Open-loop and closed-loop transmit diversity designs for multiple antenna wireless systems
This thesis focuses on the design and analysis of transmitting schemes for multiple antenna wireless systems of different system settings: open-loop; closed-loop, single-user, and multi-user. For single-user, open-loop multiple input multiple output (MIMO) links, we propose a full diversity transmitting scheme based on circulant structures. The proposed scheme has no rate loss and is applicable to any number of transmit antennas. Compared with other full diversity, rate one schemes, our scheme has very simple encoding structure. For single-user, closed-loop systems, we give a low complexity adaptive fill diversity, full rate (FDFR) design. With channel subspace knowledge at the transmitter, we adapt the open-loop FDFR codes to maintain the layer structure of the code via transmission. That layer structure enables us to reduce the detection complexity by decoupling joint detection into separate detections of lower dimensions. Adaptive power loading among the spatial subchannels is also derived. For multi-user MIMO broadcast channels (downlinks), we focus on characterizing the system throughput under partial channel knowledge at the transmitter. Three different models of partial channel state information (CSI) are considered: (i) the channel subspace feedback model where the normalized channel vector of each user is available at the basestation, (ii) the limited feedback model where the basestation has knowledge of quantized CSI for each user, and (iii) the imperfect CSI model in which the CSI is degraded by channel estimation error. We show that the channel subspace feedback is asymptotically optimal in the sum rate sense for high signal-to-noise ratio (SNR), while the limit feedback results a ceiling effect on the sum rate for a fixed feedback rate. For the third case, the asymptotic sum rate loss increases logarithmically with the SNR under fixed variance assumption of the estimation error. But for a system with orthogonal training sequences and a minimum mean square error (MMSE) channel estimators, the asymptotical sum rate loss (in bits per channel use) is just the number of transmit antennas.
Love, Purdue University.
Off-Campus Purdue Users:
To access this dissertation, please log in to our