Adaptive Communication for Wireless Massive MIMO Systems
Abstract
The demand for high data rates in wireless communications is increasing rapidly. One way to provide reliable communication with increased rates is massive multipleinput multiple-output (MIMO) systems where a large number of antennas is deployed. Weanalyze three systems utilizing a large number of antennas to provide enhancement in the performance of wireless communications. First, we consider a general form of spatial modulation (SM) systems where the number of transmitted data streams is allowed to vary and we refer to it as generalized spatial modulation with multiplexing (GSMM). A Gaussian mixture model (GMM) is shown to accurately model the transmitted spatially modulated signal using a precoding framework. Using this transmit model, a general closed-form expression for the achievable rate when operating over Rayleigh fading channels is evaluated along with a tight upper and a lower bounds for the achievable rate. The obtained expressions are flexible enough to accommodate any form of SM by adjusting the precoding set. Followed by that, we study quantized distributed wireless relay networks where a relay consisting of many geographically dispersed nodes is facilitating communication between unconnected users. Due to bandwidth constraints, distributed relay networks perform quantization at the relay nodes, and hence they are referred to as quantized distributed relay networks. In such systems, users transmit their data simultaneously to the relay nodes through the uplink channel that quantize their observed signals independently to a few bits and broadcast these bits to the users through the downlink channel. We develop algorithms that can be employed by the users to estimate the uplink channels between all users and all relay nodes when the relay nodes are performing simple sign quantization. This setup is very useful in either extending coverage to unconnected regions or replacing the existing wireless infrastructure in case of disasters. Using the uplink channel estimates, we propose multiple decoders that can be deployed at the receiver side. We also study the performance of each of these decoders under different system assumptions. A different quantization framework is also proposed for quantized distributed relay networking where the relay nodes perform vector quantization instead of sign quantization. Applying vector quantization at the relay nodes enables us to propose an algorithm that allocates quantization resources efficiently among the relay nodes inside the relay network. We also study the beamforming design at the users’ side in this case where beamforming design is not trivial due to the quantization that occurs at the relay network. Finally, we study a different setup of distributed communication systems called cell-free massive MIMO. In cell-free massive MIMO, regular cellular communication is replaced by multiple access points (APs) that are placed randomly over the coverage area. All users in the coverage area are sharing time and frequency resources and all APs are serving all UEs while power allocation is done in a central processor that is connected to the APs through a high speed backhaul network. We study the power allocation in cell-free massive MIMO system where APs are equipped with few antennas and how the distribution of the available antennas among access points affects both the performance and the infrastructure cost.
Degree
M.S.E.C.E.
Advisors
Love, Purdue University.
Subject Area
Information Technology|Engineering
Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server.