UC Irvine

Massive multiple-input multiple-output (MIMO) technology employs arrays with a large number of antennas, of the order of 100 or more, at the base station (BS) to meet the data rate and user demands of next-generation wireless systems. To cope with the power consumption problem due to an increased number of receive antennas, the idea of equipping one-bit analog-to-digital converters (ADCs) at the base station has been proposed. This thesis will focus on the topic of channel estimation which is key to exploiting the potential gains of massive MIMO. In this first part of the thesis, performance bounds on the channel estimation of one-bit millimeter-wave (mmWave) massive MIMO receivers for different types of channel models are established. The Cramer-Rao bound (CRB), which sets a benchmark for the design of channel estimators, is considered for both a structured channel model for a single user where the channel is composed of a superposition of multipaths characterized by path delays and directions-of-arrival (DOAs), and an unstructured channel model where the channel is a generic FIR filter. The Bayesian CRB when the array response is imperfectly known and is affected by perturbations in the sensor pattern or position is also derived. The CRBs are evaluated numerically and the effects of various system parameters on the CRB are studied. The results show that increasing the bandwidth or the oversampling factor decreases the estimation error variance due to improved temporal resolution. Spatial oversampling could also be used, instead of, or in addition to temporal oversampling. In the second part of this dissertation, spatial Sigma-Delta architectures, to shape the quantization noise away from users in some angular sector, are considered. A linear minimum mean squared error (LMMSE) channel estimator based on the so-called Bussgang decomposition is developed and the uplink achievable rate with linear receivers is analyzed. Finally, the problem of direction finding when the BS is equipped with a rectangular antenna array and spatial Sigma-Delta ADCs is considered. The impact of array response and noise modeling errors on the estimation errors of Bartlett beamformers and the multiple signal classification (MUSIC) algorithm is studied.