ESCUELA POLITÉCNICA SUPERIOR

Permanent URI for this collectionhttps://hdl.handle.net/10641/5230

Browse

Now showing 1 - 20 of 298

Walsh sequences as direct Error Correcting Output Code dichotomies for multiclass problems
(2026-06-01) Álvarez-Pérez, Lorena; Callejas-Ramos, Álvaro; Centro de Innovación Experimental del Conocimiento (CEIEC); Universidad Francisco de Vitoria
Resolver problemas de clasificación multiclase mediante esquemas de Código de Salida con Corrección de Errores (ECOC) ofrece flexibilidad y robustez, pero los enfoques tradicionales basados en secuencias de Rademacher sufren un aumento exponencial de dicotomías con el número de clases. Este artículo presenta dos contribuciones principales. Primero, proponemos el uso de secuencias de Walsh para construir directamente dicotomías ECOC, permitiendo abordar problemas multiclase con C clases utilizando solo C−1 clasificadores, preservando altas distancias de Hamming por pares entre los códigos de clase (C2). Segundo, presentamos un método fundamentado para estimar las probabilidades posteriores de clase a partir de las salidas de clasificadores discriminativos entrenados con divergencias de Bregman. Esto proporciona una alternativa a la decodificación de Hamming, mejora la estimación de confianza y evita la inversión de matrices gracias a la ortogonalidad de los códigos de Walsh. Los resultados experimentales en cinco conjuntos de datos multiclase —incluyendo Fashion-MNIST, Letter Recognition y Vowel— demuestran mejoras consistentes en la precisión de clasificación (por ejemplo, una reducción del error del 5% en el conjunto Letter) y en la eficiencia computacional frente a métodos de referencia como OvR y ECOC-Rademacher. El rendimiento se evaluó mediante precisión de clasificación estándar y matrices de confusión, validando las ventajas de nuestro enfoque tanto en rendimiento predictivo como en escalabilidad. [Traducción automática al español; idioma original: inglés.]
Classification of systolic murmurs in heart sounds using multiresolution complex Gabor dictionary and vision transformer
(2026-02-28) Fakhry, Mahmoud; Brery, Abeer Fath Allah; Centro de Innovación Experimental del Conocimiento (CEIEC); Universidad Francisco de Vitoria
Los soplos sistólicos son sonidos cardíacos adicionales que ocurren durante la fase de contracción del ciclo cardíaco, y a menudo indican anomalías cardíacas causadas por un flujo sanguíneo turbulento. Su intensidad, tono y calidad varían, lo que requiere una identificación precisa para el diagnóstico exacto de los trastornos cardíacos. Este estudio presenta un sistema de clasificación automática de soplos sistólicos mediante un módulo de extracción de características, seguido de un modelo de clasificación. El módulo de extracción de características emplea la búsqueda de coincidencias ortogonales complejas para proyectar uno o varios segmentos de soplo sobre un diccionario redundante compuesto por funciones base de Gabor complejas (GBF) multirresolución. Los pesos de proyección resultantes se dividen y remodelan en matrices de características tiempo-frecuencia de resolución variable. El procesamiento de múltiples segmentos de una misma grabación utilizando un diccionario compartido mitiga la variabilidad del soplo. Esto se logra aprendiendo los pesos de cada segmento a la vez que se impone que correspondan al mismo conjunto de funciones base del diccionario, favoreciendo matrices de características tiempo-frecuencia consistentes. El modelo de clasificación se construye a partir de un transformador visual para procesar múltiples matrices de entrada de distintas resoluciones, haciendo pasar cada una por una red neuronal convolucional para la tokenización en parches. Todos los tokens de incrustación se concatenan para formar una matriz y se envían a una capa codificadora que incluye atención multicabeza, conexiones residuales y una red convolucional con un tamaño de núcleo de uno. Esta integración de la extracción de características multirresolución con la clasificación de características basada en transformadores mejora la precisión y fiabilidad de la identificación de soplos cardíacos. Un análisis experimental de cuatro tipos de soplos sistólicos del conjunto de datos CirCor DigiScope demuestra la eficacia del sistema, alcanzando una precisión de clasificación del 95,96%. [Traducción automática al español; idioma original: inglés.]
A model-based solution for automated (Re-)engineering of task-oriented chatbots
(2026-01) Pérez-Soler, Sara; Guerra, Esther; de Lara, Juan; Escuela Politécnica Superior
Los chatbots son populares para acceder a todo tipo de servicios de software mediante conversaciones en lenguaje natural. La creciente demanda de chatbots orientados a tareas ha impulsado la propuesta de muchas herramientas para su construcción, como Dialogflow, Lex, Rasa o Watson. Sin embargo, seleccionar la más adecuada es difícil; el diseño conceptual detrás de un chatbot puede quedar enterrado bajo las tecnicidades de la herramienta; y la migración entre plataformas de desarrollo de chatbots debe realizarse manualmente. Para aliviar estos problemas, proponemos una notación de diseño independiente de la plataforma para chatbots orientados a tareas, basada en el análisis de quince plataformas de desarrollo de chatbots. Siguiendo los principios de la ingeniería dirigida por modelos, la implementación del chatbot se sintetiza a partir del diseño, y los diseños pueden extraerse de las implementaciones, permitiendo la migración y la reingeniería de chatbots. Además, un recomendador sugiere la plataforma más adecuada para un diseño de chatbot dado, considerando factores contextuales. Hemos materializado estas ideas en CONGA: una aplicación web extensible que incluye un editor de notación de diseño; un recomendador de plataformas de desarrollo; validadores específicos de plataforma; y generadores y analizadores para Dialogflow y Rasa. Evaluamos CONGA sobre 291 chatbots de código abierto de Dialogflow y Rasa, demostrando su expresividad, portabilidad y utilidad para encontrar problemas de calidad en los chatbots (encontrados en el 93,8 % de los chatbots). En conjunto, nuestra arquitectura permite diseños de chatbot neutrales, automatiza la migración y proporciona mecanismos para la detección de defectos a nivel de diseño. [Traducción automática al español; idioma original: inglés.]
Novel Real-Time Acoustic Power Estimation for Dynamic Thermoacoustic Control
(2025-03) Pilo de la Fuente, Eduardo; Gros, Jaime; Simón Rodríguez, María Antonia; Velasco, Ana Isabel; Iniesta, Carmen; Escuela Politécnica Superior; Ingeniería del Automóvil y Movilidad
This paper presents a new procedure for the real-time processing and analysis of data from thermoacoustic systems. The approach focuses on continuously acquiring and adjusting measurements of acoustic wave pressure, enabling the instantaneous estimation of acoustic power. This is crucial for real-time control and decision-making, especially in applications that require rapid power estimation, such as the control loop implementation in thermoacoustic engines, where conditions are constantly changing and dynamic adaptation is essential. Two methods for estimating the power delivered to the load are proposed: (method 1) instantaneous power evaluation, which calculates the power consumed by the resistance in the resistance–capacitance (RC) load, and (method 2) one-period average power calculation using the well-established two-microphones method. These methods are validated with both different synthetic signals and experimental measurements. The results reveal that the new method provides real-time accurate estimations of the power delivered to the acoustic load and, thus, has shown potential for control-based applications.
Thermodynamic Coupling Forming Performance of Short Fiber-Reinforced PEEK by Additive Manufacturing
(2024-07) Sun, Qili; Wen, Xiaomu; Yin, Guangzhong; Jia, Zijian; Yang, Xiaomei; Escuela Politécnica Superior; Advanced Sustainable Materials
Exploiting Waste towards More Sustainable Flame-Retardant Solutions for Polymers : A Review
(2024-05) Ma, De Xin; Yin, Guang Zhong; Ye, Wen; Jiang, Yan; Wang, Na; Wang, De Yi; Escuela Politécnica Superior; Advanced Sustainable Materials
Application of natural language processing techniques to network traffic processing for classification using deep learning models
(2025-12) Maitin, Ana M.; Arranz-Luque, Carlos; Alba, Emilio; García-Tejedor, Álvaro J.; Universidad Francisco de Vitoria; Escuela Politécnica Superior; Centro de Innovación Experimental del Conocimiento (CEIEC)
Antecedentes: El rápido crecimiento del tráfico de red cifrado ha aumentado la necesidad de una clasificación del tráfico de red (NTC) eficaz y no sesgada. Las técnicas tradicionales tienen dificultades con los datos cifrados, la disponibilidad limitada de características y el elevado volumen de tráfico, lo que reduce su fiabilidad en escenarios reales. Métodos: Proponemos una metodología de preprocesamiento novedosa que analiza el tráfico de red en bruto y lo transforma en un formato textual (nt2txt), lo que permite aplicar técnicas de Procesamiento del Lenguaje Natural (NLP) y de Aprendizaje Profundo. Este enfoque elimina el sesgo derivado de los metadatos de protocolo, estructura los datos en semi-flujos de tamaño fijo y utiliza una división rigurosa de los datos para evitar el solapamiento de flujos entre entrenamiento y prueba. A continuación, se entrena un modelo basado en LSTM para clasificar el tráfico utilizando únicamente los datos de la carga útil (payload). Resultados: Este trabajo proporciona un marco escalable e independiente del protocolo para la clasificación de tráfico cifrado, demostrando la eficacia de las técnicas de NLP para mejorar el rendimiento del modelo y reducir el sesgo del conjunto de datos. Nuestra metodología alcanzó una precisión del 88,87 ± 0,04% en un conjunto de datos externo ciego, superando a modelos similares LSTM e híbridos CNN-LSTM. Métricas como la Kappa de Cohen y el Coeficiente de Correlación de Matthews confirman además la robustez y la generalizabilidad de nuestro enfoque. [Traducción automática al español; idioma original: inglés.]
State-of-the-Art AI for Smarter Cities: A Global Research Showcase : Developed for Smart City Expo World Congress. 4-6 NOV 2025 Barcelona (Spain)
(2025-11-28) Ondiviela García, José Antonio; Fernández-Anez, Victoria; Escuela Politécnica Superior; Ciudades Atractivas; Persona, Inteligencia Artificial y Humanidades
Mechanical Properties and Thermal Decomposition Mechanism of Glycidyl Azide Polyol Energetic Thermoplastic Elastomer Binder with RDX Composite
(2024-09) Sun, Qili; Yang, Xiao Mei; Yin, Guang Zhong; Escuela Politécnica Superior; Advanced Sustainable Materials
Navigating Passenger Satisfaction : A Structural Equation Modeling–Artificial Neural Network Approach to Intercity Bus Services
(2024-06) Rahnama, Shaghayegh; Cortez, Adriana; Monzon, Andres; Escuela Politécnica Superior; Ingeniería del Automóvil y Movilidad; Sostenibilidad y Economía Circular en la Era Digital
The phenomenon of passenger satisfaction is an important issue for public transport services and transport companies. Clarifying the relationship between influencing attributes and passenger satisfaction significantly improves service satisfaction. This study examines passenger satisfaction with intercity buses and, in particular, the role of digital information channels (websites and mobile apps) in promoting sustainable travel choices on the Madrid–Bilbao route. This study analyzed data from 459 passengers to identify the key factors influencing the bus choice for intercity bus travel. Punctuality, safety, and ticket price are the most important determinants. We use a combined structural equation modeling (SEM) and artificial neural network (ANN) approach to capture the intricate relationships between service attributes and information channels. The results show that information channels, travel experience, and ticket prices significantly impact passenger satisfaction, which bus operators should improve. Also, inserting the SEM result as input for the ANN showed that ticket price is the most significant predictor of satisfaction, followed by information channels (84%) and travel experience (65%). This approach provides valuable insights for improving the passenger experience. This study emphasizes integrating digital transformation strategies into public transport systems to promote sustainable mobility goals.
A singular perturbation result for a class of periodic-parabolic BVPs
(2024-01-01) Cano-Casanova, Santiago; Fernández-Rincón, Sergio; López-Gómez, Julián; Escuela Politécnica Superior
Advancements of nanotechnological strategies as conventional approach for heavy metal removal from industrial wastewater : Start-of-the-art review
(2024-01) Raturi, Sakshi; Kumari, Swati; András, Kovács; Khargotra, Rohit; Sebestyén, Viktor; Singh, Tej; Escuela Politécnica Superior
Multi-faceted growth and progression of the healthy and economical society, depends upon access to clean and safe water. Rapidly over-growing population, increased in industrialization, urbanisation, and widespread practices in agricultural have all together been contributing to the production of more rapid wastewater discharge, which has not only polluted or contaminated the water but also have played a role in killing the aquatic life. One class of harmful water pollutants that is frequently found in the environment is heavy metals. Almost every transition metal has the ability to dissolve as ions in water. Heavy metals including Pb, Cd, Hg, As, Se and others can contaminate water supplies. Conventional methods for waste-water treatment have peculiar challenges including economic feasibility, energy consumption, environmental hazards, time spent, etc. To overcome these limitations, nanotechnology have been developed, which has its greater extent of application in water treatment area. Nanoparticles have a greater probability of removing heavy metals from wastewater treatment due to their effective surface characteristics and chemical activity. This review focuses on the numerous treatment procedures that have been developed recently and also been applied practically for eradication of heavy metals from waste-water of various industries.
Co-creation and evaluation of an app to support reminiscence therapy interventions for older people with dementia
(2024-01-01) De-Rosende-Celeiro, Iván; Francisco-Gilmartín, Virginia; Bautista-Blasco, Susana; Ávila-Álvarez, Adriana; Escuela Politécnica Superior; La Transformación de la Información en Inteligencia (TI3); Sociedad y Tecnología Centrada en la Persona
Optimized Autonomous Drone Navigation Using Double Deep Q-Learning for Enhanced Real-Time 3D Image Capture
(2024-12) Sánchez-Soriano, Javier; Rojo-Gala, Miguel Ángel; Pérez-Pérez, Guillermo; Bemposta Rosende, Sergio; Gordo-Herrera, Natalia; Escuela Politécnica Superior; Matemática Aplicada y Computación Cuántica; Intelligent Robotic Systems
The proposed system assists in the automatic creation of three-dimensional (3D) meshes for all types of objects, buildings, or scenarios, using drones with monocular RGB cameras. All these targets are large and located outdoors, which makes the use of drones for their capture possible. There are photogrammetry tools on the market for the creation of 2D and 3D models using drones, but this process is not fully automated, in contrast to the system proposed in this work, and it is performed manually with a previously defined flight plan and after manual processing of the captured images. The proposed system works as follows: after the region to be modeled is indicated, it starts the image capture process. This process takes place automatically, with the device always deciding the optimal route and the framing to be followed to capture all the angles and details. To achieve this, it is trained using the artificial intelligence technique of Double Deep Q-Learning Networks (reinforcement learning) to obtain a complete 3D mesh of the target.
sEMG-Controlled Soft Exo-Glove for Assistive Rehabilitation Therapies
(2024) Copaci, Dorin; Cerro, David Serrano Del; Guadalupe, Janeth Arias; Lorente, Luis Moreno; Rojas, Dolores Blanco; Escuela Politécnica Superior
The movement of the human hand offers various degrees of freedom, enabling efficient performance of dynamic tasks and robust interaction with the environment in a compliant and continuous manner. However, the rigid exoskeleton used in hand rehabilitation limits the user's freedom of movement, complicating their natural interaction with the environment. In this study, we present a soft exo-glove for assistive rehabilitation actuated by Shape Memory Alloys (SMA), controlled by a surface electromyography (sEMG) hand gesture classifier. Thanks to the actuator type, the soft exo-glove enables slow, smooth, and controlled movements when activated and provides complete control transparency when the device is not active. This advantage enhances the comfort and acceptance of the exo-glove by the patient. On the other hand, the classifier, in conjunction with the control algorithm and the soft exo-glove, offers the potential to use the exo-glove in assistive hand rehabilitation therapy. For user-friendly use, an interface has been developed, enabling the acquisition of new sEMG data from new users, retraining of the classifier, and connection with the soft exo-glove for rehabilitation therapy. The main objective of this study is to demonstrate that the proposed wearable soft exo-glove, along with the control algorithm and the employed classifier, constitutes an effective solution for assistive rehabilitation tasks, as demonstrated with healthy subjects. Furthermore, this solution can be easily adapted to the users' characteristics and requirements.
Water Architectures in the Alto Guadiana River
(2024) Santolaria Castellanos, Ana Isabel; Alderete, Jaime Ramos; Escuela Politécnica Superior
Along the Alto Guadiana River, there is a collection of architectures built at different times and with diverse uses whose raison d’être is their special relationship with water. Most of these architectural pieces are today abandoned. This research aims to underline their value, highlight their relationship with water, and reflect on the opportunities they represent, understanding that they are key pieces of the landscape. This article presents a narrative that links the Architectures of Water with the ter-ritory and time in an intimate way, a relationship that appears by walking through the landscape, giving them a new meaning. Resignification manifests itself both as a whole, considering the architectures as a collection that informs the landscape, and as individual pieces that can offer opportunities for new forms of balance between humans and nature.
Analysis of Transportation Systems for Colonies on Mars
(2024-04) de Curtò, J.; de Zarzà, I.; Escuela Politécnica Superior
The colonization of Mars poses unprecedented challenges in developing sustainable and efficient transportation systems to support inter-settlement connectivity and resource distribution. This study conducts a comprehensive evaluation of two proposed transportation systems for Martian colonies: a ground-based magnetically levitated (maglev) train and a low-orbital spaceplane. Through simulation models, we assess the energy consumption, operational and construction costs, and environmental impacts of each system. Monte Carlo simulations further provide insights into the cost variability and financial risk associated with each option over a decade. Our findings reveal that while the spaceplane system offers lower average costs and reduced financial risk, the maglev train boasts greater scalability and potential for integration with Martian infrastructural development. The maglev system, despite its higher initial cost, emerges as a strategic asset for long-term colony expansion and sustainability, highlighting the need for balanced investment in transportation technologies that align with the goals of Martian colonization. Further extending our exploration, this study introduces advanced analysis of alternative transportation technologies, including hyperloop systems, drones, and rovers, incorporating dynamic environmental modeling of Mars and reinforcement learning for autonomous navigation. In an effort to enhance the realism and complexity of our navigation simulation of Mars, we introduce several significant improvements. These enhancements focus on the inclusion of dynamic atmospheric conditions, the simulation of terrain-specific obstacles such as craters and rocks, and the introduction of a swarm intelligence approach for navigating multiple drones simultaneously. This analysis serves as a foundational framework for future research and strategic planning in Martian transportation infrastructure.
Optimizing Propellant Distribution for Interorbital Transfers
(2024-03) De Curtò, J.; De Zarzà, I.; Escuela Politécnica Superior
The advent of space exploration missions, especially those aimed at establishing a sustainable presence on the Moon and beyond, necessitates the development of efficient propulsion and mission planning techniques. This study presents a comprehensive analysis of chemical and electric propulsion systems for spacecraft, focusing on optimizing propellant distribution for missions involving transfers from Low-Earth Orbit (LEO) to Geostationary Orbit (GEO) and the Lunar surface. Using mathematical modeling and optimization algorithms, we calculate the delta-v requirements for key mission segments and determine the propellant mass required for each propulsion method. The results highlight the trade-offs between the high thrust of chemical propulsion and the high specific impulse of electric propulsion. An optimization model is developed to minimize the total propellant mass, considering a hybrid approach that leverages the advantages of both propulsion types. This research contributes to the field of aerospace engineering by providing insights into propulsion system selection and mission planning for future exploration missions to the Moon, Mars, and Venus.
Isomorphic Structures and Operator Analysis in Mimetic Discretizations
(2024) De Curtò, J.; De Zarzà, I.; Escuela Politécnica Superior
This study presents a comprehensive examination of the structural and operatorial foundations within mimetic discretizations, with a focus on bridging the gap between discrete and continuous function spaces. By scrutinizing the mimetic gradient and divergence operators-central to the discretization of the NAVIER-STOKES equations-we study their kernel and image spaces, establishing their isomorphisms through rigorous mathematical proofs. Our methodology leverages discrete scalar and vector function spaces, delineated by grid spacing, to define linear mappings that unveil the subspace relationships and quotient space structures integral to understanding these operators' roles in computational fluid dynamics. Central to our findings is the application of the first isomorphism theorem, which facilitates a deeper insight into how mimetic discretizations reflect the continuous properties of differential operators within a discrete framework. This allows for an exploration into the algebraic and topological implications of such discretizations, notably in the context of the NAVIER-STOKES equations. Furthermore, we extend our investigation to encompass subalgebras, ideals, their quotients, and the formulation of short exact sequences that mirror the continuous interplay between gradient, divergence, and LAPLACIAN operators. Significant advances include the application of the first isomorphism theorem which confirms that our mimetic discretizations preserve key properties of differential operators, thus enhancing the accuracy and reliability of computational models. Additionally, our research introduces practical extensions into subalgebras and complex operator sequences, laying groundwork for future developments in numerical methods aimed at improving the precision of engineering simulations.
Spectral Properties of Mimetic Operators for Robust Fluid–Structure Interaction in the Design of Aircraft Wings
(2024-04) de Curtò, J.; de Zarzà, I.; Escuela Politécnica Superior
This paper presents a comprehensive study on the spectral properties of mimetic finite-difference operators and their application in the robust fluid–structure interaction (FSI) analysis of aircraft wings under uncertain operating conditions. By delving into the eigenvalue behavior of mimetic Laplacian operators and extending the analysis to stochastic settings, we develop a novel stochastic mimetic framework tailored for addressing uncertainties inherent in the fluid dynamics and structural mechanics of aircraft wings. The framework integrates random matrix theory with mimetic discretization methods, enabling the incorporation of uncertainties in fluid properties, structural parameters, and coupling conditions at the fluid–structure interface. Through spectral and localization analysis of the coupled stochastic mimetic operator, we assess the system’s stability, sensitivity to perturbations, and computational efficiency. Our results highlight the potential of the stochastic mimetic approach for enhancing reliability and robustness in the design of aircraft wings, paving the way for optimization algorithms that integrate uncertainties directly into the design process. Our findings reveal a significant impact of stochastic perturbations on the spectral radius and eigenfunction localization, indicating heightened system sensitivity. The introduction of randomized singular value decomposition (RSVD) within our framework not only enhances computational efficiency but also preserves accuracy in low-rank approximations, which is critical for handling large-scale systems. Moreover, Monte Carlo simulations validate the robustness of our stochastic mimetic framework, showcasing its efficacy in capturing the nuanced dynamics of FSI under uncertainty. This study contributes to the fields of numerical methods and aerospace engineering by offering a rigorous and scalable approach for conducting uncertainty-aware FSI analysis, which is crucial for the development of safer and more efficient aircraft.

Browse

Recent Submissions