TECNALIA Publications
Welcome to TECNALIA Publications, the Institutional Repository of Fundación TECNALIA Research & Innovation. TECNALIA Publications' main objective is to collect, preserve and disseminate the intellectual production resulting from the research activity of TECNALIA to generate transferable knowledge and contribute thereby to development and social progress.
Recent Submissions
ArmAssist:
A Telerehabilitation Solution for Upper-Limb Rehabilitation at Home
(2023-03-01) Garzo, Ainara; Jung, Je H.; Arcas-Ruiz-Ruano, Javier; Perry, Joel C.; Keller, Thierry; Medical Technologies; Tecnalia Research & Innovation
ArmAssist (AA), developed by TECNALIA, is a telerehabilitation platform aiming to help poststroke subjects maintain the rehabilitation of the upper limb at home. AA includes robotic modules with multiple sensors to train and measure the users' voluntary movements. An assessment platform based on serious games is also included to not only engage the user but also perform automated evaluations of arm and hand function and their evolution over time. Moreover, AA facilitates at-home rehabilitation with limited remote supervision by the therapist. In the present article, the technical specifications and developments of AA are described. Additionally, a summary of the outcomes of a usability evaluation of AA is presented.
Nitrogen-vacancy center magnetic imaging of Fe3O4 nanoparticles inside the gastrointestinal tract of Drosophila melanogaster
(2023-12-05) Mathes, Niklas; Comas, Maria; Bleul, Regina; Everaert, Katrijn; Hermle, Tobias; Wiekhorst, Frank; Knittel, Peter; Sperling, Ralph A.; Vidal, Xavier; Quantum
Widefield magnetometry based on nitrogen-vacancy centers enables high spatial resolution imaging of magnetic field distributions without a need for spatial scanning. In this work, we show nitrogen-vacancy center magnetic imaging of Fe3O4 nanoparticles within the gastrointestinal tract of Drosophila melanogaster larvae. Vector magnetic field imaging based on optically detected magnetic resonance is carried out on dissected larvae intestine organs containing accumulations of externally loaded magnetic nanoparticles. The distribution of the magnetic nanoparticles within the tissue can be clearly deduced from the magnetic stray field measurements. Spatially resolved magnetic imaging requires the nitrogen-vacancy centers to be very close to the sample making the technique particularly interesting for thin tissue samples. This study is a proof of principle showing the capability of nitrogen-vacancy center magnetometry as a technique to detect magnetic nanoparticle distributions in Drosophila melanogaster larvae that can be extended to other biological systems.
Active Coated PLA-PHB Film with Formulations Containing a Commercial Olive Leaf Extract to Improve Quality Preservation of Fresh Pork Burgers
(2023) Fiorentini, Cecilia; Bassani, Andrea; Zaccone, Marta; Montalbano, Maria Luana; De Apodaca, Elena Díaz; Spigno, Giorgia; Alimentación Sostenible
Since fresh meat is often subject to several degradation reactions that decrease its safety and quality until it is considered unacceptable, the release of bioactive compounds into meat products may be a good option to slow down oxidation and extend its shelf-life by a few days. Therefore, this study aimed to test the application on fresh pork burgers of a PLA-PHB film coated with two different coating formulations (methylcellulose, MC and chitosan, CT), both containing a commercial olive leaf extract OL, to evaluate their effect on meat quality preservation. Samples were tested at 0, 2, 5, 7, 9, 12, and 14 days after packing for microbial, chemical, and sensory evaluations. Except for the chitosan-only formulation, all tested formulations (MC, MC+OL and CT+OL) adhered well to the PLA-PHB base without the use of specific treatments. Meat packed with the different coatings maintained a slightly brighter red colour than the control samples and, as a result, deteriorated more slowly. In the evaluation of lipid oxidation, the CT+OL coating showed lower mean values of mg MDA/kg meat, which were significantly different from the other samples, especially on the 7th and 9th day of storage. Moreover, the CT+OL coating showed a slight slowdown in Enterobacteriaceae growth, revealing promising results in maintaining the meat quality longer.
Multiobjective evolutionary pruning of Deep Neural Networks with Transfer Learning for improving their performance and robustness
(2023-11) Poyatos, Javier; Molina, Daniel; Martínez-Seras, Aitor; Del Ser, Javier; Herrera, Francisco; IA
Evolutionary Computation algorithms have been used to solve optimization problems in relation with architectural, hyper-parameter or training configuration, forging the field known today as Neural Architecture Search. These algorithms have been combined with other techniques such as the pruning of Neural Networks, which reduces the complexity of the network, and the Transfer Learning, which lets the import of knowledge from another problem related to the one at hand. The usage of several criteria to evaluate the quality of the evolutionary proposals is also a common case, in which the performance and complexity of the network are the most used criteria. This work proposes MO-EvoPruneDeepTL, a multi-objective evolutionary pruning algorithm. MO-EvoPruneDeepTL uses Transfer Learning to adapt the last layers of Deep Neural Networks, by replacing them with sparse layers evolved by a genetic algorithm, which guides the evolution based in the performance, complexity and robustness of the network, being the robustness a great quality indicator for the evolved models. We carry out different experiments with several datasets to assess the benefits of our proposal. Results show that our proposal achieves promising results in all the objectives, and direct relation are presented among them. The experiments also show that the most influential neurons help us explain which parts of the input images are the most relevant for the prediction of the pruned neural network. Lastly, by virtue of the diversity within the Pareto front of pruning patterns produced by the proposal, it is shown that an ensemble of differently pruned models improves the overall performance and robustness of the trained networks.
Definition and design of a prefabricated and modular façade system to incorporate solar harvesting technologies
(2023) Alvarez-Alava, Izaskun; Elguezabal, Peru; Jorge, Nuria; Armijos-Moya, Tatiana; Konstantinou, Thaleia; ECOEFICIENCIA DE PRODUCTOS DE CONSTRUCCIÓN
The current research presents the design and development of a prefabricated modular façade solution for renovating residential buildings. The system is conceived as an industrialised solution that incorporates solar harvesting technologies, contributing to reducing energy consumption by employing an “active façade” concept. One of the main challenges was to achieve a highly flexible solution both in terms of geometry and enabling the incorporation of different solar-capturing devices (photovoltaic, thermal, and hybrid). Therefore, to be able to provide alternative customised configurations that can be fitted to various building renovation scenarios. Guided by the requirements and specifications, the design was defined after an iterative process, concluding with a final system design validated and adopted as viable for the intended purpose. A dimensional study for interconnecting all the technologies composing the system was carried out. Potential alternative configurations were assessed under the modularity and versatility perspective, resulting in a set of alternative combinations that better fit the established requirements. Complementarily, the system also integrates an active window solution a component that incorporates an autonomous energy recovery system through ventilation. The main outcome is explicated in a highly versatile modular façade system, which gives existing buildings the possibility to achieve Nearly Zero Energy Building requirements.
Zero Waste Binder Jetting Process:
Study Of The Reusability Of The Rejected Part Powder
(2023) Lores, Asier; Azurmendi, Naiara; Leizaola, Iñaki; Agote, Iñigo; EXTREMAT
Binder Jetting Additive Manufacturing is renowned for its high powder reusability ratio, reaching near 100 % for certain alloys. This technology offers an advantage over thermal-based methods, such as laser or electron beams, which can degrade or sinter the surrounding powder particles, diminishing the reusability ratio. However, during the setup and production of parts, defects may occur, leading to the direct rejection of certain printed components. The objective of this work is to investigate the feasibility and impact of reusing powder from rejected parts, aiming to achieve zero waste and maximize the utilization of metallic powders. To accomplish this, the rejected parts underwent a series of processes including debinding, sieving, characterization, and subsequent mixing with virgin powder at various proportions. The goal was to identify suitable mixing ranges that would enable the production of high-quality material. The study focused specifically on the 17-4PH alloy and confirmed that employing separate debinding and sintering processes facilitates the complete recirculation of all used powder, thus achieving zero waste. This finding highlights the potential for implementing a closed-loop system that maximizes powder reusability and minimizes material waste in the Binder Jetting Additive Manufacturing process.
Battery Lifetime Extension in a Stand-Alone Microgrid With Flexible Power Point Tracking of Photovoltaic System
(2023-04-01) Yan, Hein Wai; Farivar, Glen G.; Beniwal, Neha; Tafti, Hossein Dehghani; Ceballos, Salvador; Pou, Josep; Konstantinou, Georgios; POWER ELECTRONICS AND SYSTEM EQUIPMENT
In stand-alone dc microgrids (dcMGs), battery energy storage systems (BESSs) are conventionally used for regulating the dc-link voltage, causing a continuous battery operation. Though operating the photovoltaic (PV) system at its maximum power point (MPP) yields minimum battery discharge current, the opposite is true for battery charging current. Therefore, reducing the battery charging current based on its state-of-charge (SoC) and the amount of available PV surplus power (which can be treated as virtually stored energy) is an opportunity for improving the battery life. The main objective of the control strategy proposed in this article is to prolong the battery lifetime by reducing the charging current and keeping the battery SoC at lower values if the PV power is enough to supply the loads. Additionally, the PV system is used as the primary asset to regulate the microgrid voltage. The dynamic performance of the proposed control strategy is validated with experimental tests under various operating conditions. Furthermore, its effectiveness in prolonging the battery lifetime is evaluated using an aging model of a lithium-ion (Li-ion) battery (without loss of generality) by simulated case studies.
A framework to improve urban accessibility and environmental conditions in age-friendly cities using graph modeling and multi-objective optimization
(2023-06) Delgado-Enales, Iñigo; Del Ser, Javier; Molina, Patricia; IA
The rapid growth of cities in recent decades has unleashed several challenges for urban planning, which have been exacerbated by their aging population. Among the most pressing problems in cities are those related to mobility and environmental quality, by which a global concern has flourished around enhancing pedestrian accessibility for both environmental and health-related reasons. To tackle this issue, this paper presents a new framework that combines multi-objective optimization with a graph model that aims to support urban planning and management to enhance age-friendly cities. The framework allows designing urban projects that improve accessibility and reduce noise and/or air pollution through the installation of urban elements (ramps and escalators, elevators, acoustic and vegetation panels), while considering the overall economic cost of the installation. To explore the trade-off between these objectives, we resort to multi-objective evolutionary algorithms, which permit to compute near Pareto-optimal interventions over the graph model of the urban area under study. We showcase the applicability of the proposed framework over two use cases in the city of Barcelona (Spain), both quantitatively and qualitatively. Results evince that the framework can help urban planners make informed decisions towards enhancing urban accessibility and the environmental quality of age-friendly cities.
Region-based evidential deep learning to quantify uncertainty and improve robustness of brain tumor segmentation
(2023-10) Li, Hao; Nan, Yang; Del Ser, Javier; Yang, Guang; IA
Despite recent advances in the accuracy of brain tumor segmentation, the results still suffer from low reliability and robustness. Uncertainty estimation is an efficient solution to this problem, as it provides a measure of confidence in the segmentation results. The current uncertainty estimation methods based on quantile regression, Bayesian neural network, ensemble, and Monte Carlo dropout are limited by their high computational cost and inconsistency. In order to overcome these challenges, Evidential Deep Learning (EDL) was developed in recent work but primarily for natural image classification and showed inferior segmentation results. In this paper, we proposed a region-based EDL segmentation framework that can generate reliable uncertainty maps and accurate segmentation results, which is robust to noise and image corruption. We used the Theory of Evidence to interpret the output of a neural network as evidence values gathered from input features. Following Subjective Logic, evidence was parameterized as a Dirichlet distribution, and predicted probabilities were treated as subjective opinions. To evaluate the performance of our model on segmentation and uncertainty estimation, we conducted quantitative and qualitative experiments on the BraTS 2020 dataset. The results demonstrated the top performance of the proposed method in quantifying segmentation uncertainty and robustly segmenting tumors. Furthermore, our proposed new framework maintained the advantages of low computational cost and easy implementation and showed the potential for clinical application.
Analysing Edge Computing Devices for the Deployment of Embedded AI
(2023-12) Garcia-Perez, Asier; Miñón, Raúl; Torre-Bastida, Ana I.; Zulueta-Guerrero, Ekaitz; HPA
In recent years, more and more devices are connected to the network, generating an overwhelming amount of data. This term that is booming today is known as the Internet of Things. In order to deal with these data close to the source, the term Edge Computing arises. The main objective is to address the limitations of cloud processing and satisfy the growing demand for applications and services that require low latency, greater efficiency and real-time response capabilities. Furthermore, it is essential to underscore the intrinsic connection between artificial intelligence and edge computing within the context of our study. This integral relationship not only addresses the challenges posed by data proliferation but also propels a transformative wave of innovation, shaping a new era of data processing capabilities at the network’s edge. Edge devices can perform real-time data analysis and make autonomous decisions without relying on constant connectivity to the cloud. This article aims at analysing and comparing Edge Computing devices when artificial intelligence algorithms are deployed on them. To this end, a detailed experiment involving various edge devices, models and metrics is conducted. In addition, we will observe how artificial intelligence accelerators such as Tensor Processing Unit behave. This analysis seeks to respond to the choice of a device that best suits the necessary AI requirements. As a summary, in general terms, the Jetson Nano provides the best performance when only CPU is used. Nevertheless the utilisation of a TPU drastically enhances the results.
