Please use this identifier to cite or link to this item:
Title: Radio engineering of relay-based cellular networks
Authors: Minelli, Mattia
Keywords: DRNTU::Engineering::Electrical and electronic engineering
Issue Date: 2017
Source: Minelli, M. (2017). Radio engineering of relay-based cellular networks. Doctoral thesis, Nanyang Technological University, Singapore.
Abstract: Data traffic demand in today's cellular networks is growing very rapidly, and operators are looking for cost-effective solutions to increase the capacity and coverage of their networks. Relay nodes have been proposed as one of the candidate solutions to achieve this target. Briefly, a relay node is a network device which is in charge of forwarding the data received from its reference base stations to a number of controlled mobile users, and forwarding the data transmitted by these mobile users to the base station. In this way, the base station can be partially offioaded, and the spatial density of network access point is increased. A similar result could be obtained by deploying more base stations. However, relays may be preferred by network operators, because they are smaller and cheaper than base stations, and they should hence allow for increasing the density of network stations without an exorbitant increment of the costs related to the deployment of new nodes and the related site rentals. Moreover, relay nodes do not require any backhaul cable, as they are connected to their reference base station via a wireless backhaul link. However, the deployment of relays entails the creation of a two-hop wireless network. This means, that interference management is more complex compared to traditional cellular networks. Moreover, transmissions on the wireless backhaul link may use a significant quota of the available radio resources, thus reducing the resources available for transmissions to users. Starting from the expected advantages and disadvantages of relays deployment, this thesis is focused on the performance evaluation and the optimization of the radio layer of relay-enhanced cellular networks. In particular, numerous aspects have been treated, including the application of the fluid network model to the modeling of relay networks, the comparison of different candidate scheduling schemes and their impact on performance, the optimization of relays placement in the cell for capacity maximization and the assessment of the energy savings that users can achieve in relay networks. This thesis analyzes the important tradeoff, between the performance gain due to the higher number of network station, and the performance loss due to the reduced amount of radio resources available at each station for transmissions to/from users. The analysis is carried out in both a static environment, i.e., focusing on a simpler model and ignoring stations buffers queues length and users arrivals and departures, and in a dynamic environment, i.e., considering queuing and arrivals/departures. The performance of two schedulers, namely the proportional fair and the round robin scheduler, is assessed in the context of relay-enhanced networks, showing that the more complex proportional fair scheduler does not necessarily perform significantly better than a round robin scheduler. This depends on the propagation environment, on the network setup and on the number of relays and users. The study is performed via a novel statistical framework for performance evaluation of relay networks, which allows to take into account fast fading without explicitly simulating it. Regarding network modeling, a part of this thesis is devoted to the applica- ' tion of the fluid network model to relay networks. The fluid model is a mathematical tool for cellular networks performance analysis, based on approximating a discrete network of transmitters with a continuum. A novel framework for the analysis of relay networks, based on the fluid model, is proposed, and its accuracy is tested in several scenarios. In particular, regular and irregular relay networks are considered, and the fluid model is applied either to the whole network of relays, or only to those relays which are far from the cell of interest. The obtained results allow to assess the accuracy of the fluid model. In particular, they show that the fluid model is accurate when the network of relays is regular, or when it is used to model far relay nodes. In this thesis, the fluid model is also applied to find the optimized relays transmit power. Indeed, the application of the fluid model allows to find the borders of the areas where relays are capable of providing a better downlink rate, <;ompared to base stations. This area is then matched as closely as possible with the area where relays provide the highest downlink received power (under the assumption that users are served by the station providing the highest received power), by tuning relays transmit power appropriately. The problem of matching the two areas is solved via a geometrical approach, and an approximate solution is proposed. A significant part of the thesis is dedicat~d to relay networks optimization, both on the downlink and on the uplink. For the downlink, a study on the optimal placement of relays for capacity increase is proposed. Indeed, relays placement has a significant impact on network performance, although most of the literature tends to adopt network models where relays are arbitrarily placed. In this thesis, the optimal placement is found under a dynamic system model, where stations buffers queues loads and their impact on interference are taken into account. The adoption of this model i~proves the accuracy of results, as in a dynamic model interference is not overestimated as, e.g., when all stations are assumed to always have data to transmit. Hence, a more precise evaluation of the mutual interference between relays, and between relays and base stations can be performed. A novel optimization algorithm for relays placement is introduced, based on a multiscale implementation of the simulated annealing algorithm with a number of ad-hoc enhancements, which are described in the text. The obtained results provide useful guidelines for relays placement, and show that the optimal placement takes into account both the need for enjoying a good signal quality on the backhaul link, and that of covering the regions close to the edge of the cell, where users experience the lowest downlink rates. For the network uplink, this thesis proposes an optimization of the network setup, in order to maximize the energy efficiency of users. Indeed, the deployment of relays shortens the average distance between users and their serving stations. This means that users can transmit to the stations with a lower power, thus reducing the consumption of batteries. Several studies have shown that a pure optimization of network energy efficiency can lead to optimal network setups which perform poorly in terms of throughput. Hence, in this work the optimization of energy efficiency is performed under a constraint on the minimum acceptable network performance. A novel optimization framework for uplink energy efficiency is introduced. The search of the best network setup over a very fragmented and difficult configuration space is tackled by means of an exterior search approach, tailored to the studied case. A customized penalty function is proposed and tested, with the aim of increasing the effectiveness of the exterior search. Results show that an optimized relays deployment allows for great users energy savings. These savings are given for different values of the traffic input and number of relays, in several deployment scenarios. A summary of the main contributions of this thesis is given in Section 1.3.
DOI: 10.32657/10356/70618
Fulltext Permission: open
Fulltext Availability: With Fulltext
Appears in Collections:EEE Theses

Files in This Item:
File Description SizeFormat 
MATTIA_MINELLI_2016.pdf22.5 MBAdobe PDFThumbnail

Page view(s)

Updated on Jun 22, 2021

Download(s) 50

Updated on Jun 22, 2021

Google ScholarTM




Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.