Public display of deposited theses

DOCTORAL DEGREE IN ARCHITECTURAL DESIGN

• GONZÁLEZ TORRADO, CRISTIAN: Aspectos ambientales en la Casa Aversú de Alejandro de la Sota: El proceso del proyecto
  
  Author: GONZÁLEZ TORRADO, CRISTIAN
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN ARCHITECTURAL DESIGN
  Department: Department of Architectural Design (PA)
  Mode: Normal
  Deposit date: 01/07/2024
  Deposit END date: 12/07/2024
  Thesis director: VALOR MONTERO, JAUME
  Committee:
       PRESIDENT: RAVETLLAT MIRA, PEDRO JUAN
       SECRETARI: BLANCO GRANADO, JAIME
       VOCAL: DEVESA DEVESA, RICARDO
       VOCAL: PRIETO GONZALEZ, EDUARDO ANTONIO
       VOCAL: PASCUAL RUBIO, ANA
  Thesis abstract: This research stems from two personal interests: climate awareness and Alejandro de la Sota. Based on this premise, the aim of the research is to find the connection between environmental aspects and architecture, particularly in the project process. Thus, the research will be used to answer the question:At what moments and in what ways do environmental aspects influence architectural form?To achieve this, the present investigation will analyze Sota's Casa Arvesú, structuring its study into four parts.The first part addresses the concept of environmental aspects, establishing the vocabulary to be used. The objective of this first part is to return to the origin of the concept and review current stances regarding the impact of environmental decisions on the project, aiming to move away from clichés and redefine architecture as a medium rather than a mere support for environmental solutions. At the end of this first section, the linkage of environmental aspects with Alejandro de la Sota will also be discussed.The second part pertains to the descriptive study of the house. It involves an exploration of the national and international context at the time of its conception and construction. Additionally, the work is recognized within Alejandro de la Sota’s professional trajectory. The house is then characterized in terms of use, location, and technique throughout the project process, analyzing its various versions—sketches—until reaching its final form.The third part pertains to the analytical study. For this, a “digital twin” of the work—demolished in 1987—has been created, allowing for the quantification of environmental impacts in terms of geometry, light, systems, or energy and fluid dynamics checks of the house. The objective of the digital model is the numerical verification of what Alejandro de la Sota applied empirically. From this starting point, the solutions are initially categorized as systems and then the house is examined through five characteristic actions that relate the work to environmental aspects. The five categories are as follows:- Elevate, densify, and distance [as implantation]- Permeate, settle, and plant [as appropriation of the environment]- Contrast, flow, and culminate [as program]- Shield, project, and extend [as definition of limits]- Shade, ventilate, and retain [as passive strategies]Finally, this third part concludes with the identification and incorporation of environmental objectives in the project, using a graph that transversally relates the design process.The fourth part corresponds to the conclusions of the research. That is, the answer to the question that initiated this research in four phases:- Characterization of the environment, the user, and the problem.- Formulation of objectives.- Establishment of the internal structure of the project.- The “effort to reduce effort.”
  
• RIUS RUIZ, MARIA: Utopia Inhabited. El Barri Gaudí by Taller de Arquitectura
  
  Author: RIUS RUIZ, MARIA
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN ARCHITECTURAL DESIGN
  Department: Department of Architectural Design (PA)
  Mode: Normal
  Deposit date: 20/06/2024
  Deposit END date: 04/07/2024
  Thesis director: MÀRIA SERRANO, MARIA MAGDALENA | SALVADÓ ARAGONÈS, NÚRIA
  Committee:
       PRESIDENT: JARZOMBEK, MARK MICHAEL
       SECRETARI: JOVER FONTANALS, CRISTINA
       VOCAL: KOCKELKORN, ANNE MARIA
       VOCAL: BAILO ESTEVE, MANUEL
       VOCAL: FREDIANI I SARFATI, ARTURO
  Thesis abstract: This text is the outcome of PhD research on El Barri Gaudí (Gaudí Neighborhood), one of the collective housing prototypes that emerged worldwide during the 1960s and 1970s embodying alternative ways of communal living in cities. The notion of 'utopia inhabited' and its apparent contradiction delves into the concept of 'past futures'—radical designs that disrupted contemporary conventions, envisioned an ideal future, were eventually constructed, and continue to be inhabited today. Among them, Gaudí Neighborhood, located in Spain and designed by El Taller de Arquitectura, represents a unique case and the first built prototype of a collective housing exploration developed by the transdisciplinary team over a decade. To unpack the story of El Barri and learn from this radical built experience, this dissertation interweaves a series of complex and interconnected narratives. Similar to a documentary, it encompasses various testimonies, gathering voices from archival material, fieldwork, and interviews with the design team members, builders, and neighbors. Echoing the theatrical qualities of the place, which, far from being a setting, is an active actor in the everyday life of the community, this dissertation is structured into independent scenes with a central character: The Neighborhood. The narrative is circular and follows the typical cycle of prototyping. There are three main chapters or groups of scenes, each dedicated to a prototyping phase: everyday scenes—the learnings extracted from contrasting expectations vs. reality—, design scenes—the process of design which includes later iterations of the system—, and built scenes—the built test which does not correspond to a defined object but an open system and its combinable elements. Each group has its structure and temporal logic, constantly traveling through time and scale, and is introduced by its distinct prelude. The everyday scenes explore the conception of the neighborhood as a social utopia: an alternative way of communal living within cities. The communal vision was embodied in an organic urban fabric, a series of domestic settings that offer inhabitants a rich landscape for freedom of movement and daily encounters. The design scenes focus on analyzing how the Gaudí Neighborhood was shaped within its political, social, and economic context. During a period of significant rural exodus, this collection of scenes delves into the team's search to provide an affordable alternative to the widely spread blocks and towers, aiming to offer a variety of spaces for encounters similar to those found in Mediterranean villages. This chapter reveals how this pursuit of creating spaces characteristic of spontaneous and organic fabric led them to work with systems of geometric growth, both for this project and subsequent prototypes. The built scenes deconstructs Gaudí Neighborhood into its essential ingredients and components that define its unique and magical landscape. This series of scenes are presented as micro-narratives concerning its colors, shapes, materials, and elements. The analysis explores the role played by each component, as well as its potential inspiration and inclusion in previous and later collective housing prototypes.Behind the scenes shifts the focus from El Barri Gaudí to the Taller de Arquitectura team, the transdisciplinary collective responsible for its design. This section details the team's formation, which is closely linked to the design and construction of the Neighborhood. It delves into their connections and collaborative working methods. Finally, it reveals the ambiguity of a team sometimes presented as a radical collective, while at other times overshadowed by the figure of Ricardo Bofill.
  
• TERÉS CASTÁN, JUDITH: La repetición en un modelo habitacional urbano: De las terraced houses londinenses tardogeorgianas a los apartment houses neoyorquinos de la Era Metropolitana
  
  Author: TERÉS CASTÁN, JUDITH
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN ARCHITECTURAL DESIGN
  Department: Department of Architectural Design (PA)
  Mode: Normal
  Deposit date: 25/06/2024
  Deposit END date: 08/07/2024
  Thesis director: GONZALEZ RAVENTOS, AQUILES | GARCIA ESCUDERO, DANIEL
  Committee:
       PRESIDENT: NAVARRA, MARCO
       SECRETARI: RUBERT DE VENTOS, MARIA
       VOCAL: LLORACH HERRERO, ENRIC
       VOCAL: MORAGAS SPA, ANTONIO
       VOCAL: FREDIANI I SARFATI, ARTURO
  Thesis abstract: Domestic architecture is the backbone of the historic city due to its ambivalent capacity, expressed in urban housing models, of providing the citizens habitat at the same time of contributing to plot the texture, to structure the morphology and to define the urban identity.The research of two a priori disparate models, the that of the late Georgian London terraced houses and that of the New York apartment houses of the "Metropolitan Era", between which, however, a formal parallelism can be detected, had made possible to expose where the essentiality of architectural repetition lies in the definition of an urban housing model. The establishment of a repetitive canon is sine qua non condition for its configuration and prevalence; so that, when the repetition is objectified and applied to practice, it appears implicit in the design codes that govern the model. It is found that the repetition covers three magnitudes: the big scale, which considers the implementation in the city; the intermediate scale, of typological nature, which contains the different typologies which make up the model; and the smaller scale of detail, that focuses the attention on those architectonic elements that, with a similar design, are regularly repeated. The result is an urban landscape in which architectural uniformity prevails, given by the rhythmic sequence that provides the city with the orderly repetition of certain parameters linked to domestic architecture; A precious, although sometimes forgotten, uniformity that allows the urban maelstrom to be calmed down and that architecture that has to be highlighted singularizes.In order to an urban housing model will become established and it persisted along the time, the repetitive canon which configures it must admit a certain degree of transgression, of flexibility in front of the different social and design conditions that may be required in each one of uncountable dwelling units that conform it. An urban housing model, governed by a repetitive canon with capacity of adaptability, could even be extrapolated contextually; as the thesis demonstrates with the translation of the model of the sober and elegant London terraced houses to New York, where after an adaptation process, including a scalar magnification, it was transformed into the model defined by the magnificent apartment houses built around Central Park during the final decades of the nineteenth century and the early twentieth century.
  

DOCTORAL DEGREE IN ARCHITECTURAL, BUILDING CONSTRUCTION AND URBANISM TECHNOLOGY

• ORTEGA DONOSO, SARA ISABEL: Luz y aprendizaje dentro de los espacios educativos. Aproximaciones desde el color.
  
  Author: ORTEGA DONOSO, SARA ISABEL
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN ARCHITECTURAL, BUILDING CONSTRUCTION AND URBANISM TECHNOLOGY
  Department: Department of Architectural Technology (TA)
  Mode: Normal
  Deposit date: 01/07/2024
  Deposit END date: 12/07/2024
  Thesis director: MUROS ALCOJOR, ADRIAN | BAUTISTA PEREZ, GUILLERMO
  Committee:
       PRESIDENT: ESCOFET ROIG, ANNA MARIA
       SECRETARI: DAUMAL DOMENECH, FRANCESC DE PAULA
       VOCAL: HIGUERA TRUJILLO, JUAN LUIS
  Thesis abstract: We live in a visually demanding environment. Most of our cognitive stimuli come from visual perception. In this context of widespread dependence on the sense of sight, studying the link between artificial light and cognitive tasks is key to interacting with our memory and attention, which are increasingly accustomed to dynamic environments.The objective of this research is to analyze the disciplines that can most influence learning: cognitive sciences, architecture, and education, to:First, analyze how some light parameters can affect learning, paying special attention to the impact of light on human perception and attention and how these affect memory.Next, analyze innovative teaching methodologies and how communication and learning technologies have been incorporated into the classroom.Finally, examine the evolution of educational spaces through case studies of schools worldwide over recent decades, delving into the ways lighting and control technologies have been integrated according to the needs of educational methods within the classroom.From the analysis of the implications of light on attention and memory, we will derive the possibilities that working with different light parameters in the classroom offers us; from the analysis of teaching methodologies, we will obtain the real needs of the current classroom; and from the case study analysis, we will gain the capacity for innovation in the classroom with new lighting strategies.The intersection of these three disciplines results in three alternative uses of lighting in educational spaces, which are subjected to an experimental analysis comparing performance on a task in the proposed scenes with performance under traditional lighting conditions to conclude new classroom solutions and future lines of research.A broad narrative review methodology has been used for the case studies, providing an overview of the problem of adapting lighting to usage and users within schools. To this end, sources published in architectural journals and significant examples with educational spaces that develop innovative teaching methodologies have been reviewed, and criteria for analyzing natural and artificial lighting considered throughout these cases have been established.An experimental methodology has been used to analyze the impact of three alternative lighting proposals based on situations not previously considered. In all three cases, a control group and a study group from two public schools in the province of Barcelona are considered. These proposals arise from the observation of a lack of consideration of an important light parameter in the transmission of information in a visual environment: color.Finally, the results reflect the need to adapt and make classrooms more flexible according to the different learning needs within the classroom and the possibility of doing so with proposals not previously considered regarding color and dynamism parameters.
  
• REFALIAN, GHAZAL: A study on the implementation of a string-rewriting system for digital modeling of the islamic geometric patterns
  
  Author: REFALIAN, GHAZAL
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN ARCHITECTURAL, BUILDING CONSTRUCTION AND URBANISM TECHNOLOGY
  Department: Department of Architectural Technology (TA)
  Mode: Normal
  Deposit date: 18/06/2024
  Deposit END date: 02/07/2024
  Thesis director: COLOMA PICÓ, ELOI | MOYA SALA, JOAQUIM NARCÍS
  Committee:
       PRESIDENT: PONS VALLADARES, ORIOL
       SECRETARI: COSTA JUTGLAR, GONÇAL
       VOCAL: LIÉBANA CARRASCO, ÓSCAR
  Thesis abstract: Islamic geometric patterns represent a unique and intricate art form within Islamic art and architecture. They use a variety of precise geometric principles to tessellate surfaces. Due to their repetitive characteristics, Islamic geometric patterns provide an economic solution for the superficial ornamentation of walls, floors, ceilings, and even objects, while also offering magnificent aesthetic appeal.These patterns have played a significant role in the history of art and architecture across a wide range of regions from East to West, experiencing a peak of growth between the 8th and 14th centuries. However, in the digital era, they demonstrate challenges in their digital modeling due to their multifaceted structure. The geometric and cultural complexity of these patterns demands precise calculations and a deep understanding of their cultural context to be accurately represented in digital form. Otherwise, the resulting patterns can readily be classified as unacceptable or incorrect by experts.To address these challenges, this research pioneers the use of a computational method, which is based on formal grammar and is known as the String Rewriting Systems method (SRS), to tackle the problem of computational simulation of Islamic geometric patterns. This method employs linguistic principles such as symbols and alphabetic characters as its fundamental components, along with syntax commands and production rules, effectively representing and manipulating complex systems in a manner analogous to language processing.SRS has been used in various fields like linguistics, logic, and computer graphics. In computer-aided design, by using this method, a pattern can be generated by step-by-step conversion of a simple initial shape to the final structure. The utilization of strings and symbols, instead of shapes or mathematical formulas, provides a direct link between geometry and machine language and facilitates their communication. Moreover, it provides new formal potentials, previously unavailable in conventional methods.The methodology of this research involved collecting samples of historical Islamic geometric patterns (IGPs) and examining existing drawing techniques. These inputs were analyzed and compared with the requirements for applying SRS to model patterns. This comparison led to the development of a new morphological approach to IGPs, which was utilized to create an algorithm for constructing IGPs using SRS.The outcome of this research is a comprehensive solution for generating and manipulating Islamic geometric patterns in a digital environment. A CAD toolkit was developed as an add-on for Grasshopper, a plug-in for RHINO3D software. This widely accessible platform for architects and designers ensures consistent availability. In addition, it eliminates the need to design an entirely new platform and allows users to combine this toolkit with other existing tools within the primary software. This integration thereby expands its potential applications.To assess the effectiveness of the method in addressing the aforementioned challenges, experimental workshops were designed to test the toolkit, demonstrating the impact of the introduced methodology, achieving an 85% success rate in pattern production compared to 55% with conventional methods. Participant feedback validated the efficacy of SRS, signaling promising avenues for future research and development. Additionally, 22 IGPs were successfully restored and modeled, now publicly available in the toolkit.This method bridges the gap between digital technologies and traditional heritage, offering a systematic approach to the digital modeling of Islamic geometric patterns (IGPs). It facilitates and accelerates the modeling of numerous patterns while staying true to the original designs, eliminating the need to compromise on geometrical details.
  

DOCTORAL DEGREE IN ARTIFICIAL INTELLIGENCE

• ZHANG, XIAO: Soft computing strategies for resolving key data challenges in organ transplantation
  
  Author: ZHANG, XIAO
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN ARTIFICIAL INTELLIGENCE
  Department: Department of Computer Science (CS)
  Mode: Normal
  Deposit date: 27/06/2024
  Deposit END date: 10/07/2024
  Thesis director: NEBOT CASTELLS, MARIA ANGELA
  Committee:
       PRESIDENT: RIBAS RIPOLL, VICENTE JORGE
       SECRETARI: VELLIDO ALCACENA, ALFREDO
       VOCAL: ARMENGOL VOLTAS, EVA
  Thesis abstract: In the field of organ transplantation, a critical gap exists: the availability of organs falls far short of the demand, resulting in numerous recipients dying before they can receive a transplant. The complexity of this field extends beyond surgical procedures, encompassing the challenges of matching organs to patients and ensuring effective post-operative care-both of which are crucial for the patients' survival and quality of life. Recently, artificial intelligence (AI) technologies, particularly machine learning (ML) and deep learning (DL) methods, have shown great potential in enhancing the accuracy of organ matching and in managing post-transplant patient risks more effectively. However, the application of these methods faces several challenges, including issues with model interpretability, data imbalance, and limitations due to small dataset sizes and insufficient labeled samples. This thesis focuses on the application and exploration of soft computing techniques in the analysis and modeling of organ transplant data. By integrating soft computing techniques with ML models, this study aims to develop new computational approaches to address the key data challenges in the organ transplant domain. Through this research, we aim to provide deeper insights into data analysis and modeling in organ transplant scenarios, thereby offering more accurate and personalized decision support for doctors and patients.This thesis demonstrates how enhancing interpretability in ML models for assessing organ transplant risks can be achieved, particularly by addressing gaps in understanding the impact of features over follow-up time and across different patient subgroups. The Extreme Gradient Boosting (XGBoost) model is shown to outperform traditional risk scores and other ML models across various follow-up periods. Using SHapley Additive exPlanations (SHAP), this thesis provides detailed insights into how specific features dynamically affect different patient subgroups over these periods, thereby enhancing both global and subgroup-specific interpretability in the context of organ transplantation.To tackle the prevalent challenge of data imbalance in the field of organ transplantation, two novel rule-based methods, Ad-RuLer and ARUST, are proposed. Ad-RuLer improves the representation of minority classes through iterative rule comparison. Building on this, ARUST further refines the segmentation of the sample space through density peak clustering (DPC), enhancing the granularity of data synthesis. Simultaneously, it enhances the detection and elimination of overlapping and noisy samples, thus improving the classification performance of the synthesized samples. These methods outperform traditional resampling techniques in predicting de novo solid malignancies in post-liver transplant recipients, showcasing their robust ability to handle imbalanced data effectively.Addressing the issues of small sample sizes and insufficient labeled samples, the thesis presents a novel neuro-fuzzy system, D-DMR-FBLS. This system integrates Deep Belief Networks (DBN) and Takagi-Sugeno-Kang (TSK) systems within the Broad Learning System (BLS) framework to enhance representation learning capacity during the unsupervised training phase (UTP). Additionally, two types of graph-based manifold regularization strategies are proposed for the system: sample-based and feature-based. Adapted into a semi-supervised learning method, it leverages the similarities among samples, including unlabeled ones, and the correlations between features within the fuzzy feature space, further improving the model's predictive performance in scenarios of small sample sizes and insufficient labeled samples.
  

DOCTORAL DEGREE IN AUTOMATIC CONTROL, ROBOTICS AND VISION

• SHEN, JIANXIONG: Incorporating Uncertainty into Neural Rendering for Interpretable 3D Modeling
  
  Author: SHEN, JIANXIONG
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN AUTOMATIC CONTROL, ROBOTICS AND VISION
  Department: Institute of Robotics and Industrial Informatics (IRI)
  Mode: Normal
  Deposit date: 21/06/2024
  Deposit END date: 05/07/2024
  Thesis director: MORENO NOGUER, FRANCESC D'ASSIS | RUIZ OVEJERO, ADRIÀ
  Committee:
       PRESIDENT: HARO ORTEGA, GLORIA
       SECRETARI: SÁNCHEZ RIERA, JORDI
       VOCAL: PORZI, LORENZO
  Thesis abstract: 3D scene modeling refers to the process of creating a digital representation of real-world environments for better understandingand further manipulation. Early works involved the utilization of pure structured representations such as meshes or voxel grids.The recent success of deep learning has enabled implicit representations of the scenes using deep neural networks, whichtypically are of high efficiency in automatically learning from data and handling high-resolution complex scenes. However, theseimplicit representations are often not explainable with an automatic feature learning process, hindering their further applicationsin practical scenarios. For example, to reduce potentially catastrophic failures in high-risky fields such as healthcare orautonomous driving, the uncertainty associated with the modeling results must be included into the decision-making process.Aiming at more interpretable 3D scene modeling, this thesis explores various methods to quantify the uncertainty in the processof 3D scene modeling that utilizes implicit neural representation. Firstly, we propose stochastic neural radiance fields to model aprobabilistic framework for capturing the uncertainty on rendered RGB images and estimated depth from the learned scenemodel, with a drastically reduction of model complexity compared to previous Bayesianbased probabilistic methods. To enhanceits capability of handling more complex scenes with varying geometry and appearance, we next extend the modelexpressiveness by incorporating a flow-based generative models to automatically learn arbitrarily complicated densitydistributions in a flexible manner. While these methods are able to achieve accurate 3D modeling with reliable uncertaintyestimation, they still suffer from the time-consuming optimization process, impeding their further applications in practicalscenarios. Therefore, we then explore a hybrid representation with incorporated voxel grids to explicitly learn the volumetricdensity of each 3D position, dramatically improving the model efficiency in both optimization and inference. So far, the currentframework is inherently designed for estimating predictive uncertainty for the areas of the scene that can be observed in thetraining images, and fail to output reliable uncertainty involving unobserved scene contents such as occlusion for roboticexploration in unknown experiments. For this purpose, we finally propose to introduce a 3D uncertainty field based on the trainedmodel, exhibiting consistently high uncertainty for predictions from those unobserved scene regions.In summary, we design an efficient and reliable probabilistic framework for 3D scene modeling to quantify the comprehensiveuncertainty associated with the predictions both from observed and unobserved scene regions. This framework provides a solidtool for uncertainty-aware decision-making and analysis in a variety of applications that rely on 3D scene modeling.
  
• SUÁREZ HERNÁNDEZ, ALEJANDRO: New methods for bridging symbolic-geometric reasoning, addressing uncertainty and action learning in task planning for robotics
  
  Author: SUÁREZ HERNÁNDEZ, ALEJANDRO
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN AUTOMATIC CONTROL, ROBOTICS AND VISION
  Department: Institute of Robotics and Industrial Informatics (IRI)
  Mode: Normal
  Deposit date: 21/06/2024
  Deposit END date: 05/07/2024
  Thesis director: TORRAS GENIS, CARMEN | ALENYÀ RIBAS, GUILLEM
  Committee:
       PRESIDENT: ANGULO BAHON, CECILIO
       SECRETARI: MORENO RIBAS, ANTONIO
       VOCAL: UGUR, EMRE
  Thesis abstract: The physical world exhibits a wide range of obstacles to the application of robotics in a large number of tasks. Scripted behavior and/or teleoperated programs are still used in many industries (e.g. car assembly) because they are robust and reliable. However, their scope is limited and falls short in less controlled environments. We explore the possibilities of task planning, or Artificial Intelligence (AI) planning, for solving tasks in a more flexible, non-scripted way. In contrast to motion planning, AI planning focuses on high-level decision-making, rather than on concerns such as computation of trajectories and dynamics. Task planning is a very powerful tool for virtual situated agents (e.g. videogames or web services). One of its main advantages is that it allows deliberative, explainable, and adaptive behavior as long as a reliable model of the environment is available. However, the physical world presents some challenges that make the application of AI planning more difficult. This thesis has the following objectives, each one tied to a different challenge: (O1) integration of AI and motion planning; (O2) handling the unintended effects of actions taken by the robot; (O3) performing tasks even when the robot is not aware of all the relevant information; and (O4) automatic learning of action models to avoid the need for handcrafted ones. Our first contributions revolve mainly around objectives O1 to O3, which involve planning and acting. We propose hierarchical paradigms of planning, exploitation of topological properties of a problem for simplifying Markov Decision Processes, and planning alongside physical simulators to minimize the risk of unintended effects. The second part of our contributions focuses on O4, and consists of different algorithms for learning and recognizing STRIPS action schemata. Published results and findings are provided to support each contribution, alongside examples from manipulation scenarios, such as automatic disassembly of electromechanical devices, and socially assistive interactions.
  

DOCTORAL DEGREE IN CHEMICAL PROCESS ENGINEERING

• VALLÉS NEBOT, VÍCTOR: Recovery of critical raw materials from saltwork by integration of sorption processes
  
  Author: VALLÉS NEBOT, VÍCTOR
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN CHEMICAL PROCESS ENGINEERING
  Department: Department of Chemical Engineering (EQ)
  Mode: Article-based thesis
  Deposit date: 21/06/2024
  Deposit END date: 05/07/2024
  Thesis director: CORTINA PALLAS, JOSE LUIS | LÓPEZ RODRÍGUEZ, JULIO
  Committee:
       PRESIDENT: SASTRE REQUENA, ANA MARIA
       SECRETARI: SIRÉS SADORNIL, IGNACIO
       VOCAL NO PRESENCIAL: KOLODYNSKA, DOROTA
  Thesis abstract: Nowadays, the increasing demand for raw materials caused by population growth and industrialisation moved the European Union (EU) to improve resource efficiency and as result the Critical Raw Materials/Strategic Raw Materials (CRMs/SRMs) list was created, which includes those raw materials with economic importance and/or supply risk for the EU. Seawater mining has emerged in the last years as a potential solution to resource scarcity. However, many valuable minerals present in seawater are considered Trace Elements (TEs) due to their low concentration (mg/L or µg/L), demanding energy-intensive and selective methods to extract them, which in the end can compromise their recovery from a technical and economic point of view. Research on Trapani saltworks (Sicily, Italy) and on the ones located at the Mediterranean basin revealed that the natural evaporation process in salt crystallisation ponds concentrates seawater components, producing a bittern (brine generated after NaCl(s) crystallisation in saltworks) with a concentration 40 times higher than seawater. Consequently, these bitterns, typically considered waste, present a potential viable source for extracting TEs identified as CRMs/SRMs.Evaluation of TEs extraction from bitterns using various sorbents demonstrated promising results. Whereas N- methylglucamine sorbents (S108, CRB03, CRB05) were highly efficient in terms of sorption/desorption for recovering B, Co, Ga and Ge, those containing aminophosphonic groups (S940, IRC747) effectively targeted Co, Ga and Sr. Similarly, one impregnated sorbent (MTX8010) exhibited outstanding selectivity towards Ga. Sorption mechanisms, based on bitterns’ speciation and sorbents’ functional group properties, were postulated to explain the extraction and selectivity patterns observed.Considering that some of the TEs were at concentrations of 0.5 mg/L in the bitterns, lab-scale packed-bed columns of these sorbents showed effective retention of TEs and high sorption capacities (e.g., 11 mg/g B for CRB03, >2.7 mg/g Co for S940, 0.9 mg/g Ga for MTX8010). The retained TEs were subsequently recovered from the sorbents by acidic desorption, resulting in a TEs-rich stream with a concentration of B up to 11-fold higher than seawater when CRB03 was used, 710-fold higher on Co when S940 was employed or 154-fold higher on Ga when MTX8010 was utilised, thus enhancing their extractionfeasibility.Diffusion Dialysis (DD) facilitated the recovery of the excess of acid used during the elution of the packed-bed resins. Flow rate and water to acid flow rate ratio effect were evaluated on acid recovery and the losses of metals. Under the optimum conditions (water to acid ratio of 1 working at 2.11 L/(m2 ·h)), 60% HCl was recovered while metallic TEs cations (i.e., Co, Ga, Sr) were completely rejected (>99%). However, certain leakage occurred for B (34%) and Ge (22%) due to their presence as neutral species (H3BO3(aq) and Ge(OH)4(aq)).The TEs could be recovered via precipitation/crystallisation from the packed-bed column eluates after treating them with DD.Co(OH)2(s), GaOOH(s) and SrSO4(s) were successfully precipitated using NaOH as precipitant agent. H3BO3(s) was crystallised from an eluate collected after processing a real bittern with N-methylglucamine packed-bed columns, leading to 70% of B recovery. In addition, a reaction with tannic acid allowed to recover 92% of Ge.Integration of sorption/desorption, DD, and precipitation/crystallisation processes enabled the definition of potential successful recovery schemes of several TEs minerals from bitterns, showcasing their potential as alternative CRMs/SRMs sources.
  

DOCTORAL DEGREE IN COMPUTER ARCHITECTURE

• SHARMA, ROBIN KUMAR: Parallelizing recurrent neural network and variants
  
  Author: SHARMA, ROBIN KUMAR
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN COMPUTER ARCHITECTURE
  Department: Department of Computer Architecture (DAC)
  Mode: Normal
  Deposit date: 01/07/2024
  Deposit END date: 12/07/2024
  Thesis director: CASAS GUIX, MARC
  Committee:
       PRESIDENT: QUINTANA ORTI, ENRIQUE SALVADOR
       SECRETARI: ARMEJACH SANOSA, ADRIÀ
       VOCAL: VALLEJO GUTIÉRREZ, ENRIQUE
  Thesis abstract: Recurrent neural networks (RNN) have succeeded remarkably in various domains, such as Automatic Speech Recognition, Sentiment Analysis, time-series prediction, and Machine Translation. Despite their versatility, RNN poses significant challenges due to their complex internal structures, which impede the effective use of model parallelism. This often leads to a reliance on data parallelism to accelerate RNN performance. Furthermore, RNN demands extensive computational resources due to their large parameter counts. This doctoral research proposes innovative High-Performance Computing (HPC) strategies to optimize RNN deployment on CPUs, enhancing their efficiency in resource-limited settings. Through algorithmic improvements and memory-efficient techniques, this work seeks to maximize the potential of parallel computing for RNN, thereby transforming AI parallel system landscapes.This thesis introduces "Wavefront-Parallelization" (W-Par), which integrates model parallelism into unidirectional RNN to enhance inference and training on CPUs. W-Par utilizes fine-grained pipeline parallelism through wavefront computations, which are particularly effective for multi-layer RNNs on multi-core CPUs. These techniques allow for efficient workload distribution across parallel tasks while managing the dependencies of each RNN cell. Empirical results show that W-Par significantly outperforms existing implementations, achieving speed-ups of up to 6.6x times on contemporary multi-core CPU architectures, and maintains robust performance across various core and memory configurations without requiring source code modifications.Additionally, the thesis presents "Bidirectional-Parallelization" (B-Par), a novel execution model for Bidirectional Recurrent Neural Networks (BRNN). B-Par leverages inherent data and control dependencies in forward and reverse-order RNN in BRNN, dividing workloads efficiently across parallel tasks without needing layer-specific synchronization barriers. Tests on the TIDIGITS speech database and Wikipedia dataset demonstrate that B-Par significantly exceeds the performance of leading frameworks like TensorFlow-Keras and PyTorch, with speed-ups of up to 2.34x and 9.16x times, respectively, while maintaining accuracy.Finally, this thesis introduces the "Semi-Bidirectional RNN" (SB-RNN), a novel architecture that synergistically integrates the strengths of both unidirectional and bidirectional RNN. SB-RNN maintains the parameter count of unidirectional RNN while incorporating backward connections across layers to enhance the capability for information retention. This architecture enables SB-RNN to match and potentially exceed the accuracy of unidirectional RNN and bidirectional RNN (BRNN) across both CPU and GPU environments. Specifically, on the sentiment analysis task of the Stanford Sentiment Treebank (SST) dataset, SB-RNN demonstrates superior performance with 56.61% fewer parameters than their unidirectional counterparts, leading to a significant reduction in training time by 52.94%.Overall, this thesis introduces three advanced techniques: W-Par, B-Par, and SB-RNN - that significantly improve the efficiency and performance of RNN and BRNN models on multi-core CPUs and GPUs, facilitating enhanced processing across various applications without extensive code alterations.
  

DOCTORAL DEGREE IN COMPUTING

• YUN, HAORAN: Real-time Avatar Animation Synthesis in Virtual Reality
  
  Author: YUN, HAORAN
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN COMPUTING
  Department: Department of Computer Science (CS)
  Mode: Normal
  Deposit date: 25/06/2024
  Deposit END date: 08/07/2024
  Thesis director: PELECHANO GOMEZ, NURIA | ANDUJAR GRAN, CARLOS ANTONIO
  Thesis abstract: The rapid development of consumer-grade virtual reality (VR) systems has changed how users interact within virtualenvironments. This opens up exciting opportunities for various domains, such as education, social interactions, and gaming.However, the current VR technology often provides only head and hand tracking, which significantly restricts the potential foroffering realistic full-body animation to create immersive VR experiences. This is especially problematic for activities in VR thatrequire users to use their whole body, like human factors engineering, fitness, rehabilitation, and training. In addition to the sparsetracking limitation, the field faces many other significant challenges, including the real-time performance requirement, not havingenough VR motion datasets, and challenges regarding how we evaluate these technologies.Therefore, as technical contributions, this thesis presents two novel data-driven solutions for full-body avatar animation in VRfrom sparse tracking data. The first method breaks down the animation process into three parts: body orientation, lower body, andupper body, solving them by different modules. A lightweight neural network is used to estimate the body direction from theHead-Mounted Display (HMD) and controllers. Then, customized Motion Matching finds the motion from a dataset that bestmatches the user's movement, avoiding fixed walking animations. Inverse kinematic solvers are used to animate the upper bodyand to refine the final pose. The second method uses a novel deep-learning framework to reconstruct the full-body motion from thepositions and rotations of six tracking devices without separating different body parts. Once trained, the model takes live datafrom VR devices and outputs an accurately animated full-body avatar that the user can control as their physical body. We alsocaptured several datasets featuring interaction movements and locomotion most relevant to VR avatar animation which have beenmade publicly available.In addition to its technical advancements, this thesis also contributes to a better understanding of user experiences in virtualreality through user studies. Two in-depth user studies have been conducted to evaluate the impact of animation quality andcollision feedback on how users perceive and interact within VR environments. Through objective metrics, subjectiveassessments, and interviews, insights were gained into improving avatar animation and virtual interactions. Findings include thatsparse trackers with high-quality inverse kinematics can match the embodiment of advanced motion capture suits for certain tasksbut fall short for tasks requiring accurate poses. Moreover, our studies indicate that achieving a realistic interaction with othervirtual humans requires not only advanced animation methods but also believable collision feedback, such as inducing participantsto expect that a physical bump against a real person might occur. Together, this research has advanced the state-of-the-art infull-body avatar animation in VR and deepened our understanding of potential improvements in this field.
  

DOCTORAL DEGREE IN CONSTRUCTION ENGINEERING

• MAZZATURA, ISABELLA: An integral methodology for the special inspection of concrete bridges with bonded post-tensioned cables.
  
  Author: MAZZATURA, ISABELLA
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN CONSTRUCTION ENGINEERING
  Department: Department of Civil and Environmental Engineering (DECA)
  Mode: Change of supervisor
  Deposit date: 20/06/2024
  Deposit END date: 04/07/2024
  Thesis director: CASAS RIUS, JUAN RAMON | CAPRILI, SILVIA | SALVATORE, WALTER
  Committee:
       PRESIDENT: BARTOLI, GIANNI
       SECRETARI: MORELLI, FRANCESCO
       VOCAL: SAS, GABRIEL
       VOCAL: STRAUSS, ALFRED
       VOCAL: SILVA CARVALHO CAMPOS E MATOS, JOSE ANTONIO
  Thesis abstract: The inspection and the consequent assessment of prestressed concrete bridges with post-tensionedcables is a complex issue. The system was conceived in an era in which there were no doubts about thedurability of the concrete. Moreover, avoiding the cracking of the concrete meant the total protection ofthe high-strength steel cables against corrosion. There was also the additional protection grouting layer.Thus, the first designers did not worry about providing inspection methodologies. When the first suddencollapses occurred, engineers questioned the possible damage mechanisms. Nowadays, a consolidatedinspection method is still missing. The research focuses on two major subjects, i.e. how to perform theinspection and how many tests to carry out for a reliable result. Some Non-Destructive Tests (NDTs)were selected and tested both in the laboratory and in situ. The required characteristics must be theeasiness of application, the low effort in terms of time and money, and their reliability. The lab campaignswere executed mainly to calibrate the procedures and collect data. Ground Penetrating Radar (GPR) wasselected for individuating tendons, while Ultrasonic Tomography (UT) for locating void regions. Even ifthe absence of grout is not sufficient for the corrosion trigger, the hypothesis of correspondence of voidand corrosion is conservative. Seven specimens were realised to state the accuracy of the twotechniques. The GPR was reasonably accurate in locating the tendons. It was possible to obtain a Normaldistribution with a good fit to the data, with parameters mean of 0.702 cm and a standard deviation of1.52 cm, describing the error. The UTs’ reliability was assessed by the Probability of Detection model. Itresulted that, stated 95% confidence and 90% probability of detection, the system can detect voidslonger than 30 cm. The remaining prestress estimation allows to state the losses' entity and indirectlyidentify damage. The X-ray diffraction technique was selected for the stress estimation on the tendons.During the first laboratory campaign, only wires were tested. The laboratory results showed that themethodology is accurate if the residual stresses are known and the load level does not influence theresults. Given that outcome, another campaign was performed, and it showed that the variations inresidual stresses within similar sample groups are not negligible (for wires residual stresses from -30 to-190 MPa, and strands from 120 to 240 MPa). If the residual stresses are considered, a constant relativeerror in the order of 20% is obtained. The saw-cut method was chosen for estimating the stresses onthe concrete. The samples tested in the laboratory were very simple, aiming to assess the accuracy ofthe procedure. The results were not satisfactory; out of twelve tests, only in one case did the testcorrectly estimate the actual stress. The in-situ testing campaign was executed on seven similarstructures to assess the applicability of the NDTs in real conditions. The research deals also with the issueof the number of samples at different levels of evaluation, based on the procedure proposed by theFederal Highway Administration (FHWA), adapted to the Italian Standards, and to different levels ofconfidence for more accurate evaluations. Finally, an overall inspection protocol is thus proposed, tryingto summarize the main outcomes of the research, by the definition of a multi-phase methodology for thespecial inspection in the assessment of PT bridges. The proposal follows the same philosophy as theItalian Guidelines, providing more effort (and tests) for the riskiest structures while tryingto optimise resources. The last part of the thesis concerns the accurate assessment, i.e. how to obtainand use inspection results for bridges requiring accurate evaluation because the special inspection didnot provide a conclusive classification of the actual safety of the bridge.
  

DOCTORAL DEGREE IN GEOTECHNICAL ENGINEERING

• AVILES MURCIA, LUIS ANGEL: Numerical modelling of unsaturated soils with the material point method
  
  Author: AVILES MURCIA, LUIS ANGEL
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN GEOTECHNICAL ENGINEERING
  Department: Department of Civil and Environmental Engineering (DECA)
  Mode: Normal
  Deposit date: 18/06/2024
  Deposit END date: 02/07/2024
  Thesis director: PINYOL PUIGMARTI, NURIA MERCE
  Committee:
       PRESIDENT: LIZCANO, ARCESIO
       SECRETARI: OLIVELLA PASTALLE, SEBASTIAN
       VOCAL: CUOMO, SABATINO
  Thesis abstract: The Material Point Method (MPM) is presented as an advanced numerical method used to simulate geotechnical problems subjected to large deformations and soil-structure interaction problems, such as landslides, penetration issues, and collapse of geotechnical structures. Its main advantage is the ability to simulate large movements without the mesh-related problems typical of the standard Finite Element Method (FEM).Simulating different material states and external conditions is a challenging aspect in geotechnical engineering. This process requires a detailed analysis of how to address phenomena such as rainfall and water flow through the soil. Additionally, it is crucial to incorporate material constitutive relationships to properly replicate or predict stress-strain behavior.The modelling of unsaturated soils is necessary in various fields, especially in geotechnical works employing compacted materials such as embankments, dams, fills, and natural slopes. The study and analysis of these soils requires modelling the entire process of the work, from its construction, if applicable, to post-failure behavior.Until now, simulating the unsaturated state of materials has mainly relied on FEM, assuming small deformations at a theoretical level. This involves developing coupled mechanical and hydraulic constitutive models and utilizing advanced numerical tools. This Thesis presents advancements in modelling unsaturated soils in large deformation problems through the development of the MPM. The formulation of a single-point, multi-phases has been used, showing its potential for improving the accuracy and reliability of geotechnical simulations involving large deformations.Firstly, the developments for simulating boundary conditions for unsaturated soils (imposed flow and seepage condition) are presented. These conditions are complex, particularly since the boundary where they are applied is not fixed and must be determined during the calculations. Validation of these conditions is consistently performed using a soil column subjected to infiltration flow. The results are compared with those obtained from a widely used finite element code, giving similar results.Next, for the first time, the modelling of wetting collapse in MPM is proposed. A constitutive model is implemented to capture the key characteristics of unsaturated material behavior. This model is based on critical state theory and employs two alternative formulations, depending on whether Bishop's stress or net stress is used, along with suction. Validation is conducted through conventional laboratory tests.The loss of volume due to wetting is analysed, a characteristic associated with unsaturated soils with an open structure. The loss of suction in the material leads to a loss of strength and, consequently, the possibility of failure. This is illustrated by a physical test conducted in a centrifuge on a slope subjected to a rising water table from the base, which results in the evolution of the saturation front, collapse-induced deformations, and ultimately a landslide when critical saturation conditions are reached.Finally, one of the objectives of this Thesis is to model the entire process of a geotechnical project. To achieve this, a numerical scheme was implemented to simulate construction stages, representing a significant advancement in the application of the MPM for analysing material behavior during the pre-failure stage.Modelling of the construction process is crucial because it not only represents the construction process and the displacements occurring during it, but also establishes the initial state of the structure for subsequent analysis stages. The proposed scheme effectively captures deformations and failure processes during construction, unlike conventional methods based on Eulerian formulations, which cannot accommodate large material displacements during construction
  
• ZHOU, YUNFENG: Multiphase fluid flow in heterogeneous / anisotropic deformable geomaterials
  
  Author: ZHOU, YUNFENG
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN GEOTECHNICAL ENGINEERING
  Department: Department of Civil and Environmental Engineering (DECA)
  Mode: Normal
  Deposit date: 21/06/2024
  Deposit END date: 05/07/2024
  Thesis director: RODRIGUEZ DONO, ALFONSO | OLIVELLA PASTALLE, SEBASTIAN
  Committee:
       PRESIDENT: CARRERA RAMIREZ, JESUS
       SECRETARI: SAALTINK, MAARTEN WILLEM
       VOCAL: DE SIMONE, SILVIA
  Thesis abstract: Deep geological disposal is currently the preferred option for managing long-lived and heat-emitting radioactive waste. This method involves confining the waste for an extremely long period, potentially severa! hundreds of thousands of years, by placing it in a deep geological formation. Therefore, understanding the migration of gases produced by metal corrosion, microbial degradation, and radiolysis of water within a deep geological repository is crucial for assessing the repository's performance and long-term evolution. The primary objective of this study is to enhance the predictive capability of numerical models for understanding the processes and mechanisms of fracture initiation and growth in claystones with inherent heterogeneity and anisotropy, especially under rapid gas overpressure increases. To achieve this, geostatistics were applied into the finite element method software CODE_BRIGHT to generate a spatially correlated heterogeneous field for porosity. In CODE_BRIGHT, properties such as Young's modulus, intrinsic permeability, and thermal conductivity are functions of porosity, making these properties heterogeneous as well. Additionally, the solution of Eshelby's problem was implemented to introduce both heterogeneity and anisotropy in Young's modulus. lt is crucial to select the appropriate constitutive laws to effectively introduce anisotropy in intrinsic permeability. Simulations were conducted to validate the proposed framework, and sensitivity analyses were performed at both laboratory and in-situ experiment scales to assess the model's performance. The results indicate that the proposed model successfully captures the main observed features of heterogeneous anisotropic claystones during gas injection, including the formation of preferential gas pathways.
  

DOCTORAL DEGREE IN MECHANICAL, FLUIDS AND AEROSPACE ENGINEERING

• TERRONES FERNÁNDEZ, INÉS: Innovative Modular Pour Plating Microbiology Culture Media Technology
  
  Author: TERRONES FERNÁNDEZ, INÉS
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN MECHANICAL, FLUIDS AND AEROSPACE ENGINEERING
  Department: Department of Fluid Mechanics (MF)
  Mode: Normal
  Deposit date: 27/06/2024
  Deposit END date: 10/07/2024
  Thesis director: GAMEZ MONTERO, PEDRO JAVIER | CASTILLA LOPEZ, ROBERTO
  Committee:
       PRESIDENT: GRILLO DOLSET, MARIA JESUS
       SECRETARI: ESCALER PUIGORIOL, FRANCESC XAVIER
       VOCAL: VERNET PEÑA, ANTON
  Thesis abstract: The current methods used in microbiological quality analysis rely on the use of traditional and manual methods, such as the pour plate method. The need to perform more analyses of different matrices makes it necessary to improve these methods. Thus, this thesis focuses on the design of systems in which the manual work of laboratory technicians is reduced. In order to meet the main objective, this thesis has been separated into two subjects, i.e., microbiology and mechanical and fluid engineering. The union of these two fields is essential for the achievement of this need.The starting point for microbiology focused on the separation of culture media into their different phases. This separation would allow both easier maintenance of the media and a simpler dosage of the same. In addition, studies and the establishment of a protocol for sterilizing the culture medium in a microwave oven were also carried out.By means of Computational Fluid dynamics (CFD), studies were carried out to design a static homogenizer so that the culture medium would be completely mixed when pouring onto the plates. In addition, a device was designed with which the dosing of the culture medium, once separated into its components, was possible. The favorable results obtained in both microbiological and engineering studies demonstrate that a pour plate method automation system is possible. Hence, the development of this new line of research will continue, as this study is only the starting point for the development of such a system.
  

DOCTORAL DEGREE IN NATURAL RESOURCES AND THE ENVIRONMENT

• CUBIDES PAEZ, DAVID FERNANDO: Biological technologies for nitric oxide abatement
  
  Author: CUBIDES PAEZ, DAVID FERNANDO
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN NATURAL RESOURCES AND THE ENVIRONMENT
  Department: Department of Mining, Industrial and ICT Engineering (EMIT)
  Mode: Normal
  Deposit date: 01/07/2024
  Deposit END date: 12/07/2024
  Thesis director: JUBANY GUELL, IRENE | GAMISANS NOGUERA, XAVIER
  Committee:
       PRESIDENT: RAMÍREZ MUÑOZ, MARTÍN
       SECRETARI: DORADO CASTAÑO, ANTONIO DAVID
       VOCAL: MARTIN SANCHEZ, MARIA JOSE
  Thesis abstract: Tackling the widespread challenge of air pollution, particularly the reduction of nitrogen oxides (NOx), is critical to improving public health and environmental quality in Europe. Annually, air quality-related problems contribute to some premature deaths across Europe, underlining the urgent need for effective pollution control strategies. This PhD thesis explores innovative biotechnological processes for NOx removal, focusing on the potential of biological treatments as sustainable and cost-effective alternatives to conventional methods.Conducted at the Department of Mining, Industrial Engineering and ICT of the Universitat Politècnica de Catalunya (UPC) and the Eurecat Water, Air and Soil Technology Unit, this research is part of a collaborative effort to bridge the gap between academic studies and industrial applications. The thesis investigates the effectiveness of ionic liquids (ILs) and non-aqueous phase liquids (NAPs) as mass transfer vectors to enhance nitrogen oxide (NO) solubility and in turn improve NO bioavailability for microbial degradation. These vectors have the potential to revolutionize the design and operation of biological treatment systems of low water-soluble gases by improving their efficiency and scalability.The thesis systematically reviews existing NO control technologies to lay the groundwork for the introduction of bio-based alternatives. It delves into the selection and optimization of materials and methods, with emphasis on experimental designs that facilitate robust and reliable results. By improving mass transfer from the gas to the liquid phase, the research aims to address one of the main limitations faced by current biological treatments when treating hydrophobic gases such as NO.This research was supported by several academic and government grants, reflecting its importance and potential impact. The results are expected to open a door to the study of a new industrial technology, providing a solid scientific basis for further research into new alternatives for gas treatment.
  

DOCTORAL DEGREE IN NETWORK ENGINEERING

• SSEMAKULA, JOHN BOSCO: Contribution to Network Management of Beyond 5G Networks: Management and Orchestration Architecture to Support Microservice-based Services
  
  Author: SSEMAKULA, JOHN BOSCO
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN NETWORK ENGINEERING
  Department: Department of Network Engineering (ENTEL)
  Mode: Normal
  Deposit date: 18/06/2024
  Deposit END date: 02/07/2024
  Thesis director: GORRICHO MORENO, JUAN LUIS
  Committee:
       PRESIDENT: BALIOSIAN, JAVIER ERNESTO
       SECRETARI: HESSELBACH SERRA, XAVIER
       VOCAL: RUBIO LOYOLA, JAVIER
  Thesis abstract: The thesis has contributed to the research on network management for the provisioning of future services, which we refer to as the Future Service Deployment Problem (FSDP). In this thesis, different deployment techniques facing the FSDP are investigated to achieve solving the problem in a cost-effective and resource-efficient manner under different constraints while meeting the requirements of selected representative future applications, as a problem to be solved in a time frame appropriate for the particular working scenario considered. Depending on the operational behaviour and network and resource requirements, four representative applications have been identified and classified into two different categories: i) Forwarding Applications (FAs), such as some smart city applications, and ii) Closed Loop Applications (CLAs), such as Virtual Reality (VR) or Vehicle Collision Avoidance (VCA).The FSDP can generally be viewed as consisting of two interlinked sub-problems that should always be addressed in a coordinated manner in order to achieve adequate deployments. In particular, the two sub-problems are: i) Where to deploy a service; and ii) How to deploy the service. For the former, we mainly consider a given set of coordinated and cooperating resources that form the substrate network. The solution to the problem is to determine an optimal compute node or subset of compute nodes and the associated links between those nodes to allocate the various microservices of a service request. While the latter deals with the appropriate provisioning of microservices that takes into account the various key attributes arising from the network state and QoS requirements of the service requests, in a resource-efficient and cost-effective manner, to be executed in a reasonable time while meeting the associated constraints. In the following, a brief mentioning of the different chapters of the thesis is given according to the proposed technique and the considered working scenario for each of them: chapters 1 and 2 present the thesis objectives, the working scenarios, the state-of-the-art regarding the addressed problem, the description of considered application use cases, a mathematical formulation of the problem, and a summary of proposed techniques to solve it. In Chapter 3, an Optimization Deployment Algorithm (ODA) is presented to obtain near-optimal deployment solutions in a reasonable amount of time, a Multiaccess Edge Computing (MEC) working scenario is assumed, where different MEC State Features (MSFs) are used. Chapter 4introduces a deployment technique that uses a combination of heuristic and artificial intelligence (AI) to flexibly and optimally fulfil a wider range of requirements and to accommodate the various service requests on a shared network. Chapter 5 explores the use of a more complex working scenario: a hybrid edge-cloud system, where resources at the edge tier are geographically distributed.Moreover, an AI technique based on reinforcement learning (RL) is proposed. In addition, a new heuristic is proposed for the allocation of microservices within the edge network. Chapter 6 presents a metaheuristic algorithm, namely a customized Genetic Algorithm, tailored for the deployment of future microservice-based applications in a multi-tier network. Chapter 7 concludes the thesis with a summary of the main results, future work, and drawn conclusions.
  

DOCTORAL DEGREE IN OPTICAL ENGINEERING

• ROVIRA GAY, CRISTINA: Objective Evaluation on the Effectiveness of Vergence Vision Training Based on the Analysis of Eye Movements
  
  Author: ROVIRA GAY, CRISTINA
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN OPTICAL ENGINEERING
  Department: Department of Optics and Optometry (OO)
  Mode: Normal
  Deposit date: 01/07/2024
  Deposit END date: 12/07/2024
  Thesis director: MESTRE FERRER, CLARA | ARGILÉS SANS, MARC
  Committee:
       PRESIDENT: PIÑERO LLORENS, DAVID PABLO
       SECRETARI: VILASECA RICART, MERITXELL
       VOCAL: JOLLY, JASLEEN K
  Thesis abstract: When an object of interest is detected in the visual field, a saccadic movement is made and the attention and gaze are centred on it. During this visual fixation, optokinetic eye movement occurs to maintain the retinal image position stable and, at the same time, small fixational eye movements take place. Vestibular eye movements also take part in this action to guarantee the accurate stabilization of the eyes during rotations and translations of the head. If the object of interest moves in the frontal plane, smooth pursuit eye movements allow the eyes to follow the object as it is moving. If the object moves in depth, vergence eye movements align the two eyes with the target at different distances. Specifically, the vergence system, which controls binocular alignment, aligns both foveae with the object of interest allowing fusion of the retinal images into a single percept. Vergence eye movements will be described in a greater detail in this Thesis as they are the most relevant types of eye movements for the purpose of this Thesis.Phoropter rotary prisms and prism bars are commonly used to evaluate the amplitude of both the positive (convergence) and the negative (divergence) fusional vergence amplitudes, measuring the blur, break, and recovery of single vision points. The blur point informs about the amount of relative fusional vergence that can be stimulated without the intervention of accommodation. The break point assesses the total amplitude of fusional vergence amplitudes, and the recovery point indicates the patient’s ability to recover single binocular vision after diplopia occurs. These two tests are subjective and depend on the answers of the patients, optometrist’s experience, manual dexterity, and examination criteria, which leads to variability of the results and poor repeatability. One of the purposes of this PhD Thesis is to evaluate objectively the fusional vergence amplitudes and compare the results with the conventional subjective clinical tests. Vision therapy or vision training, also called, orthoptics has been demonstrated to be an effective treatment for convergence (CI), among other binocular dysfunctions. The objective of this treatment is to increase the amplitude, speed, accuracy of accommodative and vergence responses, and the visual comfort. Since 1855, vision therapy has been the primary treatment for CI and there are validated protocols to improve the abilities of the binocular vision system. The objective monitoring of the effectiveness of this treatment for normal binocular vision participants is a novelty of this PhD Thesis. In this PhD Thesis, vergence responses are evaluatep in a subjective and an objective way using conventional procedures and the eye tracker EyeLink 1000 Plus (SR Research Ltd., Ontario, Canada). The main purpose of this PhD Thesis is to assess objectively the effectiveness of a conventional vision therapy protocol for training vergence responses
  

DOCTORAL DEGREE IN PHOTONICS

• CIRAUQUI GARCÍA, DAVID: Optimization with spin glass models
  
  Author: CIRAUQUI GARCÍA, DAVID
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN PHOTONICS
  Department: Institute of Photonic Sciences (ICFO)
  Mode: Normal
  Deposit date: 20/06/2024
  Deposit END date: 04/07/2024
  Thesis director: LEWENSTEIN, MACIEJ | MARTÍNEZ SAAVEDRA, JOSÉ RAMÓN | RYSZARD GRZYBOWSKI, PRZEMYSLAW
  Committee:
       PRESIDENT: MAZZANTI CASTRILLEJO, FERNANDO PABLO
       SECRETARI: PRUNERI, VALERIO
       VOCAL: DELLANTONIO, LUCA
  Thesis abstract: With applicability on almost every aspect of our lives, optimization problems are ubiquitous to a broad range of fields within both scientific research and industrial environments. As such, these are growing in size and complexity at a fast pace, and are only expected to continue to do so. Accordingly, the urgency for better methods that can yield more optimal solutions in shorter times is increasing and, while the development of quantum computing technologies that are capable of tackling these problems evolves steadily, it does so too slowly for the challenges that nowadays society's demands represent. Consequently, a lot of effort is being invested to further develop classical methods and machines that are specially designed to solve optimization problems of relevant enough sizes. The present thesis is framed within this paradigm: classical optimization techniques are studied from various different perspectives, with the goal of improving their efficiency. To this end, we first dive into basic concerns related to the physical properties of the systems that allow for the convenient formulation of industrially-relevant optimization problems, namely spin glasses with quenched disorders. The understanding of such properties is of utmost importance for the correct designing of the annealing schedules used by thermally-based optimization methods. We then study the impact that the hidden correlations of the pseudo random number streams used in their simulations have in the results by comparing simulations using PRNGs of various qualities and perfectly random QRNGs. To conclude, we investigate novel ways, inspired by quantum-mechanical systems, to efficiently navigate the energy landscapes of spin glasses in classical algorithms, which has the potential of preventing the simulations getting stuck into local energy minima and thus reaching more optimal solutions.
  
• HÄGELE, SEBASTIAN: Compact Phase Imaging Platform and its Application to Material Science and Manufacturing
  
  Author: HÄGELE, SEBASTIAN
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN PHOTONICS
  Department: Institute of Photonic Sciences (ICFO)
  Mode: Normal
  Deposit date: 20/06/2024
  Deposit END date: 04/07/2024
  Thesis director: PRUNERI, VALERIO | TERBORG, ROLAND ALFONSO
  Committee:
       PRESIDENT: FELIX PEREIRA, SILVANIA
       SECRETARI: ARTIGAS GARCIA, DAVID
       VOCAL: VOGL, ULRICH
  Thesis abstract: As the world moves towards increasingly miniaturized and complex technologies and devices, the need for imaging and metrology tools for precise material characterization and fabrication process control is rising accordingly. For highly transparent and ultra-thin structures and samples (e.g., optical coatings, lithographic structures or biological cells), intensity-based imaging techniques fall short due to insufficient contrast, as well as failing to provide quantitative information.To overcome these limitations, the field of phase imaging, based on superposition and interference of light, has emerged. In order to create image contrast, phase imaging does not leverage changes in intensity, but rather, as the name implies, changes in the phase of the electro-magnetic wave. With a long-standing history, and Nobel prizes awarded in 1953 to Zernike’s “phase contrast microscope” and 1971 to Gabor’s holographic methods, the field has evolved to “quantitative phase imaging” (QPI), using sophisticated methods and setups to control and manipulate the state of light in order to recover the phase information quantitatively. Herein, the category of “common-path” techniques promises adaptable, compact, robust, and cost-efficient imaging devices, enabling use in industrial applications outside of a well-controlled lab environment.In this thesis, we will describe the development and technological innovations of a “common-path” phase imaging platform based on the “lateral-shearing interferometric microscopy” (LIM) technology. We will implement and adapt the platform to various optical setups, e.g., for large-area lens-free imaging and for high-resolution microscopic imaging. We will also demonstrate the performance and versatility of the platform by exploring a range of applications, with a focus given to material science and manufacturing. Specifically, we will perform volumetric imaging of the tiniest femtosecond laser-written refractive index (RI) changes inside glass. This is followed by the characterization of semi-transparent ultra-thin gold films using multispectral intensity and phase imaging, enabling us to determine the complex RIs of the films of varying thickness. Lastly, we will apply the platform to the imaging of curing grades and RI changes in photopolymers, such as those used in resin-based 3D printing. Further applications of the platform could include surface metrology, imaging of 2D materials, as well as quantitative phase imaging for bio- and cell-imaging applications, with the possibility of integrating the whole platform into a compact add-on which could be added to any commercial microscope. In summary, this thesis will make evident the significant potential of phase imaging in both research and industrial settings, enabled by the proposed compact phase imaging platform. The work builds the foundation for future innovations and developments with a potentially lasting impact on the photonics industry.
  
• HERKERT, EDIZ: Advanced Nanoantenna Platforms for Enhanced Single-Molecule Detection at High Concentrations
  
  Author: HERKERT, EDIZ
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN PHOTONICS
  Department: Institute of Photonic Sciences (ICFO)
  Mode: Normal
  Deposit date: 25/06/2024
  Deposit END date: 08/07/2024
  Thesis director: GARCÍA PARAJO, MARÍA
  Committee:
       PRESIDENT: ACUNA, GUILLERMO
       SECRETARI: VAN HULST, NIEK
       VOCAL: ZIJLSTRA, PETER
  Thesis abstract: The ability to study the dynamics of individual biomolecules is crucial to understanding the complex organization of biological systems beyond what can be learned from ensemble averages. These single-molecule dynamics often occur at high micro- to millimolar concentrations, where conventional optical techniques cannot isolate single molecules anymore due to fundamental physical laws. This thesis explores the design, fabrication, and application of advanced nanoantenna platforms to detect individual fluorescent molecules at such high concentrations with increased sensitivity.Here, the theoretical groundwork is provided to understand the interactions between fluorescent molecules and nanoantennas. It is discussed how the single-molecule detection sensitivity of nanoantenna platforms can be quantitatively assessed through analytical models and numerical simulations. Based on these quantitative models, antenna-in-box platforms are identified to provide superior sensing performance and suitable lithography processes for their fabrication are established.Both computational and experimental evidence are presented that cleverly combining materials in hybrid antenna-in-box platforms enhances single-molecule detection sensitivity at micromolar concentrations. This improvement is attributed to decreased background signals and the use of previously unexplored coupling mechanisms inherent in the antenna-in-box architecture. Furthermore, hexagonal close-packed antenna-inbox platforms are introduced to enable highly parallelized single-molecule detection at micromolar concentrations. Notably, these hexagonally ordered platforms constitute the first demonstration of antenna-in-box platforms capable of single-molecule detection across the visible spectral range.Lastly, a correlative approach is presented that combines nonlinear fluorescence and vibrational spectroscopy to study the organization of receptor proteins in the cell membrane of living cells using nanoantennas. Measures to protect both the nanoantennas and the living cells are discussed and their effectiveness is validated.Overall, this thesis presents novel approaches for studying single-molecule dynamics at high concentrations with enhanced sensitivity. The development of these approaches was enabled through analytical and numerical modeling, the creation of new fabrication processes, and the use of appropriate experimental methods. These advancements promise to offer previously inaccessible insights into dynamics within biological systems.
  

DOCTORAL DEGREE IN SIGNAL THEORY AND COMMUNICATIONS

• PALACIOS ARIAS, CESAR AUGUSTO: Design, Implementation, and Experimental Validation of Microsystems for Near-field Communications and Sensing at X-waves
  
  Author: PALACIOS ARIAS, CESAR AUGUSTO
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN SIGNAL THEORY AND COMMUNICATIONS
  Department: Department of Signal Theory and Communications (TSC)
  Mode: Article-based thesis
  Deposit date: 01/07/2024
  Deposit END date: 12/07/2024
  Thesis director: JOFRE ROCA, LUIS | JOFRE CRUANYES, MARC
  Committee:
       PRESIDENT NO PRESENCIAL: FERRANDO BATALLER, MIGUEL
       SECRETARI: RIUS CASALS, JUAN-MANUEL
       VOCAL NO PRESENCIAL: RAMÍREZ ARROYAVE, GERMÁN AUGUSTO
  Thesis abstract: This thesis delves into the intricate process of designing and implementing micrometric scale systems for "x-waves" applications, aterm used in this thesis to encompass microwaves, millimeter waves, optics, and terahertz waves. The initial results have beenachieved with microsystems operating at microwave frequencies, integrated with microfluidics technologies for bioparticle near-fieldcommunication and sensing. In particular, microwave and microfluidic technologies has emerged as a promising approachfor the detection and analysis of bioparticles. This interdisciplinary field, situated at the intersection of communication engineering,microscale physics,and biology, leverages the unique properties of microwaves and the precise control offered by microfluidicsto explore the behavior and functionality of living cells. This thesis encompasses three interconnected studies that in conjunctionexplore the integration of microwave and microfluidic technologies for the detection, differentiation, and analysis of bioparticles.The first study focuses on the design and optimization of a high-sensitivity measurement system capable of detecting bioparticlesover a frequency range of 0.01–10 GHz using various configurations of coplanar electrodes on a microfluidic platform. The proposedmeasurement setup addresses the detection gap at microwaves with real-time superheterodyne microwave detection systembased on the optimization of a Lock-In-Amplifier (LIA) for single particle detection. The second study demonstrates the system’scapability to differentiate single live/dead bioparticles. It incorporates a Transimpedance Amplifier (TIA) in the measurement systemto improve the signal-to-noise ratio (SNR) by 4 dB. In addition, the superheterodyne receiver was optimized in terms of localoscillator power and operation frequency. The microfluidic system has been adjusted to confine bioparticles at the centerline of themicrochannel. The third study presents a system that measures the electromagnetic nonlinear susceptibility variation of livingorganisms. The technique is based in the measurement of the intermodulation products at microwaves frequencies produced byvarious samples, including pure ethanol, a mixture of ethanol and dimethyl-sulfoxide (DMSO), live Escherichia coli (E. coli), andheat-killed E. coli. Collectively, these studies highlight the potential of this integrated microwave-microfluidic platform in industrialprocesses, biomedical research, and environmental monitoring applications. The research opens up an unexplored avenue in themicrowave field for understanding the behavior and functionality of living cells.
  
• SEDAR, MOHOTTIGE ROSHAN MADHUSANKA: Misbehaviour Detection and Trustworthy Collaboration in Vehicular Communication Networks
  
  Author: SEDAR, MOHOTTIGE ROSHAN MADHUSANKA
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN SIGNAL THEORY AND COMMUNICATIONS
  Department: Department of Signal Theory and Communications (TSC)
  Mode: Normal
  Deposit date: 20/06/2024
  Deposit END date: 04/07/2024
  Thesis director: ALONSO ZARATE, JESUS | VAZQUEZ GALLEGO, FRANCISCO
  Committee:
       PRESIDENT: MARQUEZ BARJA, JOHANN MARCELO
       SECRETARI: PEREZ ROMERO, JORGE
       VOCAL: GOZÁLVEZ SEMPERE, JAVIER
  Thesis abstract: The integration of advanced wireless technologies, e.g., cellular and IEEE 802.11p, in modern vehicles enables vehicle-to-everything (V2X) communication, fostering the next-generation Internet-of-Vehicles (IoV). The rise of IoV leads to more connected vehicles on roads, capable of making informed and coordinated decisions through real-time information sharing among vehicles, communication infrastructure, pedestrians, or roadside units (RSUs). However, V2X and IoV technologies inadvertently bring unprecedented challenges involving security and privacy vulnerabilities. Security threats and attacks can emerge from both malicious outsiders and insiders in V2X communication. Detecting and containing misbehaviours, particularly those initiated by rogue insiders, present challenging yet critical tasks for ensuring road safety. Furthermore, the pervasive use of artificial intelligence and machine learning (AI/ML) tools across various aspects poses potential threats to secure V2X operations. Motivated by these challenges, this doctoral thesis focuses on enhancing the security, robustness, and trustworthiness of V2X communications by enabling efficient and effective misbehaviour detection and fostering trustworthy collaboration. Specifically, we focus on (i) achieving effective and efficient misbehaviour detection with high accuracy and minimal false alarms, leveraging diverse spatiotemporal characteristics in vehicular data, and (ii) facilitating trustworthy information sharing for collaborative misbehaviour detection, with an emphasis on generalisability and the ability to detect previously unseen and partially observable attacks.The absence of standardised approaches to address misbehaviours calls for advanced AI/ML-based solutions capable of handling the surging volume of data, enhancing robustness and generalisability, and meeting the real-time demands of V2X applications. To this end, we propose a generic deep RL (DRL) misbehaviour detection methodology capable of dynamically improving detection through interactions and experiences by leveraging various spatiotemporal behaviours present in the ambient vehicular measurement space. The scarcity of labelled vehicular data exacerbates the effective training of AI/ML-based models. Motivated by this challenge, we propose an ensemble learning framework for misbehaviour detection, coupled with unsupervised learning and a DRL model. This enables the detection of attacks from unlabelled vehicular data, facilitating the generalisation and detection of new and unseen attack variants. Additionally, addressing adversarial attacks poses a significant challenge, requiring enhanced solutions to make AI/ML-based misbehaviour detection more resilient against such threats. Towards this, we introduce and evaluate a tailored DRL approach designed to protect V2X communication systems against adversarial attacks, as well as mitigate issues stemming from inappropriate formatting of input training data due to vehicular sensor malfunctions or reading errors. By implementing data poisoning adversarial attacks, we demonstrate the resilience of the DRL-based misbehaviour detection approach even under severe conditions of sophisticated adversarial manipulation.Building upon the proposed DRL-based misbehaviour detection approach, we introduce a novel scheme for collaborative misbehaviour detection. This scheme involves deploying a DRL-based misbehaviour detection model in an RSU at the network edge. It leverages transfer learning principles to share the knowledge learned about misbehaviours at the source RSUs with the target RSU, enabling the reuse of relevant expertise for collaborative misbehaviour detection. Considering data poisoning attacks aimed at influencing misbehavior detection, we implement selective knowledge transfer from trustworthy RSUs to avoid adversarial interference. We introduce a semantic relatedness metric to quantify each RSU's trust level for collaborative misbehavior detection.
  
• SHAABANZADEH, SEYEDEH SOHEILA: Contribution to the Development of Wi-Fi Networks through Machine Learning based Prediction and Classification Techniques
  
  Author: SHAABANZADEH, SEYEDEH SOHEILA
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN SIGNAL THEORY AND COMMUNICATIONS
  Department: Department of Signal Theory and Communications (TSC)
  Mode: Normal
  Deposit date: 18/06/2024
  Deposit END date: 02/07/2024
  Thesis director: SANCHEZ GONZALEZ, JUAN
  Committee:
       PRESIDENT: ADELANTADO FREIXER, FERRAN
       SECRETARI: PEREZ ROMERO, JORGE
       VOCAL: KOUTLIA, KATERINA
  Thesis abstract: The growing number of Wi-Fi users and the emergence of bandwidth-intensive services have necessitated an increase in Access Point (AP) density, resulting in more complex network configuration, optimization, and management tasks. Concurrently, advancements in data monitoring and analytics technologies in wireless networks offer opportunities to extract valuable insights into network and user behavior, facilitating more efficient network management. In this thesis, we propose different Machine Learning based techniques to enhance Wi-Fi network management, focusing on three aspects: user connectivity prediction, Wi-Fi traffic prediction, and Wi-Fi traffic classification. The first aspect of our work focuses on predicting the next Access Point (AP) a user will connect to in a Wi-Fi network. We propose a methodology based on historical information of the AP to which a user has been connected, extracting connectivity patterns at different time scales (hourly, daily, weekly). Predictions are done using techniques based on Neural Networks and Random Forest algorithms. This approach is evaluated using real data from a university campus Wi-Fi network. Predicting the next AP of users allows for proactive network reconfiguration, enhancing the efficiency of techniques like Pairwise Master Key caching and Opportunistic Key Caching, which reduce re-authentication times. Additionally, this prediction helps to identify the geographical region of the User Equipment (UE) and can be used for commercial purposes, such as targeted advertising, by customizing messages based on locations of the users. Secondly, we propose a methodology for predicting the aggregated traffic at access points (APs) by leveraging spatial and temporal correlations from neighboring APs to enhance prediction accuracy. Using real measurements, we evaluate various Deep Learning methods, including Convolutional Neural Network (CNN), Simple Recurrent Neural Network (SRNN), Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM), and Transformer, and present a hybrid approach combining CNN for spatial processing and RNN for temporal prediction. This hybrid method improves accuracy with minimal additional training time and negligible impact on prediction time. Accurate traffic forecasting at each AP enables better load distribution and can inform resource management techniques such as admission control, congestion control, and load balancing. Additionally, predicting low traffic periods can aid in energy-saving strategies by allowing APs with minimal traffic to be switched off during specific times. Finally, traffic classification is essential for enhancing network performance by allowing better resource allocation and prioritization of services with stringent latency requirements. The increasing demand for Virtual Reality (VR) services poses a significant challenge for Wi-Fi networks to meet strict latency needs, crucial for VR to ensure immediate response and avoid user discomfort. To improve VR Quality of Service (QoS), distinguishing interactive VR traffic from Non-VR traffic is key. We propose a machine learning-based method to identify interactive VR traffic in a Cloud-Edge VR environment by analyzing downlink and uplink data correlations and extracting features from single-user traffic characteristics. Six classification techniques (i.e., Logistic Regression, Support Vector Machines, k-Nearest Neighbors, Decision Trees, Random Forest, and Naive Bayes) are compared. The result of the classification is used for the prioritization of VR traffic over Non VR traffic. We evaluate our method using datasets from various VR applications and Wi-Fi network simulations. Our results show a significant reduction in VR traffic delays with minimal impact on Non-VR service latency.
  

DOCTORAL DEGREE IN SUSTAINABILITY

• MORALES VERGARA, IVETHEYAMEL: The Impact of Urban Living Labs in Sustainable Development Education, Case Study: Faculty of Architecture and Urbanism, University of Guayaquil - Ecuador.
  
  Author: MORALES VERGARA, IVETHEYAMEL
  Thesis file: (contact the Doctoral School to confirm you have a valid doctoral degree and to get the link to the thesis)
  Programme: DOCTORAL DEGREE IN SUSTAINABILITY
  Department: University Research Institute for Sustainability Science and Technology (IS.UPC)
  Mode: Article-based thesis
  Deposit date: 27/06/2024
  Deposit END date: 10/07/2024
  Thesis director: MASSECK, TORSTEN ANDREAS | SEGALAS CORAL, JORDI | MASSECK, TORSTEN ANDREAS
  Thesis abstract: The following doctoral thesis presents an investigation into the application and effectiveness of Urban Living Laboratories (ULLs) as transformative educational strategies within Higher Education Institutions (HEIs) to promote Education for Sustainable Development (ESD). Through the application of a diverse methodological framework that includes comprehensive literature reviews, expert validations and extensive case study analyses, with a focus on the Delta Project of the Faculty of Architecture and Urbanism of the University of Guayaquil, this research significantly advances the understanding of how ULLs can be integrated into higher education curricula.The thesis is structured in seven chapters with its core part made up of three scientific articles. The first part presents a literature review and conceptual framework. This section provides a comprehensive review of the literature on ESD within HEIs, detailing the conceptualisation and operationalisation of ULLs. It explores the fundamental theories and practices of ESD, emphasising the role of HEIs in leading sustainability initiatives. The review also covers the characteristics and typologies of ULLs, placing them within the broader context of sustainable urban development and educational innovation.In the second part, the thesis presents an innovative evaluation tool (E-ULL-HEIs), designed to measure the impact and effectiveness of ULLs in promoting sustainable practices within the academic environment. This tool evaluates ULLs through three constructs: Synergy, Strategy and Pedagogy, which describe key interactions for ESD. Synergy emphasises cross-sectoral and disciplinary cooperation; Strategy addresses clarity of purpose and lasting outcomes; and Pedagogy combines the incorporation of new knowledge with learning and research impact.These constructs are developed into seven key indicators covering the essential elements of lifelong learning, ESD and higher education institutions. By applying this assessment tool to fifty case studies, the research validates the reliability and consistency of these indicators. The results of these studies, confirmed by exploratory factor analysis and Cronbach's alpha coefficient, demonstrate the robustness of the tool in capturing the nuanced impacts of ULL on educational strategies and community engagement. This evaluation framework not only facilitates the assessment of ULL contributions to sustainability, but also ensures their alignment with educational learning outcomes and the broader Sustainable Development Goals (SDGs).In the third part, this assessment tool is applied in the Delta Project, a ULL at the University of Guayaquil, Ecuador. This tool was crucial to identify both the strengths and weaknesses of the project. The findings highlight the impact on diversity and stakeholder participation, interdisciplinarity, curricular integration, learning and research. However, it also identified challenges in terms of cross-sectoral collaboration and student engagement, providing valuable lessons on the importance of strategic planning and adaptability.The results of this doctoral research highlight the transformative potential of ULLs in HEIs, highlighting them as central to promoting sustainable urban transitions. However, the thesis also identifies challenges in cross-sectoral collaboration and student engagement that provide critical insights into the strategic planning necessary for the successful adaptation and replication of ULL models in different cultural and academic contexts.