-theoretic results indicate that, under certain fading statistics, the available data rates are proportional to the antenna array sizes at both ends of the link. This dissertation addresses space-time processing approaches that can utilize the high data rates while achieving the benefits of the maximal diversity order. Information-theoretic arguments and code performance criteria are used to develop methods that can be practically implemented even for large transmitter antenna arrays.; For systems where the channel state information is available at the transmitter, a preprocessing method is presented that maximizes the mutual information by adjusting the covariance of transmitted vectors. The result is improved bandwidth and/or energy efficiency when the number of transmitters exceeds the number of receivers. The effect of correlated and non-Rayleigh fading on achievable rates and effective SNR is investigated.; When the channel parameters are not known at the transmitter, space-time coding can be used to maximize the effective number of degrees of freedom against fading. Certain structural features of concatenated codes can naturally lead to satisfying the necessary conditions for maximal diversity, making them good candidates for space-time coding applications. A serially concatenated space-time encoder for block transmission is presented along with block length design rules that allow maximal diversity benefits to be obtained.; The concatenated encoder is complemented by an iterative decoding architecture where the modulator and channel are treated as an additional encoder in serial concatenation, offering both performance improvement and a natural opportunity to reduce decoding complexity. It is shown that the performance of full-complexity decoding can be very closely approximated using a subset algorithm with only quadratic complexity. The proposed space-time coding method is a flexible processing scheme where the transmission rate can be scaled proportionally to the array size, without requiring a manual re-design of the system, achieving maximal diversity benefits even for very large transmitter arrays.