Master's degree in Telecommunications Engineering (MET)
http://hdl.handle.net/2099.1/22913
2024-03-28T20:01:42ZSemiconductor quantum dot lasers for frequency comb generation
http://hdl.handle.net/2117/403848
Semiconductor quantum dot lasers for frequency comb generation
Palacios Tomás, Pablo Isidro
The rapid evolution of machine learning and artificial intelligence, coupled with the exponential growth in data traffic, has heightened the demand for faster and more energy- efficient computational solutions. Addressing this need, the integration of photonic integrated circuits into emerging devices, along with the implementation of the Dense Wave- length Division Multiplexing (DWDM) technique, has gained prominence in the growing field of High Performance Computing (HPC). For such purpose, a laser source is required. Among all the possible solutions, optical frequency combs have emerged as a promising candidate due to their fixed channel spacing, narrow linewidth, and high efficiency performance. More specifically, optical frequency combs utilizing a quantum dot gain medium exhibit exceptional characteristics, including high thermal stability, tolerance to optical feedback, low threshold current density, ultrafast carrier dynamics, and minimal relative intensity and phase noise among others. This thesis undertakes a comprehensive investigation into the performance and potential advantages of quantum dot comb lasers integrated in silicon, laying a rigorous theoretical foundation that aims to explain the obtained experimental results. These results come from the characterization of a set of 1.3 μm lasers provided by Hewlett- Packard (HP) labs. In particular, after a search for the highest performance laser, the selected one has been tested under different conditions of bias current, reverse bias voltage of the saturable absorber that was incorporated in the laser and different temperatures. Several maps, such as the beatnote power, number of modes, beatnote width, spectral flatness, etc. dependency on the different conditions have been produced. Furthermore, this thesis includes a transmission experiment aimed at demonstrating the potential of such laser for communications. However, the results deviated from expectations, as the formation of the comb does not necessarily coincide with the best eye diagram. Further investigations are required at this stage and left for future research. In conclusion, this research sheds light on the promising prospects of quantum dot frequency comb lasers for high performance computing, providing a foundational understanding of their operation, and highlighting avenues for future exploration and optimization by the experimental investigation of a cutting-edge HP quantum dot frequency comb laser integrated on silicon.
2024-03-06T12:06:05ZPalacios Tomás, Pablo IsidroThe rapid evolution of machine learning and artificial intelligence, coupled with the exponential growth in data traffic, has heightened the demand for faster and more energy- efficient computational solutions. Addressing this need, the integration of photonic integrated circuits into emerging devices, along with the implementation of the Dense Wave- length Division Multiplexing (DWDM) technique, has gained prominence in the growing field of High Performance Computing (HPC). For such purpose, a laser source is required. Among all the possible solutions, optical frequency combs have emerged as a promising candidate due to their fixed channel spacing, narrow linewidth, and high efficiency performance. More specifically, optical frequency combs utilizing a quantum dot gain medium exhibit exceptional characteristics, including high thermal stability, tolerance to optical feedback, low threshold current density, ultrafast carrier dynamics, and minimal relative intensity and phase noise among others. This thesis undertakes a comprehensive investigation into the performance and potential advantages of quantum dot comb lasers integrated in silicon, laying a rigorous theoretical foundation that aims to explain the obtained experimental results. These results come from the characterization of a set of 1.3 μm lasers provided by Hewlett- Packard (HP) labs. In particular, after a search for the highest performance laser, the selected one has been tested under different conditions of bias current, reverse bias voltage of the saturable absorber that was incorporated in the laser and different temperatures. Several maps, such as the beatnote power, number of modes, beatnote width, spectral flatness, etc. dependency on the different conditions have been produced. Furthermore, this thesis includes a transmission experiment aimed at demonstrating the potential of such laser for communications. However, the results deviated from expectations, as the formation of the comb does not necessarily coincide with the best eye diagram. Further investigations are required at this stage and left for future research. In conclusion, this research sheds light on the promising prospects of quantum dot frequency comb lasers for high performance computing, providing a foundational understanding of their operation, and highlighting avenues for future exploration and optimization by the experimental investigation of a cutting-edge HP quantum dot frequency comb laser integrated on silicon.Efficiently learning the system model for microservice-based applications
http://hdl.handle.net/2117/403254
Efficiently learning the system model for microservice-based applications
Macià Coll, Laura
Microservice-based applications are increasing in popularity. They offer numerous advantages and qualities (i.e., decentralized architecture, ease of monitoring and deployment, or scalability). However, controlling a virtualized environment for microservice-based applications is difficult due to its dynamic configuration. Previous research shows that it is possible to control the system response and manage and optimize the system's performance through an accurate system model. Nonetheless, to learn the system model and the operating region, many monitored samples from the system are usually required, rendering these approaches unpractical (or practically infeasible). In this study, we demonstrate that learning the system model and operating region for a given accuracy range with a reduced number of samples is possible to make monitoring feasible. Moreover, we present the effect of different background configuration settings on the behavior of the services running on the testbed. In our approach, the system model is learned from a laboratory testbed where we execute information and computing services on a service mesh supported by Istio and Kubernetes. By observing the system and applying supervised learning methods, we estimate the service metric from service statistics for each measurement and study the trade-off between the number of training samples and the model's accuracy. Our results are novel in that they present a system model and the operating region of the system that allows us to perform near-optimal predictions of the system while reducing the monitoring time from seven to four days while presenting efficient actions to control the system.
2024-02-27T08:38:08ZMacià Coll, LauraMicroservice-based applications are increasing in popularity. They offer numerous advantages and qualities (i.e., decentralized architecture, ease of monitoring and deployment, or scalability). However, controlling a virtualized environment for microservice-based applications is difficult due to its dynamic configuration. Previous research shows that it is possible to control the system response and manage and optimize the system's performance through an accurate system model. Nonetheless, to learn the system model and the operating region, many monitored samples from the system are usually required, rendering these approaches unpractical (or practically infeasible). In this study, we demonstrate that learning the system model and operating region for a given accuracy range with a reduced number of samples is possible to make monitoring feasible. Moreover, we present the effect of different background configuration settings on the behavior of the services running on the testbed. In our approach, the system model is learned from a laboratory testbed where we execute information and computing services on a service mesh supported by Istio and Kubernetes. By observing the system and applying supervised learning methods, we estimate the service metric from service statistics for each measurement and study the trade-off between the number of training samples and the model's accuracy. Our results are novel in that they present a system model and the operating region of the system that allows us to perform near-optimal predictions of the system while reducing the monitoring time from seven to four days while presenting efficient actions to control the system.Image and radar fusion for autonomous driving
http://hdl.handle.net/2117/403218
Image and radar fusion for autonomous driving
De Gibert Duart, Francesc Xavier
2024-02-26T15:15:44ZDe Gibert Duart, Francesc XavierOptimal transmit array processing: a study for satellite communications
http://hdl.handle.net/2117/403092
Optimal transmit array processing: a study for satellite communications
Carreras Mallol, Sergi
Low Earth Orbit satellite communications nowadays are an increasingly popular method to provide coverage and Internet access globally. Worldwide companies such as SpaceX (via Starlink) or ViaSat are launching full constellations of LEO nanosatellites to ensure correct coverage levels all around the globe. The aim of this work is to study the benefits of array processing and beamforming, especially for downlink satellite communications, in order to provide proper received power to the serving user while keeping interferences to other nearby users as low as possible. This scenario is particularly significant when considering densely user-populated deployments or, in general, in situations where multiple end users are located in the same satellite footprint area.
2024-02-23T18:45:09ZCarreras Mallol, SergiLow Earth Orbit satellite communications nowadays are an increasingly popular method to provide coverage and Internet access globally. Worldwide companies such as SpaceX (via Starlink) or ViaSat are launching full constellations of LEO nanosatellites to ensure correct coverage levels all around the globe. The aim of this work is to study the benefits of array processing and beamforming, especially for downlink satellite communications, in order to provide proper received power to the serving user while keeping interferences to other nearby users as low as possible. This scenario is particularly significant when considering densely user-populated deployments or, in general, in situations where multiple end users are located in the same satellite footprint area.Combination of fully focused sentinel-6 and interferometric CryoSat-2 to enhance sea ice retrievals
http://hdl.handle.net/2117/403091
Combination of fully focused sentinel-6 and interferometric CryoSat-2 to enhance sea ice retrievals
Guerra Bernal, Michel Enrique
Radar altimetry has proven to be key in Earth observation when retrieving remote and systematic data for looking at long term trends, this is the case of the cryosphere monitoring. In this context, important missions addressed for altimetry are Sentinel-6 and CryoSat-2, launched by ESA. This study conveys methods for retrieval and estimation of surface ice parameters and features by combining data from both of them. In a first phase, Fully Focused SAR data processed in Athabasca Lake, along with interferometric data from CryoSat-2, are combined to classify the visible features and estimate the signal source location. Here, the advantages of Fully Focused processing over Delay-Doppler processing are highlighted. In a second stage, four methods for sea ice concentration estimation are employed using Sentinel-6 data from the year 2021 in a small region of the antarctic Weddell Sea, one based in histogram clustering identification and the others based on curve fitting and local maxima isolation, waveform per waveform. The results show a good convergence between the fitting filters and the local maxima detection. Finally, the validation data from AMSR2 shows agreement with the time series trend in 2021.
2024-02-23T18:37:51ZGuerra Bernal, Michel EnriqueRadar altimetry has proven to be key in Earth observation when retrieving remote and systematic data for looking at long term trends, this is the case of the cryosphere monitoring. In this context, important missions addressed for altimetry are Sentinel-6 and CryoSat-2, launched by ESA. This study conveys methods for retrieval and estimation of surface ice parameters and features by combining data from both of them. In a first phase, Fully Focused SAR data processed in Athabasca Lake, along with interferometric data from CryoSat-2, are combined to classify the visible features and estimate the signal source location. Here, the advantages of Fully Focused processing over Delay-Doppler processing are highlighted. In a second stage, four methods for sea ice concentration estimation are employed using Sentinel-6 data from the year 2021 in a small region of the antarctic Weddell Sea, one based in histogram clustering identification and the others based on curve fitting and local maxima isolation, waveform per waveform. The results show a good convergence between the fitting filters and the local maxima detection. Finally, the validation data from AMSR2 shows agreement with the time series trend in 2021.UWB compact probe microwave imaging antenna
http://hdl.handle.net/2117/403068
UWB compact probe microwave imaging antenna
Singh, Chaman
The medical imaging has undergone significant advancements, leading to the development of human imaging systems with antennas working at microwave frequencies, thanks to its low-cost, compact size, and the use of non-ionizing waves, to visualize the internal parts of the human body by producing micro-wave images. One of the challenging tasks of a microwave imaging (MWI) system is to design compact miniaturized Ultra-wide band (UWB) antennas at microwave frequency. These antennas are used on-body, so they need to be compact in size. To get a better imaging resolution wide bandwidth is one of the requirements. In this master's thesis an UWB compact Parabolic Back Cavity Waveguide (PBCW) antenna is designed using CST studio. The antenna is matched to the human brain-modelled medium. The PBCW antenna is working with a frequency range from 0.8 GHz to 2.75 GHz, with an overall size of 44.5 mm × 34 mm × 15 mm. The probe antenna is fabricated and measured in the lossy liquid mimicking the average permittivity of the brain medium, the scattering parameters are measured with a network analyzer ZNB40. The result shows a good agreement between the numerical and experimental validation of the antennas.
2024-02-23T15:26:53ZSingh, ChamanThe medical imaging has undergone significant advancements, leading to the development of human imaging systems with antennas working at microwave frequencies, thanks to its low-cost, compact size, and the use of non-ionizing waves, to visualize the internal parts of the human body by producing micro-wave images. One of the challenging tasks of a microwave imaging (MWI) system is to design compact miniaturized Ultra-wide band (UWB) antennas at microwave frequency. These antennas are used on-body, so they need to be compact in size. To get a better imaging resolution wide bandwidth is one of the requirements. In this master's thesis an UWB compact Parabolic Back Cavity Waveguide (PBCW) antenna is designed using CST studio. The antenna is matched to the human brain-modelled medium. The PBCW antenna is working with a frequency range from 0.8 GHz to 2.75 GHz, with an overall size of 44.5 mm × 34 mm × 15 mm. The probe antenna is fabricated and measured in the lossy liquid mimicking the average permittivity of the brain medium, the scattering parameters are measured with a network analyzer ZNB40. The result shows a good agreement between the numerical and experimental validation of the antennas.Satellite Laser Ranging network contact planning and performance assessment under adverse meteorological conditions
http://hdl.handle.net/2117/402873
Satellite Laser Ranging network contact planning and performance assessment under adverse meteorological conditions
Cugota Latorre, Carles
In recent timeps, SLR (Satellite Laser Ranging) is being used lately for orbit determination in many satellite constellations. It improves the accuracy of traditional radiofrequency systems by using laser transmitters, but the short-wavelength signals that are used in the lasers make them very weak against the atmospheric effects between the ground station and the satellite. In this thesis, the study on how the clouds affect the contact plan of the SLR stations will be done. For this, LEO, MEO and GEO constellations will be considered, as well as three ground station scenarios: cloudless, cloudy and hybrid. The contact plan for each ground station scenario with each satellite constellation will be executed to analyze the results, and draw conclusions.
2024-02-23T08:19:11ZCugota Latorre, CarlesIn recent timeps, SLR (Satellite Laser Ranging) is being used lately for orbit determination in many satellite constellations. It improves the accuracy of traditional radiofrequency systems by using laser transmitters, but the short-wavelength signals that are used in the lasers make them very weak against the atmospheric effects between the ground station and the satellite. In this thesis, the study on how the clouds affect the contact plan of the SLR stations will be done. For this, LEO, MEO and GEO constellations will be considered, as well as three ground station scenarios: cloudless, cloudy and hybrid. The contact plan for each ground station scenario with each satellite constellation will be executed to analyze the results, and draw conclusions.Enhancing random access success probability for NB-IoT over NTN
http://hdl.handle.net/2117/402428
Enhancing random access success probability for NB-IoT over NTN
Moliner Préjano, Álvaro
Narrow band Internet of Things (NB-IoT) is a technology that has been growing in recent years and will continue to play an important role in fifth (5G) and sixth (6G) generations of mobile communications in upcoming years. NB-IoT covers a big part of the communication market segment with several examples including monitoring of smart cities, agriculture, cattle and forest or a variety of industrial applications. Some of these scenarios are remote places where there is not an accessible terrestrial network and there is also not a great demand of service to deploy a new one. One solution to address this issue is the use of Non-Terrestrial Networks (NTN) which would guarantee satellite connectivity to these remote areas. On this project, a MATLAB code has been developed to simulate a satellite scenario where several NB-IoT devices will try to access the satellite using a Random Access (RA) procedure. With the help of the simulations, it has been determined which parameters of the RA are the most critical to the system capacity. Four methods are proposed which can help to equalize the capacity across the NTN cell to guarantee fairness for all User Equipment (UEs).; Internet de las Cosas de banda estrecha (NB-IoT) es una tecnología que ha estado creciendo en los últimos años y seguirá desempeñando un papel importante en la quinta (5G) y sexta (6G) generaciones de comunicaciones móviles en los próximos años. NB-IoT cubre una gran parte del segmento del mercado de la comunicación con numerosos os ejemplos que incluyen la monitorización de ciudades inteligentes, agricultura, ganadería y áreas forestales o una amplia variedad de aplicaciones industriales. Algunos de estos escenarios son lugares remotos donde no hay una red terrestre accesible y tampoco hay una gran demanda de servicio para desplegar una nueva. Una solución para abordar este problema es el uso de Redes No Terrestres (NTN) que garantizarían la conectividad satelital a estas áreas remotas. En este proyecto se ha desarrollado un código MATLAB para simular un escenario de satélite donde varios dispositivos NB-IoT intentarán acceder al satélite mediante un procedimiento de Acceso Aleatorio (RA). Con la ayuda de las simulaciones, se ha determinado qué parámetros del RA son los más críticos para la capacidad del sistema. Se proponen cuatro métodos que pueden ayudar a igualar la capacidad en la celda NTN para garantizar la equidad para todos los equipos de usuario (UE).; Internet de les coses de banda estreta (NB-IoT) és una tecnologia que ha anat creixent en els darrers anys i que continuarà jugant un paper important en la cinquena (5G) i la sisena (6G) generacions de comunicacions mòbils en els propers anys. NB-IoT cobreix una gran part del segment del mercat de la comunicació amb diversos exemples, com ara el monitoratge de ciutats intel·ligents, l'agricultura, la ramaderia i les àrees forestals o una varietat d'aplicacions industrials. Alguns d'aquests escenaris són llocs remots on no hi ha una xarxa terrestre accessible i tampoc hi ha una gran demanda de servei per desplegar-ne una de nova. Una solució per abordar aquest problema és l'ús de xarxes no terrestres (NTN) que garantirien la connectivitat per satèl·lit a aquestes zones remotes. En aquest projecte, s'ha desenvolupat un codi MATLAB per simular un escenari de satèl·lit on diversos dispositius NB-IoT intentaran accedir al satèl·lit mitjançant un procediment d'accés aleatori (RA). Amb l'ajuda de les simulacions, s'ha determinat quins paràmetres de la RA són els més crítics per a la capacitat del sistema. Es proposen quatre mètodes que poden ajudar a igualar la capacitat de la cèl·lula NTN per garantir l'equitat per a tots els equips d'usuari (UE).
2024-02-21T11:56:35ZMoliner Préjano, ÁlvaroNarrow band Internet of Things (NB-IoT) is a technology that has been growing in recent years and will continue to play an important role in fifth (5G) and sixth (6G) generations of mobile communications in upcoming years. NB-IoT covers a big part of the communication market segment with several examples including monitoring of smart cities, agriculture, cattle and forest or a variety of industrial applications. Some of these scenarios are remote places where there is not an accessible terrestrial network and there is also not a great demand of service to deploy a new one. One solution to address this issue is the use of Non-Terrestrial Networks (NTN) which would guarantee satellite connectivity to these remote areas. On this project, a MATLAB code has been developed to simulate a satellite scenario where several NB-IoT devices will try to access the satellite using a Random Access (RA) procedure. With the help of the simulations, it has been determined which parameters of the RA are the most critical to the system capacity. Four methods are proposed which can help to equalize the capacity across the NTN cell to guarantee fairness for all User Equipment (UEs).
Internet de las Cosas de banda estrecha (NB-IoT) es una tecnología que ha estado creciendo en los últimos años y seguirá desempeñando un papel importante en la quinta (5G) y sexta (6G) generaciones de comunicaciones móviles en los próximos años. NB-IoT cubre una gran parte del segmento del mercado de la comunicación con numerosos os ejemplos que incluyen la monitorización de ciudades inteligentes, agricultura, ganadería y áreas forestales o una amplia variedad de aplicaciones industriales. Algunos de estos escenarios son lugares remotos donde no hay una red terrestre accesible y tampoco hay una gran demanda de servicio para desplegar una nueva. Una solución para abordar este problema es el uso de Redes No Terrestres (NTN) que garantizarían la conectividad satelital a estas áreas remotas. En este proyecto se ha desarrollado un código MATLAB para simular un escenario de satélite donde varios dispositivos NB-IoT intentarán acceder al satélite mediante un procedimiento de Acceso Aleatorio (RA). Con la ayuda de las simulaciones, se ha determinado qué parámetros del RA son los más críticos para la capacidad del sistema. Se proponen cuatro métodos que pueden ayudar a igualar la capacidad en la celda NTN para garantizar la equidad para todos los equipos de usuario (UE).
Internet de les coses de banda estreta (NB-IoT) és una tecnologia que ha anat creixent en els darrers anys i que continuarà jugant un paper important en la cinquena (5G) i la sisena (6G) generacions de comunicacions mòbils en els propers anys. NB-IoT cobreix una gran part del segment del mercat de la comunicació amb diversos exemples, com ara el monitoratge de ciutats intel·ligents, l'agricultura, la ramaderia i les àrees forestals o una varietat d'aplicacions industrials. Alguns d'aquests escenaris són llocs remots on no hi ha una xarxa terrestre accessible i tampoc hi ha una gran demanda de servei per desplegar-ne una de nova. Una solució per abordar aquest problema és l'ús de xarxes no terrestres (NTN) que garantirien la connectivitat per satèl·lit a aquestes zones remotes. En aquest projecte, s'ha desenvolupat un codi MATLAB per simular un escenari de satèl·lit on diversos dispositius NB-IoT intentaran accedir al satèl·lit mitjançant un procediment d'accés aleatori (RA). Amb l'ajuda de les simulacions, s'ha determinat quins paràmetres de la RA són els més crítics per a la capacitat del sistema. Es proposen quatre mètodes que poden ajudar a igualar la capacitat de la cèl·lula NTN per garantir l'equitat per a tots els equips d'usuari (UE).Medium Access Protocol for AmBC
http://hdl.handle.net/2117/402399
Medium Access Protocol for AmBC
Puerto Martínez, Miquel
This thesis will study different medium access techniques for Ambient Backscatter Communications (AmBC), which takes profit of the ambient RF signals in the environment to enable wireless devices communicate and harvest energy. The proposed protocols will be studied to improve the efficiency and reliability of AmBC systems, while minimizing interference. This will be based on random access protocol, starting from the pure ALOHA protocol, which will incorporate code techniques to enable multiple devices to share the same communication channel. The performance of the proposed protocol is evaluated through simulations in the communications laboratory. The results show that the proposed protocol achieves higher throughput and lower packet loss rate. The proposed protocol has the potential to improve the efficiency and reliability of AmBC systems, making them more suitable for a wide range of applications in the Internet of Things (IoT) and other wireless networks.
2024-02-21T10:21:17ZPuerto Martínez, MiquelThis thesis will study different medium access techniques for Ambient Backscatter Communications (AmBC), which takes profit of the ambient RF signals in the environment to enable wireless devices communicate and harvest energy. The proposed protocols will be studied to improve the efficiency and reliability of AmBC systems, while minimizing interference. This will be based on random access protocol, starting from the pure ALOHA protocol, which will incorporate code techniques to enable multiple devices to share the same communication channel. The performance of the proposed protocol is evaluated through simulations in the communications laboratory. The results show that the proposed protocol achieves higher throughput and lower packet loss rate. The proposed protocol has the potential to improve the efficiency and reliability of AmBC systems, making them more suitable for a wide range of applications in the Internet of Things (IoT) and other wireless networks.Development of an advanced automotive radar system: adaptive design, simulation, measurements, and data utilization
http://hdl.handle.net/2117/402256
Development of an advanced automotive radar system: adaptive design, simulation, measurements, and data utilization
Birnstengel, Daniel Louis
2024-02-19T17:02:14ZBirnstengel, Daniel Louis