Location-aided Beam Alignment and Handover in Millimeter-wave Networks
Time: Thu 2023-05-25 14.00
Location: F3, Lindstedtsvägen 26 & 28, Stockholm
Video link: https://kth-se.zoom.us/j/63271456810
Language: English
Doctoral student: Sara Khosravi , Kommunikationssystem, CoS
Opponent: Professor Sofie Pollin, KU Leuven, Leuven, Belgium
Supervisor: Marina Petrova, Kommunikationssystem, CoS; Jens Zander, Kommunikationssystem, CoS
QC 20230504
Abstract
Ever-increasing data rate demands in mobile networks and the spectrum scarcity at the microwave bands have resulted in the exploration of millimeter-wave (mmWave) frequencies for the next generation of wireless networks. While mmWave frequencies offer large bandwidth, communication at these frequencies is not straightforward due to the challenging propagation characteristics. One approach to overcome the propagation challenge is the use of directional communication with narrow beams at the base station and the user to prepare a link with sufficient received power. Hence, communications with narrow beams pose a new challenge in link establishment and channel estimation based on fine angular scanning. Current mmWave systems apply analog phase arrays that can scan one angle at a time, resulting in high latency and overhead during link establishment. Moreover, mmWave links are sensitive to blockages that lead to the high probability of beam misalignment and the frequent updating of beam scanning, especially in mobile scenarios. Hence, it is desirable to design a low overhead beam selection by exploiting the unique properties of the mmWave channel. Furthermore, to provide adequate coverage and capacity, the density of the base stations in mmWave networks is usually higher than the conventional sub-6 GHz network. This leads to frequent handovers and establishing and maintaining the mmWave links more challenging. Therefore, fast base station discovery (finding the target base station in the handover process), and efficient handover execution techniques, will be required to use the full promise of mmWave cellular networks.
Motivated by the mentioned challenges, this thesis considers the beam alignment and handover problems. Specifically, in the first thread of the thesis, inspired by the unique properties of the spatial channel response of mmWave links, we propose a location-aided beam alignment method based on dividing the user trajectory into regions and storing a set of candidate beams for each region in mobile mmWave networks. Our analysis reveals that our proposed method can achieve a high signal-to-noise ratio, and low overhead while being more robust to the location information accuracy compared with some location-aided baselines in the literature.
In the second thread of the thesis, we focus on the handover problem. To this end, we formulate the association problem that maximizes the aggregate data rate along the trajectory while guaranteeing a predefined data rate threshold. We also consider the effect of choosing beam tracking or handover when the serving link quality drops on the user throughput along the trajectory. We apply reinforcement learning to learn whether and when the handover and beam tracking should be performed and choose the target base stations. We evaluate the proposed methods through numerical simulations and show promising results in terms of the achieved rate and throughput compared with the comparable methods in the literature.
In the third thread of the thesis, we focus on joint link configuration and resource allocation in a multi-user mobile scenario. We consider two kinds of link configuration: base station selection and beamwidth optimization in the case of handover. We apply reinforcement learning to approximate the solution of the association problem. In general, the main objective of our proposed method is to maximize the aggregate throughput of all the users along trajectories and ensure that their throughput in each location is higher than a threshold. Our numerical results demonstrate that the improved policy obtained from joint link configuration and resource allocation significantly outperforms other baseline policies in both throughput and achieved rate.