Constellation rotation based enhanced space time block code

Abstract

In a multiple-input multiple-output (MIMO) channel, created by employing multiple antennas at the transmitter and receiver, the system capacity and diversity dramatically increase. In a view of diversity, the space time block code (STBC) was proposed to enhance diversity gain. On the other hand, the bell labs layered space time architecture (BLAST) can achieve full data rate. To achieve both advantages, full-diversity full-rate (FDFR)code was introduced. However, such schemes of open-loop MIMO system, show some problems which are system complexity and performance degradation. To solve those, closed-loop MIMO system is introduced, in which channel state information (CSI) is available at transmitter as well as the receiver. In such systems, several schemes are proposed, which are antenna shuffling, antenna selection, beamforming, closed-loop FDFR code and so on. In this thesis, we consider open-loop and closed-loop FDFR code, which is constructed based on linear constellation rotation. The rotation matrix has independent signal experience all channel coefficients, that increases diversity gain. Even though open-loop FDFR code can achieve both high data rate and good reliability, it needs signicantly increased decoding complexity. Especially, in MIMO system with large number of transmit antennas, the decoding complexity is more critical issue than increasing diversity gain. Thus, we propose modified STBC which can reduce decoding complexity at the expense of diversity loss. It is possible that we design code matrix with $2\It{N_{t}}$ symbols, while conventional FDFR code consists of $\It{N^{2}_{t}}$ symbols, where $\It{N_{t}}$ is the number of transmit antenna. Since the proposed STBC can not achieve full diversity, it shows performance loss. However, its effect is slight and decoding complexity can be reduced efficiently at sufficient diversity order. In another view, to mitigate the decoding complexity, low-complexity FDFR code was designed for closed-loop ...