Cell-free massive multiple antenna systems: estimation techniques, scalability challenges, and aerial access point deployments

Abstract

Cell-free (CF) massive multiple-input multiple-output (MIMO) networks are envisioned as promising technologies for next-generation wireless networks. These networks stand out for their high spectral efficiency, lower susceptibility to blocking and shadowing, and uniform performance among users. As the CF massive MIMO scenario spreads a large number of access points (APs) over a large area serving a much smaller number of user equipment (UE) devices, problems related to power control, signal precoding, resources, decoding strategies, and interference management will be key issues. Furthermore, to achieve a scalable and efficient CF massive MIMO system, a user-centric (UC) approach is used to define the subset of APs serving each UE device, such that the complexity and resource requirements of each AP must remain finite when the number of UE devices is infinite. In this regard, this doctoral dissertation aims to present scalable methodologies suitable for efficiently implementing UC CF massive MIMO networks. Specifically, radio resource allocation, pilot assignment methods, downlink (DL) pilot based, and blind effective channel estimation schemes are developed. Furthermore, a dynamic strategy is proposed to choose the best channel estimation method for each UE device. The results showthat the proposed solutions can improve the spectral efficiency of the 50%-likely UE devices by up to 91% compared with the solutions presented in the literature. This doctoral dissertation by also investigates how to implement a massive UC CF network MIMO enabled by aerial APs, which are useful to improve wireless connectivity in ultradense environments or flexibly extend coverage in more isolated areas. For this purpose, a trajectory optimization of a swarm of cooperative unmanned aerial vehicles (UAVs) is proposed. The results show that it is possible to optimize UAV deployment and trajectory without knowledge of the UE locations, reaching up to a 47%increase in average spectral efficiency. Finally, this doctoral dissertation investigates UC CF massive MIMO systems with a limited processing capability. More specifically, AP cluster adjustment methods are proposed to guarantee that the computational complexity (CC) of signal processing does not increase with the number of APs. The results show that it is possible to reduce the CC by up to 96% while maintaining spectral efficiency with minimal degradation.