Date of Award
Doctor of Philosophy (PhD)
Electrical Engineering and Computer Science
Gursoy, M. Cenk
Cell-free Massive MIMO, Intelligent Reflecting Surface (IRS), Internet-of-Things (IoT), Massive Connectivity, Wireless Communications
Recent years have witnessed unprecedented growth in wireless networks in terms of both data traffic and number of connected devices. How to support this fast increasing demand for high data traffic and connectivity is a key consideration in the design of future wireless communication systems. With this motivation, in this thesis, we focus on the analysis of wireless networks with massive connectivity.
In the first part of the thesis, we seek to improve the energy efficiency (EE) of single-cell massive multiple-input multiple-output (MIMO) networks with joint antenna selection and user scheduling. We propose a two-step iterative procedure to maximize the EE. In each iteration, bisection search and random selection are used first to determine a subset of antennas with the users selected before, and then identify the EE-optimal subset of users with the selected antennas via cross entropy algorithm. Subsequently, we focus on the joint uplink and downlink EE maximization, under a limitation on the number of available radio frequency (RF) chains. With the Jensen's inequality and the power consumption model, the original problem is converted into a combinatorial optimization problem. Utilizing the learning-based stochastic gradient descent framework and the rare event simulation method, we propose an efficient learning-based stochastic gradient descent algorithm to solve the corresponding combinatorial optimization problem.
In the second part of the thesis, we focus on the joint activity detection and channel estimation in cell-free massive MIMO systems with massive connectivity. At first, we conduct an asymptotic analysis of single measurement vector (SMV) based minimum mean square error (MMSE) estimation in cell-free massive MIMO systems with massive connectivity. We establish a decoupling principle of SMV based MMSE estimation for sparse signal vectors with independent and non-identically distributed (i.n.i.d.) non-zero components. Subsequently, using the decoupling principle, likelihood ratio test and the optimal fusion rule, we obtain detection rules for the activity of users based on the received pilot signals at only one access point (AP), and also based on the cooperation of the received pilot signals from the entire set of APs for centralized and distributed detection. Moreover, we study the achievable uplink rates with zero-forcing (ZF) detector at the central processing unit (CPU) of the cell-free massive MIMO systems.
In the third part, we focus on the performance analysis of intelligent reflecting surface (IRS) assisted wireless networks. Initially, we investigate the MMSE channel estimation for IRS assisted wireless communication systems. Then, we study the sparse activity detection problem in IRS assisted wireless networks. Specifically, employing the generalized approximate message passing (GAMP) algorithm, we obtain the MMSE estimates of the equivalent effective channel coefficients from the base station (BS) to all users, and transform the received pilot signals into additive Gaussian noise corrupted versions of the equivalent effective channel coefficients. Likelihood ratio test is used to acquire decisions on the activity of each user based on the Gaussian noise corrupted equivalent effective channel coefficients, and the optimal fusion rule is used to obtain the final decisions on the activity of all users based on the previous decisions on the activity of each user and the corresponding reliabilities. Finally, we conduct an asymptotic analysis of maximizing the weighted sum rate by joint beamforming and power allocation under transmit power and quality-of-service (QoS) constraints in IRS assisted wireless networks.
Guo, Mangqing, "Analysis of Wireless Networks With Massive Connectivity" (2022). Dissertations - ALL. 1557.