No / Over the past decade, wireless communications has experienced tremendous growth, and this growth is likely to multiply in the near future. The proliferation of mobile users and an ever increasing demand for multimedia services has resulted in greater capacity requirements. Radio frequency spectrum is scarce and cannot meet this ever increasing demand and the required Quality of Service (QoS) will no longer be achieved if efficient Radio Resource Management (RRM) solutions are not found. Conventional Radio Access Networks (RAN) have standalone Base Stations (BS) with capacity preconfigured for peak loads. These RANs have high call blocking and dropping rates since BSs resources cannot be shared. Cloud based RANs (C-RAN) have been proposed as a cost and energy efficient way of meeting high capacity demand of future wireless access networks by consolidating BSs to the cloud. Instead of relying on rejection of new call requests due to limited BS resources, C-RAN takes benefit of the cloud elasticity, which allows dynamic provisioning of cloud BS resources. This paper presents a novel C-RAN Call Admission Control (C-RAN CAC) to ensure Grade of Service (GoS) by improving blocking probability and improvement of call waiting times. Call blocking probability, call average waiting time and system utilization are used to evaluate the performance of the proposed CAC algorithm.

Fifth Generation (5G) cellular networks will experience a thousand-fold increase in data traffic with over 100 billion connected devices by 2020. In order to support this skyrocketing traffic demand, smaller base stations (BSs) are deployed to increase capacity. However, more BSs increase energy consumption which contributes to operational expenditure (OPEX) and CO2 emissions. Also, an introduction of a plethora of 5G applications running in the mobile devices cause a significant amount of energy consumption in the mobile devices. This thesis presents a novel framework for energy efficiency in 5G cloud radio access networks (C-RAN) by leveraging cloud computing technology. Energy efficiency is achieved in three ways; (i) at the radio side of H-C-RAN (Heterogeneous C-RAN), a dynamic BS switching off algorithm is proposed to minimise energy consumption while maintaining Quality of Service (QoS), (ii) in the BS cloud, baseband workload consolidation schemes are proposed based on simulated annealing and genetic algorithms to minimise energy consumption in the cloud, where also advanced fuzzy based admission control with pre-emption is implemented to improve QoS and resource utilisation (iii) at the mobile device side, Mobile Edge Computing (MEC) is used where computer intensive tasks from the mobile device are executed in the MEC server in the cloud. The simulation results show that the proposed framework effectively reduced energy consumption by up to 48% within RAN and 57% in the mobile devices, and improved network energy efficiency by a factor of 10, network throughput by a factor of 2.7 and resource utilisation by 54% while maintaining QoS.

The increasing demand for mobile data has placed significant strain on radio access networks (RANs), leading to a continuous need for increased network capacity. In keeping with that, a significant advancement in modern RANs is the ability to utilize several receivers and transmitters, to allow for beamforming. One way to increase the capacity of the network is therefore to optimize the resource allocation by preprocessing the transmitted signals, which involves several costly matrix multiplications (MMs). The aim of the project was to investigate the potential of accelerating Ericsson's RAN simulations by using AI based approximate matrix multiplication (AMM) algorithms. The main focus was on the multiply additionless (MADDNESS) algorithm, a product quantization technique that has achieved speedups of up to 100 times compared to exact MM, and 10 times faster than previous AMM methods. A complex matrix handling version of MADDNESS was implemented in Java and Python respectively, and its speed and accuracy were evaluated against Ericsson's current MM implementation. The proposed implementation did not beat the benchmark with respect to speed, instead resulting in a 4-10 times slowdown in runtime. However, this may largely be due to the fact that the used languages do not allow for complete control over memory resource allocation. As such, the implementations at hand do not incorporate all the crucial features of the algorithm. Particularly, the handicapped version does not fully leverage the vectorization potential, which is one of the key contributors to the speed of the algorithm. Consequently, further improvements are necessary before employing the techniques in an end-to-end implementation. / Den växande efterfrågan på mobildata har ökat belastningen på dagens radionätverk (RAN) och har medfört ett behov av att utvidga dess kapacitet. En betydande innovation inom RAN är beamforming, vilket är förmågan att fokusera digitala signaler mot mottagaren och på så vis öka singalstyrkan. En metod för att öka kapaciteten i ett nätverk är att optimera både kvaliteten av och resursallokeringen mellan nätverkets digitala kanaler, vilket medför tidskrävande matrismultiplikationer. Syftet med denna studie var att utforska om AI-baserade approximativa matrismultiplikationsalgoritmer har potentialen att accelerera Ericssons digitala tvilling-simuleringar. Studien fokuserade i huvudsak på produktkvantiseringsalgoritmen MADDNESS som påvisat potentialen att accelerera exakta matrismultiplikationer med en faktor 100, samt en faktor 10 snabbare än jämförbara approximativa metoder. En modifierad version av MADDNESS, som behandlar komplexa matriser, implementerades i Java samt Python, varefter precisionen och hastigheten utvärderades. Den föreslagna implementationen resulterade i en försämring med avseende på hastigheten med en faktor 4-10 jämfört med Ericssons nuvarande algoritmer. Den föreslagna implementationen saknar effektiv minnesallokering och misslyckas följaktligen att till fullo ta tillvara på vektoriseringspotentialen i MADDNESS. Detta indikerar att det är nödvändigt för ytterligare förbättringar innan algoritmen är användbar i den givna simuleringsmiljön.

Fifth Generation (5G) cellular networks will experience a thousand-fold increase in data traffic with over 100 billion connected devices by 2020. In order to support this skyrocketing traffic demand, smaller base stations (BSs) are deployed to increase capacity. However, more BSs increase energy consumption which contributes to operational expenditure (OPEX) and CO2 emissions. Also, an introduction of a plethora of 5G applications running in the mobile devices cause a significant amount of energy consumption in the mobile devices. This thesis presents a novel framework for energy efficiency in 5G cloud radio access networks (C-RAN) by leveraging cloud computing technology. Energy efficiency is achieved in three ways; (i) at the radio side of H-C-RAN (Heterogeneous C-RAN), a dynamic BS switching off algorithm is proposed to minimise energy consumption while maintaining Quality of Service (QoS), (ii) in the BS cloud, baseband workload consolidation schemes are proposed based on simulated annealing and genetic algorithms to minimise energy consumption in the cloud, where also advanced fuzzy based admission control with pre-emption is implemented to improve QoS and resource utilisation (iii) at the mobile device side, Mobile Edge Computing (MEC) is used where computer intensive tasks from the mobile device are executed in the MEC server in the cloud. The simulation results show that the proposed framework effectively reduced energy consumption by up to 48% within RAN and 57% in the mobile devices, and improved network energy efficiency by a factor of 10, network throughput by a factor of 2.7 and resource utilisation by 54% while maintaining QoS.

Fifth Generation (5G) cellular networks will experience a thousand-fold increase in data traffic with over 100 billion connected devices by 2020. In order to support this skyrocketing traffic demand, smaller base stations (BSs) are deployed to increase capacity. However, more BSs increases energy consumption which contributes to operational expenditure (OPEX) and CO2 emissions. Also, an introduction of a plethora of 5G applications running in the mobile devices causes a significant amount of energy consumption in the mobile devices. This thesis presents a novel framework for energy efficiency in 5G cloud radio access networks (C-RAN) by leveraging cloud computing technology. Energy efficiency is achieved in three ways; (i) at the radio side of H-C-RAN (Heterogeneous C-RAN), a dynamic BS switching off algorithm is proposed to minimise energy consumption while maintaining Quality of Service (QoS), (ii) in the BS cloud, baseband workload consolidation schemes are proposed based on simulated annealing and genetic algorithms to minimise energy consumption in the cloud, where also advanced fuzzy based admission control with pre-emption is implemented to improve QoS and resource utilisation (iii) at the mobile device side, Mobile Edge Computing (MEC) is used where computer intensive tasks from the mobile device are executed in the MEC server in the cloud. The simulation results show that the proposed framework effectively reduced energy consumption by up to 48% within RAN and 57% in the mobile devices, and improved network energy efficiency by a factor of 10, network throughput by a factor of 2.7 and resource utilisation by 54% while maintaining QoS.

Driven by the surging need for seamless connectivity, research in the wireless communication area has dramatically evolved over the years to meet the increasing demand for data rate and seamless coverage. Such evolvement concurs with a notable increase in data traffic and the widespread of data-hungry devices, thereby inflicting stringent requirements on terrestrial networks. Despite the tremendous advances achieved through the past generations of wireless systems, almost half of the world's population remains unconnected, leading to an accentuated digital divide problem. Therefore, this work invigorates a new connectivity solution that integrates aerial and terrestrial communications with a high-altitude platform station (HAPS) to promote a sustainable connectivity landscape. The connectivity solution adopted in this thesis specifically integrates terrestrial base stations with hot-air balloons under the framework of a cloud-enabled HAPS via a data-sharing fronthauling strategy. The aerial (hot-air balloons) and terrestrial base stations, grouped into disjoint clusters, coordinate their mutual transmission to serve aerial (i.e., drones) and terrestrial users. This work studies the downlink communication from the cloud-enabled HAPS to the aerial and terrestrial users under practical system considerations, namely the limited transmit power and the limited-capacity fronthaul link, per-base station. To this end, the first part of the thesis devises a specific optimization problem that maximizes the network sum-rate while accounting for system design constraints to determine the user association strategy, i.e., user to terrestrial clusters or user to air clusters, and the associated beamforming vectors. The second part of the thesis, then, designs a different resource allocations optimization problem that accounts for the fairness among the users, thus adopting a proportionally fair scheduling scheme to assign users on frequency tones to maximize the log of the long-term average rate. On this account, the work solves a handful of non-convex intricate optimization problems using techniques from optimization theory, namely, fractional programming and $\ell_0$-norm approximation. The work consequently outlines the gains realized by providing on-demand coverage in crowded and unserved areas. Moreover, the thesis illustrates the benefits of coordinating the operations of aerial and terrestrial base stations for interference management, load-balancing, and fairness measures.

Les réseaux mobiles de la cinquième génération (5G) sont conçus pour introduire de nouveaux services nécessitant des débits de données extrêmement hauts et une faible latence. 5G sera un changement de paradigme qui comprend des réseaux hétérogènes densifiés, des réseaux d'accès radio virtualisés, des fréquences porteuses à ondes millimétrées et des densités de périphériques très élevées. Cependant, contrairement aux générations précédentes, 5G sera un réseau holistique, intégrant n'importe quelle nouvelle technologie radio avec les technologies LTE et WiFi existant. Dans ce contexte, on se concentre sur de nouvelles stratégies d'allocation de ressources capables de bénéficier du multihoming dans le cas d'accès double au réseau. On modélise ces algorithmes au niveau du flux et analyse leurs performances en termes de débit, de stabilité du système et d'équité entre différentes catégories d'utilisateurs. On se concentre tout d'abord sur le multihoming dans les réseaux hétérogènes LTE/WiFi. On considère les allocations centrées sur le réseau où un planificateur central effectue des allocations d'équité proportionnelle (PF) locale et globale pour différentes classes d'utilisateurs, utilisateurs individuels (single-homed) et multi-domiciliés (multihomed). Par rapport à un modèle de référence sans multihoming, on montre que les deux stratégies améliorent la performance et la stabilité du système, au détriment d'une plus grande complexité pour la stratégie PF globale. On étudie également les stratégies d'allocation centrées sur l'utilisateur, dans lesquelles les utilisateurs multihomed décident la partition de la demande d'un fichier en utilisant soit la maximisation du débit crête, soit la stratégie assistée par réseau. On montre que cette dernière stratégie maximise le débit moyen dans l'ensemble du réseau. On montre également que les stratégies centrées sur le réseau permettent d'obtenir des débits de données plus élevés que ceux centrés sur l'utilisateur. Ensuite, on se concentre sur les réseaux d'accès radio virtuels (V-RAN) et en particulier sur l'allocation de multi-ressources. On étudie la faisabilité de la virtualisation sans diminuer ni la performance des utilisateurs, ni la stabilité du système. On considère un réseau hétérogène 5G composé de cellules LTE et mm-wave afin d'étudier comment les réseaux hauts fréquence peuvent augmenter la capacité du système. On montre que la virtualisation du réseau est réalisable sans perte de performance lors de l'utilisation de la stratégie « dominant resource fairness » (DRF). On propose une stratégie d'allocation en deux phases (TPA) qui montre un indice d'équité plus élevé que DRF et une stabilité du système plus élevée que PF. On montre également des gains importants apportés par l'adoption des fréquences mm-wave au lieu de WiFi. Finalement, on considère l'efficacité énergétique et compare les stratégies DRF et TPA avec une stratégie éconergétique basée sur l'algorithme de Dinklebach. Les résultats montrent que la stratégie éconergétique dépasse légèrement DRF et TPA à charge faible ou moyenne en termes de débit moyen plus élevé avec une consommation d'énergie comparable, alors qu'elle les surpasse à une charge élevée en termes de consommation d'énergie moins élevée. Dans ce cas de charge élevée, DRF surpasse TPA et la stratégie éconergétique en termes de débit moyen. En ce qui concerne l'indice d'équité de Jain, TPA réalise l'indice d'équité le plus élevé parmi d'autres stratégies / Fifth generation mobile networks (5G) are being designed to introduce new services that require extreme broadband data rates and utlra-reliable latency. 5G will be a paradigm shift that includes heterogeneous networks with densification, virtualized radio access networks, mm-wave carrier frequencies, and very high device densities. However, unlike the previous generations, it will be a holistic network, tying any new 5G air interface and spectrum with the currently existing LTE and WiFi. In this context, we focus on new resource allocation strategies that are able to take advantage of multihoming in dual access settings. We model such algorithms at the flow level and analyze their performance in terms of flow throughput, system stability and fairness between different classes of users. We first focus on multihoming in LTE/WiFi heterogeneous networks. We consider network centric allocations where a central scheduler performs local and global proportional fairness (PF) allocations for different classes of users, single-homed and multihomed users. By comparison with a reference model without multihoming, we show that both strategies improve system performance and stability, at the expense of more complexity for the global PF. We also investigate user centric allocation strategies where multihomed users decide the split of a file using either peak rate maximization or network assisted strategy. We show that the latter strategy maximizes the average throughput in the whole network. We also show that network centric strategies achieve higher data rates than the user centric ones. Then, we focus on Virtual Radio Access Networks (V-RAN) and particularly on multi-resource allocation therein. We investigate the feasibility of virtualization without decreasing neither users performance, nor system's stability. We consider a 5G heterogeneous network composed of LTE and mm-wave cells in order to study how high frequency networks can increase system's capacity. We show that network virtualization is feasible without performance loss when using the dominant resource fairness strategy (DRF). We propose a two-phase allocation (TPA) strategy which achieves a higher fairness index than DRF and a higher system stability than PF. We also show significant gains brought by mm-wave instead of WiFi. Eventually, we consider energy efficiency and compare DRF and TPA strategies with a Dinklebach based energy efficient strategy. Our results show that the energy efficient strategy slightly outperforms DRF and TPA at low to medium load in terms of higher average throughput with comparable power consumption, while it outperforms them at high load in terms of power consumption. In this case of high load, DRF outperforms TPA and the energy efficient strategy in terms of average throughput. As for Jain's fairness index, TPA achieves the highest one

[ES] El estándar de quinta generación (5G) es la clave potencial para satisfacer el aumento exponencial en la demanda de nuevas aplicaciones, servicios y usuarios. La tecnología 5G ofrecerá una latencia extremadamente baja de 1 ms, una velocidad máxima de datos de 10 Gbit/s, una alta densidad de conexión de hasta 106 dispositivos/km2 y permitirá una alta movilidad de los dispositivos de hasta 500 km/h. En esta Tesis se proponen varias soluciones basadas en tecnologías habilitadoras para el despliegue de redes 5G. La arquitectura de la red de acceso de radio en la nube (C-RAN) se emplea junto con las técnicas de Fotónica de Microondas como una solución prometedora para generar y transmitir señales de ondas milimétricas (mmW) en la próxima generación de comunicaciones móviles. La tecnología radio sobre fibra (RoF) ha demostrado ser una buena opción para enfrentarse al desafío de la distribución inalámbrica mmW debido a la gran distancia de transmisión, el gran ancho de banda y la inmunidad a las interferencias electromagnéticas, entre algunas de las principales ventajas. Además, esta tecnología se puede ampliar con comunicaciones ópticas de espacio libre (FSO) en sistemas de radio sobre FSO (RoFSO) en las redes inalámbricas. En esta Tesis, las señales mmW se generan fotónicamente mediante modulación externa de doble banda lateral con supresión de portadora (CS-DSB) y se distribuyen a través de enlaces fronthaul híbridos RoF/FSO. Además, la generación múltiple de señales permite la distribución reconfigurable en canales multiplexados por división de longitud de onda (WDM) desde una oficina central hasta las estaciones base, y se ha evaluado el impacto de las turbulencias producidas en los canales FSO sobre las señales mmW generadas fotónicamente en términos de fluctuaciones de potencia y ruido de fase de la señal. Se propone la técnica de modulación directa de un láser (DML) como solución principal para la transmisión de datos a través de enlaces ópticos híbridos que emplean un esquema de multiplicación de frecuencias ópticas, es decir, CS-DSB, para la generación de señales de mmW. En concreto, se evalúan teórica y experimentalmente los esquemas de generación fotónica local y remoto de señales mmW y se comparan para su implementación práctica en la red frontal de la C-RAN y, además, se estudia experimentalmente el impacto de la distorsión armónica y de la intermodulación en la transmisión de datos. Igualmente, con el fin de obtener la capacidad que ofrece el DML en términos de ancho de banda, también se presenta una evaluación teórica y experimental del efecto de la dispersión de la fibra y el chirp sobre diferentes anchos de banda de señales de M-modulación de amplitud en cuadratura (QAM). No obstante, la Tesis también incluye otro enfoque para la transmisión de datos basado en el uso de otro modulador externo. En este caso, la demostración experimental de la generación de señales ópticas empleando CS-DSB y la transmisión de señales a través de fibra híbrida y red frontal FSO se completa con un enlace de antena que permite transmitir señales 5G 64/256-QAM. La investigación realizada con los sistemas CS-DSB y DSB también permiten comparar la robustez frente al desvanecimiento inducido por la dispersión cromática de la fibra. Además, se ha realizado una evaluación experimental impacto las turbulencias producidas en los canales FSO sobre las señales mmW generadas fotónicamente con diferentes distribuciones térmicas y se ha cuantificado la degradación de la señal de datos de acuerdo con las condiciones de la turbulencia. Como demostradores finales, esta Tesis incluye un sistema de transmisión full-dúplex que emplea señales 5G en enlace descendente (DL) a 39 GHz y en enlace ascendente (UL) a 37 GHz; y la transmisión de señales OFDM LTE de 60 GHz (DL) y 25 GHz (UL) sobre una infraestructura heterogénea de frontal óptico que consiste en fibra óptica de 10 km, un canal FSO de 100 m y un enlace de radio inalámbrico de 2 m. / [CA] L'estàndard de quinta generació (5G) és la clau potencial per a satisfer l'augment exponencial en la demanda de noves aplicacions, serveis i usuaris. La tecnologia 5G oferirà una latència extremadament baixa d'1 ms, una velocitat màxima de dades de 10 Gbit/s, una alta densitat de connexió de fins a 106 dispositius/km2 i permetrà una alta mobilitat dels dispositius de fins a 500 km/h. En aquesta tesi es proposen diverses solucions basades en tecnologies habilitadores per al desplegament de xarxes 5G. L'arquitectura de la xarxa d'accés de ràdio en el núvol (CRAN) s'empra junt amb les tècniques de Fotònica de Microones com una solució prometedora per a generar i transmetre senyals d'ones mil·limètriques (mmW) en la pròxima generació de comunicacions mòbils. La tecnologia ràdio sobre fibra ( RoF) ha demostrat ser una bona opció per a enfrontar-se al desafiament de la distribució sense fil mmW a causa de la gran distància de transmissió, el gran ample de banda i la immunitat a les interferències electromagnètiques, entre alguns dels principals avantatges. A més, aquesta tecnologia es pot ampliar amb comunicacions òptiques d'espai lliure (FSO) en sistemes de ràdio sobre FSO (RoFSO) en les xarxes sense fil. En aquesta Tesi, els senyals mmW es generen fotònicament per mitjà de modulació externa de doble banda lateral amb supressió de portadora (CS-DSB) i es distribueixen a través d'enllaços frontals híbrids RoF/FSO.. A més, la generació múltiple de senyals permet la distribució reconfigurable en canals multiplexats per divisió de longitud d'ona ( WDM) des d'una oficina central fins a les estacions base, i s'ha avaluat l'impacte de les turbulències produïdes en els canals FSO sobre els senyals mmW generades fotònicament en termes de fluctuacions de potència i soroll de fase del senyal. Aquest treball proposa la tècnica de modulació directa d'un làser (DML) com solució principal per a la transmissió de dades a través d'enllaços òptics híbrids que fan servir un esquema de multiplicació de freqüències òptiques, és a dir, CS-DSB, per a la generació de senyals de mmW. En concret, s'avalua teòric i experimentalment els esquemes de generació fotònica local i remota de senyals mmW i es comparen per a la seua implementació pràctica a la xarxa frontal de la C-RAN i a més, s'estudia experimentalment l'impacte de la distorsió harmònica i de la intermodulació en la transmissió de dades. Igualment, amb el fi d'obtindre la capacitat que ofereix el DML en termes d'amplada de banda, també es presenta una avaluació teòrica i experimental de l'efecte de la dispersió de la fibra i el chirp sobre diferents amples de banda de senyals de M-modulació d'amplitud en quadratura (QAM). No obstant això, la Tesis també inclou altre enfocament per a la transmissió de dades basat amb l¿ús d'altre modulador extern. En aquest cas, la demostració experimental de la generació de senyals òptics emprant CS-DSB i la transmissió de senyals a través de fibra híbrida i xarxa frontal FSO es completa com un enllaç d'antena que permet transmetre senyals 5G 64/256-QAM. La investigació realitzada amb els sistemes CS-DSB i DSB també permet comparar la seua robustesa davant l¿esvaïment induït per la dispersió cromàtica. A més, s'ha avaluat experimentalment l'impacte de les turbulències produïdes en els canals FSO sobre els senyals mmW generades fotònicament amb diferents distribucions tèrmiques i s'ha quantificat la degradació del senyal de dades d'acord amb les condicions de la turbulència. Com a demostradors finals, aquesta Tesi inclou un sistema de transmissió full-dúplex que empra senyals 5G en enllaç descendent (DL) a 39 GHz i en enllaç ascendent (UL) a 37 GHz; i la transmissió de senyals OFDM LTE de 60 GHz (DL) i 25 GHz (UL) sobre una infraestructura heterogènia de frontal òptic que consisteix en fibra òptica de 10 km, un canal FSO de 100 m i un enllaç de ràdio sense fil de 2 m. / [EN] The fifth generation (5G) standard is the potential key to meet the exponentially increasing demand of the emerging applications, services and mobile end users. 5G technology will offer an extremely low latency of 1 ms, peak data rate of 10 Gbit/s, high contention density up to 106 devices/km2 and enable high mobility up to 500 km/h. This Thesis proposes several solutions based on enabling technologies for deploying 5G networks. Cloud-radio access network (C-RAN) architecture is employed in conjunction with microwave photonics techniques as a promising solution to generate and transmit millimeter wave (mmW) signals in the next generation of mobile communications. Radio over fiber (RoF) has been demonstrated as a good option to face the challenge of mmW wireless distribution, due to long transmission distance, large bandwidth and immunity to electromagnetic interference, as some of the main advantages. Moreover, this technology can be extended with free-space optical (FSO) communications in Radio over FSO systems (RoFSO) as wireless networks. In this Thesis, mmW signals are photonically generated by carrier suppressed double sideband (CS-DSB) external modulation and distributed over hybrid RoF/FSO fronthaul links. Moreover, multiple generated signals allow reconfigurable distribution in wavelength-division multiplexed (WDM) channels from a central office to the base stations, and the impact of turbulent FSO channels on photonically generated mmW signals has been evaluated in terms of power signal fluctuations and phase noise. A directly modulated laser (DML) is proposed as a major solution for signal transmission over hybrid optical links employing optical frequency multiplication scheme, i.e. CS-DSB, for mmW signal generation. Moreover, local and remote photonic mmW signal generation schemes are theoretically and experimentally evaluated and compared for practical deployment in C-RAN fronthaul network while the impact of harmonic and intermodulation distortion on data transmission is also experimentally studied. Furthermore, for the sake of obtaining the DML usability in terms of bandwidth, theoretical and experimental evaluation of the effect of fiber dispersion and chirp over different M-quadrature amplitude modulation (QAM) signals bandwidth is also presented. Another data transmission approach based on the cascade of two external modulators is also employed in the Thesis. In this case, the experimental demonstration of optical signal generation employing CS-DSB and signal transmission over hybrid fiber and FSO fronthaul network is completed with a seamless antenna link leading to successful transmission of 64/256-QAM 5G signals. The CS-DSB and DSB schemes are also investigated for the sake of comparison in terms of robustness against fiber chromatic dispersion-induced fading. Furthermore, experimental evaluation of the impact of turbulent FSO links on photonically generated mmW signals with different thermal distributions has been performed and data signal degradation has been quantified according to the turbulence conditions. As final demonstrators, the Thesis includes a full-duplex transmission system employing 39 GHz downlink (DL) and 37 GHz uplink (UL) 5G signals over hybrid links; and 60 GHz (DL) and 25 GHz (UL) OFDM LTE signal transmission over an heterogeneous optical fronthaul infrastructure consisting of 10 km optical fiber, 100 m FSO channel and 2 m wireless radio link. / I would like to acknowledge the financial support given by Research Excellence Award Programme GVA PROMETEO 2017/103 Future Microwave Photonics and European Network for High Performance Integrated Microwave Photonics (EUIMWP) CA16220. / Vallejo Castro, L. (2022). Photonic Millimeter Wave Signal Generation and Transmission Over Hybrid Links in 5G Communication Networks [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/190253