(A) study on MIMO communication systems in line-of-sight channel environments

Abstract

To provide the high data rates required for fixed wireless applications such as cellular backhaul, we needto increase the channel capacity of line-of-sight (LoS) multiple-input multiple-output (MIMO) communicationsystems and one key solution is to exploit spatial multiplexing gains. In contrast to the conventional view that LoSMIMO channels are rank-deficient, it has been shown that that LoS MIMO systems can achieve a full multiplexinggain through the optimization of antenna placement. Specially, LoS MIMO communication systems with uniformcircular-arrays (UCAs) have an advantage over uniform linear arrays (ULAs) and uniform planar arrays (URAs),that is, its channel matrix can be modeled as a circulant matrix. However, the UCA-based LoS MIMO systemshave not been fully investigated, and their use for practical applications is limited.In the first one-third of this dissertation, we consider the design of UCA system under array misalignmentover LoS channels. UCAs have been considered as useful array structures since their LoS channels can be diagonalized without the channel state information at the transmitter (Tx). However, such a characteristic holds onlywhen transceiver arrays are perfectly aligned. To overcome this difficulty, we propose an optimal design methodand transceiver architectures for UCA systems under array misalignment that achieve channel capacity withoutknowledge of misalignment angles. To this end, we first verify that the singular values of the misaligned UCA system are independent of tilting and center-shift and small relative array rotation, but fluctuate with the Radii Productto Distance Ratio (RPDR). Using the result, we present an optimal design method for UCA systems that performsan one-dimensional search for the RPDR to maximize channel capacity, and we show that a simple zero-forcingreceiver can achieve the maximum channel capacity because of the channel characteristic at the optimal designcriteria. Additionally, we propose a low-complexity precoding scheme for UCA systems in which the optimaldesign criteria cannot be fulfilled. The simulation results demonstrate the validity of the proposed design methodand transceiver architectures.In the second half of this dissertation, we consider the design of transceivers for the LoS MIMO system usingUCA with subarrays. Using a large number of antennas in LoS MIMO system is a burden to implementation as itrequires wide antenna spacings, and expensive radio frequency (RF) chains. To relieve this problem, we propose theuse of UCA with subarrays in hybrid MIMO systems. Specifically, we present the optimal precoder and combinerof UCA-SA system in the case of antenna misplacement. It is shown that the RF beamformer consists of all onevalued beamforming coefficients can be a near optimal at the transmitter, which is effective for implemention ofby the RF phase shifter whose performance is limited by quantization. In addition, the optimal baseband processercan be implemented simple iterative search over a small number of codebook vectors. At the receiver, the optimalbaseband processer is the DFT matrix and only searching RF beamformer is required. The simulation resultsillustrate that the performance of the proposed transceiver is closed to the optimal transceiver.In the last one-third of this dissertation, we consider the design of transceiver arrays for uplink MIMO systemsover strong LoS channels. Spatial multiplexing gain attained by proper antenna spacings can provide high data ratesover LoS channels

however, this has only been considered in the point-to-point MIMO systems. To fully exploit LoS spatial multiplexing gain in uplink MIMO systems, we propose the use of multi-layer UCAs at a receiver and single-layer UCAs at transmitters, and we present design criteria for transceiver UCAs. The simulation resultsdemonstrate the validity of the proposed multi-layer UCA system over strong LoS channel environments.