Capítulos (Arquitectura y Tecnología de Computadores)

URI permanente para esta colecciónhttps://hdl.handle.net/11441/11293

Examinar

Mostrando 1 - 20 de 46

Spiking convolution engine for spiking convolution neural networks
(Springer, 2024-08) Curra-Sosa, Dagnier A.; Tapiador Morales, Ricardo; Gómez Rodríguez, Francisco de Asís; Linares Barranco, Alejandro; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Ministerio de Ciencia e Innovación (MICIN). España; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
A promising alternative in artificial vision tasks that considerably re-duces computational cost and power is neuromorphic event-based processing. In this context, we employed a multiconvolution event-based system on a FPGA, inspired by the Leaky Integrate-and-Fire (LIF) neuron, to demonstrate its viabil-ity in implementing a Spiking Convolutional Neural Network (SCNN). In this work, we show that the convolution layers of the LeNet-5 network, trained with MNIST, can be implemented in a spiking manner, and discuss the necessary mod-ifications to the architecture to offer this solution in real-time.
Performance analysis of CNNs deployed on low power arquitectures
(Springer, 2024-08) Pérez-Peña, Antonio Manuel; Linares Barranco, Alejandro; Casanueva-Morato, Daniel; Barranco, Francisco; Romero García, Samuel F.; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
Artificial intelligence is getting a more important role in our lives day by day. It is present in many daily used devices. In the majority of cases, the use of these algorithms implies having an internet connection so as to work due to their high computational demand. With the lately increase in the computing power of embedded systems, the execution of neural networks is focusing on the so-called edge computing. This favors the execution of neural network algorithms in environments where is not possible an internet connection. This article accomplishes an analysis of efficiency in the execution of a convolutional neural network (CNN) in a dedicated embedded device with graphical processing units (GPU) and inside two different multi-processors programmable system-on-chip with FPGA devices (MPSoC) from Xilinx. The first motivation for this work is the study of the viability of the deployment of an application to help visual functional diversity people with embedded systems not requiring an internet connection and oriented to Edge-AI, and the second is the search for a low-power device for longer battery life. At the end of the study, it is concluded that the deployment of these algorithms inside the selected low-power devices makes it possible to maintain a high frameper- second (FPS) rate in some architectures with an energy demand considerably inferior to the general-purpose devices.
Design, Implementation and Testing of Wearable System and Application for Sleep Study
(Springer, 2024-08) Pérez-Ornedo, Elena; Ponce-González, Carmen; Luna Perejón, Francisco; Domínguez Morales, Manuel Jesús; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
Sleep-related problems are currently an important field of study. Regular sleep drastically reduces mental illness. In this respect, more than half of the population wakes up with the feeling of not having had a good night's sleep, in many cases because they have sleep disorders that prevent them from getting a good night's sleep. One of the most common is obstructive sleep ap-noea which, if not treated in time, can lead to short- and long-term problems. However, only 10\% of the population is diagnosed. To perform a poly-somnographic study, it is necessary to undergo a sleep study session in a spe-cialised centre or to borrow a heavy and invasive device that obstructs one's own sleep and interferes with the study. Therefore, in this work we design, im-plement and evaluate a non-invasive wearable system for detecting sleep anom-alies, containing a user application that allows real-time monitoring of the pa-tient, saves the information in a database and detects various alarm markers.
Towards the design and prototyping of next generation Advance Warning Triangle V-16 and V-27 smart and connected devices
(Springer, 2024-08) Díaz-Romero, Santiago; Linares Barranco, Alejandro; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
Traditionally, vehicle breakdown signaling devices such as Advance Warning Triangles require the user to get out of the vehicle, which increases the risk of an accident. According to the Spanish Directorate-General for Traffic (DGT), a significant number of accidents occur in Spain while a vehicle is being repaired or waiting for roadside assistance. In order to try to reduce the numbers, the DGT has proposed some connected solutions such as the V-16 light signal and the V-27 virtual triangle. This paper proposes an approach to the software and hardware implementation of these devices. For the hardware design, a modular system based on MCU and the real-time operating system FreeRTOS has been chosen. This would allow power saving and future modifications or upgrades in simpler and remote way. On the other hand, the designed software platform allows to collect and process both the incidents generated by this device and those published by the DGT in order to send them to the V-27 application on an Android system. It has been possible to develop a modular hardware device with a relatively low cost and a potentially scalable functional software platform.
Design and development of a sensorised sports waistcoat
(Springer, 2024-08) Vicente-García, Javier; Muñoz Saavedra, Luis; Domínguez Morales, Manuel Jesús; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
In the field of sports, the monitoring of physiological parameters is essential to assess the physical performance of athletes and to detect possible health-related pathologies. With this information, it is possible to obtain data about the state of the body and the needs it requires. In 2022, according to data collected by the Statistics and Studies Division of the General Technical Secre-tariat of the Ministry of Culture and Sport together with the Higher Sports Council, sports practice in Spain has increased by almost 4% compared to 2015, with 57.3% of the Spanish population now having practised sport in the last year, either periodically or occasionally. While there are expensive solutions available for professional athletes, the cheap equivalent for amateur athletes is not available. This project involves the design and implementation of a weara-ble device attached to a T-shirt and composed of a low-cost microcontroller and various physiological sensors, whose information is transmitted to a mobile ap-plication to be monitored in real time and analysed, either by the user or by spe-cialists, in order to establish improvement parameters for future training, as well as to detect anomalies that require the intervention of specialists.
A computational approach to a neuromorphic sequential memory bio-inspired on the Hippocampus and Entorhinal Cortex formation
(Springer, 2024-08) Casanueva Morato, Daniel; Ayuso Martínez, Álvaro; Pérez-Peña, Antonio Manuel; Domínguez Morales, Juan Pedro; Jiménez Moreno, Gabriel; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Ministerio de Ciencia e Innovación (MICIN). España; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
The brain is considered one of the most powerful and efficient machines in existence. This is why neuromorphic engineering is trying to mimic biology to develop new systems that incorporate these superior capabilities. Within this field, bio-inspired learning and memory systems are still a challenge to be solved, and this is where the hippocampus and entorhinal cortex is involved. This brain formation acts as a short-term memory capable of learning by self-association of memories from different sources of information. In this work, we propose a fully functional bio-inspired spike-based hippocampus and entorhinal cortex sequential memory system capable of learning memories, recalling individual and sequential memories from a single fragment, and forgetting them. This model has been implemented on SpiNNaker using Spiking Neural Networks, and a set of experiments were performed to demonstrate its correct operation. This model will pave the road for the development of future more complex neuromorphic systems with a large amount of applicability.
Detection of gait-related activities from accelerometer signals obtained from the ankle
(Springer, 2024-08) Martín Romero, Ricardo Andrés; Civit Masot, Javier; Luna Perejón, Francisco; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
Accurate activity monitoring systems are essential to differentiate between everyday activities and events that could potentially endanger physical safety. This work analyzes the performance in terms of effectiveness of Deep Learning models, specifically Dense Neural Networks and Recurrent Neural Networks models, in the identification of different activities collected in the AnkFall dataset. This dataset contains different gait-related scenarios, including daily living activities, risk events and falls simulations, recorded with a triaxial accelerometer placed on the ankle. Different classification problems were considered, using different groupings of the activities contemplated in the dataset. To train the models, data augmentation techniques were employed to create modified versions of the original signals without causing excessive distortion. The results indicate difficulty in distinguishing between similar daily activities with minimal leg movement, as well as in recognizing fall risks in the event of a trip. However, promising performance was observed in distinguishing when the user walks while dizzy, which increases the risk of losing balance and falling.
A comprehensive Bird ecosystem monitorization and analysis tool from sound recognition using Machine Learning
(Springer, 2024-08) Vena Campos, Jesús; Muñoz Saavedra, Luis; Luna Perejón, Francisco; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
This project aims to develop a system for bird identification based on their songs, targeting ecosystem researchers, ornithologists, and nature enthusiasts. The system integrates the BirdNet model to process bird audio for continuous monitorization. It incorporates automatic mechanisms to record the audio from a capture device, pass it to the Machine Learning model, and store the data for later access and analysis. The system features a graphical interface via a web application, enhancing the user experience and presenting data in a clear format. The interface displays analyses in a table format with dates and times, and it offers dynamic statistics and charts that update with new analyses and data. This application can be scaled to contribute to the creation of a distributed system that collects data from broader ecosystems. As it is based on a widely used machine learning model, it can be integrated with other utilities that make use of this model.
Study and analysis of the heterogeneity of a prostate cancer dataset: first steps on the release of a multicenter strongly-annotated dataset
(Springer, 2024-08) Marrón-Esquivel, José M.; Durán López, Lourdes; Domínguez Morales, Juan Pedro; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Junta de Andalucía; European Commission (EC). Fondo Europeo de Desarrollo Regional (FEDER); Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
Prostate cancer (PCa) is a prevalent and deadly disease, necessitating advancements in diagnostic approaches with the advent of clinical digitization. This shift involves the application of Deep Learning and Convolution Neural Networks for automated PCa diagnosis through image processing. The Gleason scale is pivotal in assessing PCa aggressiveness, but the notable inter-observer variability among pathologists in assigning Gleason scores poses a challenge. To address this, leveraging diverse and sufficiently heterogeneous datasets becomes crucial for training artificial intelligence systems to generalize effectively. In this study, a pivotal step is taken by introducing a dataset comprising labeled PCa samples from three distinct medical centers. The outcomes unveil insights into image variability and establish a ground truth using supervised learning, serving as a benchmark for researchers. This initiative aims to empower AI systems to provide precise diagnoses across diverse PCa samples, ultimately enhancing clinical decision-making and mitigating observer-dependent discrepancies.
Design and development of a sensorised sports waistcoat
(Springer, 2024-08) Vicente-García, Javier; Muñoz Saavedra, Luis; Domínguez Morales, Manuel Jesús; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
In the field of sports, the monitoring of physiological parameters is essential to assess the physical performance of athletes and to detect possible health-related pathologies. With this information, it is possible to obtain data about the state of the body and the needs it requires. In 2022, according to data collected by the Statistics and Studies Division of the General Technical Secre-tariat of the Ministry of Culture and Sport together with the Higher Sports Council, sports practice in Spain has increased by almost 4% compared to 2015, with 57.3% of the Spanish population now having practised sport in the last year, either periodically or occasionally. While there are expensive solutions available for professional athletes, the cheap equivalent for amateur athletes is not available. This project involves the design and implementation of a weara-ble device attached to a T-shirt and composed of a low-cost microcontroller and various physiological sensors, whose information is transmitted to a mobile ap-plication to be monitored in real time and analysed, either by the user or by spe-cialists, in order to establish improvement parameters for future training, as well as to detect anomalies that require the intervention of specialists.
A feedback sensor based on spiking neural networks for real-time robot adaption
(Springer, 2024-08) López Osorio, Pablo; Pérez Peña, Fernando; Domínguez Morales, Juan Pedro; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Ministerio de Ciencia e Innovación (MICIN). España; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
In recent years, it has been observed how the locomotion systems of vertebrate animals serve as inspiration to enhance the performance of robotic systems. These animal systems are characterized by their ability to adapt to environmental changes detected by their biological sensors. With this model in mind, our objective is to replicate this adaptability in robotics through the use of a Central Pattern Generator (CPG). We present an advanced robotic system based on Spiking Neural Networks (SNNs), implemented in both Spinnaker and Field Programmable Gate Arrays (FPGAs). This system is capable of modifying its locomotion pattern based on information provided by Force Sensitive Resistors (FSRs). Our experiments demonstrate that this platform can adapt in real-time to the terrain in which it operates.
Smart cane for the visually impaired
(Springer, 2024-08) Paredes Orozco, Diana Lucía; Domínguez Morales, Manuel Jesús; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores
Visual impairment is understood as the partial or total loss of sight even with the use of optical aids, impacting notoriously on daily life, most of the time making them dependent on others, with mobility being one of the most affected aspects. There are support tools to facilitate it, being the white cane the most used; however, it has a series of disadvantages such as the detection range and the lack of attention of the upper body plane. This work presents the design and development of a functional prototype of a smart cane, which adapts to the environment of the blind person and solves the problems previously mentioned, as well as a mobile application to which it is connected as a guide and localisation aid in case of accident. The results show a reliability of 98% in the obstacle detection distance, with a vertical detection angle of 13.2% and a horizontal detection angle of 21.46%, as well as a correct activation of the alarm states in the event of falls, warning the family member and indicating the GPS position; all of this with a battery life of 32 hours.
Indoor Positioning System for Nursing Homes
(Springer, 2024-08) Camacho-del-Valle, Azucena; Santos-Gómez, Manuel; Jiménez Fernández, Ángel Francisco; Domínguez Morales, Manuel Jesús; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores
Global positioning systems using satellites have proven to be extremely useful, not only for route guidance, but also for knowing the exact whereabouts in the event of an accident or emergency. However, this system has major shortcomings indoors due to the error it contains (between 3 and 10 meters depending on the number of visible satellites), but mostly due to the weak signal obtained indoors. This is a challenge in large buildings where the exact location of a person needs to be known: for example, in nursing homes, where the number of caregivers does not allow all the elderly to be always monitored. This is why, in this type of situation, the usual Global Positioning System is dispensed with, and local solutions based on cameras or sensor networks are used. In this work, a study is carried out on indoor positioning using wearable devices attached to the people to be located, which serve as positioning beacons and communicate via Zigbee networks with a series of fixed nodes located in the rooms. Unlike camera control, it is a non-intrusive mechanism and the system's power consumption is significantly reduced. The results of the tests carried out in a house with 3 rooms and 2 fixed nodes demonstrate its correct operation in real time.
Prediction of the microinvasion of melanoma using supervised deep convolutional neural networks
(Springer, 2024-08) Durán López, Lourdes; Hernández-Rodríguez, Juan Carlos; Domínguez Morales, Juan Pedro; Pereyra-Rodríguez, José-Juan; Universidad de Sevilla. Departamento de Medicina; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores; Junta de Andalucía; European Commission (EC). Fondo Europeo de Desarrollo Regional (FEDER); Academia Española de Dermatología y Venereología; Universidad de Sevilla. TEP108: Robótica y Tecnología de Computadores; Universidad de Sevilla. CTS-1088: Enfermedades Inmunomediadas (Sistema Sanitario Público de Andalucía. Fundación Pública Andaluza para la Gestión de la Investigación en Salud de Sevilla (FISEVI))
Melanoma poses a significant global health threat, contributing to over 90% of skin cancer-related deaths. Dermoscopy aids in early melanoma diagnosis, but distinguishing between in situ and invasive melanomas remains challenging, even for experienced dermatologists. Recent strides in artificial intelligence (AI) for medical image analysis suggest its potential to support and offer a secondary opinion to dermatologists in dermoscopy. Utilizing annotated images is crucial for Deep Learning systems to learn associations between images and ground-truth labels. This study employs datasets from diverse sources to train and evaluate Deep Transfer Learning algorithms, specifically using a DenseNet121 convolutional neural network, to the Breslow thickness of melanomas. Supervised learning, employing a stratified 5-fold cross-validation technique, demonstrates strong performance in this melanoma classification task. The research underscores the potential of AI-based automatic diagnosis systems as valuable aids for medical professionals, serving as secondary opinions or triage tools in healthcare settings.
Potenciación de las competencias transversales en titulaciones técnicas
(Servicio de Publicaciones de la UEX, 2007) Fernández Vega, Francisco; Díaz Álvarez, Josefa; Guisado Lizar, José Luis; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores
En este artículo presentamos una experiencia docente en el marco del nuevo Espacio Europeo de Educación Superior: la potenciación de las competencias transversales en un Plan Piloto de Ingeniería Técnica Informática de Sistemas (ITIS). Para ello se seleccionó una asignatura de tercer curso de ITIS, Sistemas Operativos, en la que los alumnos han podido trabajar competencias transversales de carácter interdisciplinar y multidisciplinar. Además, se seleccionaron algunos de los alumnos matriculados en la asignatura con la idea de promover experiencias extra-curriculares que permitieran a los mismos afrontar situaciones reales que encontrarán en su futuro profesional. Con este objetivo se planteó la implementación de dos videojuegos con la participación, además de profesionales externos, de alumnos de 4 titulaciones diferentes. Como reconocimiento a la labor desarrollada los videojuegos han sido integrados en una distribución Linux y como consecuencia de ello se ha conseguido motivar en mayor grado a los alumnos de la titulación universitaria, utilizando el videojuego como práctica en alguna de las asignaturas del plan de estudios.
Planificación de la experiencia piloto para la adaptación del segundo curso de Ingeniería Técnica en Informática de Sistemas de la Universidad de Extremadura al Espacio Europeo de Educación Superior
(Universidad de Extremadura, 2006) Guisado Lizar, José Luis; Bravo Trinidad, José Luis; Astillero, Antonio; Cwierz, Halina; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores
La universidad europea se encuentra inmersa en un proceso de convergencia hacia un Espacio Europeo de Educación Superior que posibilite el reconocimiento mutuo de titulaciones y el establecimiento de equivalencias, competencias y cualificaciones comunes y que fomente un cambio tanto en las metodologías y actividades docentes como en las actitudes y expectativas del alumnado. La Universidad de Extremadura, anticipándose a estos cambios, llevará a cabo durante el curso académico 2006/2007 una experiencia piloto en el segundo curso de la titulación de Ingeniería Técnica en Informática de Sistemas del Centro Universitario de Mérida, que continuará a la realizada durante el curso anterior en el primer curso. En este trabajo presentamos los planes docentes elaborados, según la metodología ECTS, de las asignaturas de segundo curso recogidas en el actual plan de estudio, así como la agenda del estudiante, atendiendo a las actividades programadas por cada asignatura y a la distribución del horario.
Remote Services for Advanced Problem Optimization
(John Wiley & Sons, Inc., 2008) Gómez, J.A.; Vega, M.A.; Sánchez, J.M.; Guisado Lizar, José Luis; Lombraña, D.; Fernández, F.; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores
In this chapter we present some remote services for advanced problem optimization. These tools permit us to get a set of resources useful to approach problems easily using the Internet as a channel of experimentation with the optimization algorithms involved.
Application of Cellular Automata Algorithms to the Parallel Simulation of Laser Dynamics
(John Wiley & Sons, Inc., 2008) Guisado Lizar, José Luis; Jiménez Morales, Francisco; Guerra, J.M.; Fernández, F.; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores
In this chapter we review the use of a biologically inspired heuristic technique— cellular automata (CA)—as a problem solver for one of the most paradigmatic complex systems: the laser. CAs are a class of mathematical system that can be used to model spatiotemporal phenomena, characterized by the discreteness of all of its variables: space, time, and normally state variables. An important property of CAs is their intrinsic parallel character. Therefore, they are specially well suited to be implemented very efficiently on parallel computers.
Simulating Laser Dynamics with Cellular Automata
(Springer Verlag, 2018) Jiménez Morales, Francisco; Guisado Lizar, José Luis; Guerra, José Manuel; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores
The long-established approach to study laser dynamics uses a set of differential equations known as the laser rate equations. In this work we present an overview of an alternative model based on a cellular automaton (CA). We also present a panorama of different variants of the model: the original one, designed to simulate general laser dynamics; an additional one, that was proposed to simulate pulsed pumped lasers; and finally a new model to simulate lasers that exhibit antiphase dynamics, which is proposed here. Despite its simplicity, the CA model reproduces qualitatively the phenomenology encountered in many real laser systems: (i) the existence of a threshold value of the pumping rate Rt;(ii)the exact dependence of Rt on the life times of the photons and the inversion population; (iii) the two main laser regimes: a steady state and an oscillatory one
Derecho de las niñas y niños de educación secundaria de la escuela Don Bosco Mbuji Mayi a una educación de calidad y con enfoque de género a través de una mejor formación y mayor motivación en las asignaturas de las área de conocimiento STEM (ciencia, tecnología, ingeniería y matemáticas)
(Comares, 2023) Alexandre, Betsaida; Paillacho Bolaños, Lisbeth; Pérez López, José Ángel; García Robles, Rocío; Universidad de Sevilla. Departamento de Tecnología Electrónica; Universidad de Sevilla. Departamento de Economía Financiera y Dirección de Operaciones; Universidad de Sevilla. Departamento de Contabilidad y Economía Financiera; Universidad de Sevilla. Departamento de Arquitectura y Tecnología de Computadores
Este proyecto tiene como objetivo mejorar la calidad de la educación de las niñas y niños del primer y segundo curso de la escuela secundaria Don Bosco Mbuji Mayi en la República Democrática del Congo a través de la enseñanza de programación y robótica, promoviendo el interés de los chicos y especialmente de las chicas en las áreas de conocimiento STEM (Ciencia, Tecnología, Ingeniería y Matemáticas), de tal forma que al finalizar el segundo curso opten por seguir la especialidad de Mecánica General que ofrece esta escuela secundaria pues actualmente dicha especialidad cuenta con pocos alumnos y ninguna chica, promoviendo que a futuro cursen estudios universitarios en grados de ingeniería. Así mismo se prevé contribuir a que las niñas y niños mejoren su rendimiento escolar en las asignaturas STEM asociadas disminuyendo así el índice de repitientes pues se ha detectado que casi el 50% de los alumnos del primer curso no logran aprobar. Se plantea integrar la robótica educativa dentro de la asignatura de «Tecnología» que es obligatoria en ambos cursos y cuyo contenido programático se imparte solo de manera teórica ya que no cuentan con laboratorios equipados ni otro tipo de elementos educativos que les permita un mejor aprendizaje.

Examinar

Envíos recientes