Düşük çözünürlüklü nicemleyicilere sahip kitlesel çok-girdili çok-çıktılı haberleşme sistemleri

Low resolution analog-to-digital converters (ADC) attracted much attention for their use in massive multiple-input multiple-output (MIMO) systems due to their low power consumption and cost. In this thesis, we question whether large number of antennas present in massive MIMO is sufficient to provide an ultimate performance or additional sampling in time (temporal oversampling) will provide significant performance advantages. To begin with, we illustrate the benefits of oversampling in time for uplink massive MIMO systems with low-resolution ADCs in terms of symbol error rate (SER) and achievable rate for both single-carrier (SC) and multi-carrier modulation scenarios by deriving analytical bounds and with simulations. We also propose a sequantial linear minimum-mean-square error (LMMSE) based receiver as a low-complexity detector, which is much more feasible to implement compared to the zero-forcing (ZF) detector for temporally oversampled massive MIMO systems. We also examine and illustrate the benefits of temporal oversampling for quantized massive MIMO systems under adjacent channel interference caused by the non-linearity of the quantizers. According to the results that we obtain, it seems that temporal oversampling can be very beneficial and should always be considered for use in quantized massive MIMO. Finally, we examine whether a low-complexity MIMO detector, which can outperform the existing detectors of similar complexity, can be proposed even without resorting temporal oversampling. For that purpose, we propose a near optimal factor-graph based Ungerboeck type detector with bi-directional decision feedback, along with the derivation of LMMSE channel estimator for quantized wideband SC-MIMO systems. The proposed detector is shown to outperform a representative detector of comparable complexity from the literature in terms of SER and achievable rate per user performance metrics.