ESTCB - Artigos em revistas com arbitragem científica
Permanent URI for this collection
Browse
Recent Submissions
- Computer vision for reducing food waste in an institutional canteen: A literature review and performance analysisPublication . Correia, Ana; Aidos, Clara; Caldeira, J.M.L.P.; Soares, V.N.G.J.Food waste in today's society has been the subject of growing interest and discussion, given its economic, environmental, social, and nutritional implications. Although food waste is present throughout the food supply chain, in developed countries it tends to be higher in the final stages of consumption (e.g., households and food services). This study focuses on institutional canteens, where food waste includes prepared meals that have not been sold (i.e., leftovers), as well as food served that is left on plates after the meal has been consumed (i.e., scraps). It presents a first step towards developing a prototype/solution based on computer vision techniques to identify and quantify food waste in an institutional canteen. It begins by introducing the related concepts. It then surveys the state-of-the-art and categorizes existing solutions, presenting their main characteristics, strengths, and limitations. Inception-V3 and ResNet-50 are identified as the most promising computer vision techniques, and their performance has been evaluated. Information is also provided on open questions and research directions in this area
- Indoor microclimate monitoring and forecasting: Public Sector building use casePublication . Sudniks, Ruslans; Ziemelis, Arturs; Nikitenko, Agris; Soares, V.N.G.J.; Supe, AndisThis research aims to demonstrate a machine learning (ML) algorithm-based indoor air quality (IAQ) monitoring and forecasting system for a public sector building use case. Such a system has the potential to automate existing heating/ventilation systems, therefore reducing energy consumption. One of Riga Technical University’s campus buildings, equipped with around 128 IAQ sensors, is used as a test bed to create a digital shadow including a comparison of five ML-based data prediction tools. We compare the IAQ data prediction loss using Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) error metrics based on real sensor data. Gated Recurrent Unit (GRU) and Kolmogorov–Arnold Networks (KAN) prove to be the most accurate models regarding the prediction error. Also, GRU proved to be the most efficient model regarding the required computation time.
- Generative jazz chord progressions: A statistical approach to harmonic creativityPublication . Raposo, Adriano; Soares, V.N.G.J.Jazz music has long been a subject of interest in the field of generative music. Traditional jazz chord progressions follow established patterns that contribute to the genre’s distinct sound. However, the demand for more innovative and diverse harmonic structures has led to the exploration of alternative approaches in music generation. This paper addresses the challenge of generating novel and engaging jazz chord sequences that go beyond traditional chord progressions. It proposes an unconventional statistical approach, leveraging a corpus of 1382 jazz standards, which includes key information, song structure, and chord sequences by section. The proposed method generates chord sequences based on statistical patterns extracted from the corpus, considering a tonal context while introducing a degree of unpredictability that enhances the results with elements of surprise and interest. The goal is to move beyond conventional and well-known jazz chord progressions, exploring new and inspiring harmonic possibilities. The evaluation of the generated dataset, which matches the size of the learning corpus, demonstrates a strong statistical alignment between distributions across multiple analysis parameters while also revealing opportunities for further exploration of novel harmonic pathways.
- Hybrid B5G-DTN architecture with federated learning for contextual communication offloadingPublication . Jesús-Azabal, Manuel; Zheng, Meichun; Soares, V.N.G.J.In dense urban environments and large-scale events, Internet infrastructure often becomes overloaded due to high communication demand. Many of these communications are local and short-lived, exchanged between users in close proximity but still relying on global infrastructure, leading to unnecessary network stress. In this context, delay-tolerant networks (DTNs) offer an alternative by enabling device-to-device (D2D) communication without requiring constant connectivity. However, DTNs face significant challenges in routing due to unpredictable node mobility and intermittent contacts, making reliable delivery difficult. Considering these challenges, this paper presents a hybrid Beyond 5G (B5G) DTN architecture to provide private context-aware routing in dense scenarios. In this proposal, dynamic contextual notifications are shared among relevant local nodes, combining federated learning (FL) and edge artificial intelligence (AI) to estimate the optimal relay paths based on variables such as mobility patterns and contact history. To keep the local FL models updated with the evolving context, edge nodes, integrated as part of the B5G architecture, act as coordinating entities for model aggregation and redistribution. The proposed architecture has been implemented and evaluated in simulation testbeds, studying its performance and sensibility to the node density in a realistic scenario. In high-density scenarios, the architecture outperforms state-of-the-art routing schemes, achieving an average delivery probability of 77%, with limited latency and overhead, demonstrating relevant technical viability.
- Determination of dynamic elastic properties of 3D-printed nylon 12CF using impulse excitation of vibrationPublication . Garcia, Pedro F.; Ramalho, Armando; Vasco, Joel; Rubén, Rui; Capela, Carlos; MDPIMaterial Extrusion (MEX) process is increasingly used to fabricate components for structural applications, driven by the availability of advanced materials and greater industrial adoption. In these contexts, understanding the mechanical performance of printed parts is crucial. However, conventional methods for assessing anisotropic elastic behavior often rely on expensive equipment and time-consuming procedures. The aim of this study is to evaluate the applicability of the impulse excitation of vibration (IEV) in characterizing the dynamic mechanical properties of a 3D-printed composite material. Tensile tests were also performed to compare quasi-static properties with the dynamic ones obtained through IEV. The tested material, Nylon 12CF, contains 35% short carbon fibers by weight and is commercially available from Stratasys. It is used in the fused deposition modeling (FDM) process, a Material Extrusion technology, and exhibits anisotropic mechanical properties. This is further reinforced by the filament deposition process, which affects the mechanical response of printed parts. Young’s modulus obtained in the direction perpendicular to the deposition plane (E33), obtained via IEV, was 14.77% higher than the value in the technical datasheet. Comparing methods, the Young’s modulus obtained in the deposition plane, in an inclined direction of 45 degrees in relation to the deposition direction (E45), showed a 22.95% difference between IEV and tensile tests, while Poisson’s ratio in the deposition plane (v12) differed by 6.78%. This data is critical for designing parts subject to demanding service conditions, and the results obtained (orthotropic elastic properties) can be used in finite element simulation software. Ultimately, this work reinforces the potential of the IEV method as an accessible and consistent alternative for characterizing the anisotropic properties of components produced through additive manufacturing (AM).
- Dynamic elastic properties of E-Glass randomly oriented fiber reinforced SR GreenPoxy composite - Experimental and numerical analysisPublication . Ramalho, Armando; Gaspar, Marcelo; Correia, Mário; Vasco, Joel; Capela, Carlos; Rubén, RuiIn this article, the in-plane dynamic elastic properties of an E-glass randomly oriented fiber-reinforced SR GreenPoxy 56 composite were obtained based on the procedure specified in the ASTM E1876-21 standard. The experimental frequencies and the ones predicted by the simulation of the experimental procedure using a finite element analysis developed in the Patran/Nastran 2021 package were used in an iterative algorithm using sensitivity analysis to improve the first approaches of the dynamic elastic properties obtained by the impulse excitation technique. These experimental results are compared with the ones obtained by the 2D Short Fiber Composite model of the E-glass randomly oriented fiber-reinforced SR GreenPoxy 56 composite, developed in the Patran/Nastran 2021 software.
- Using drones to estimate and reduce the risk of wildfire propagation in wildland–urban interfacesPublication . Santos, Osvaldo; Santos, NatérciaForest fires have become one of the most destructive natural disasters worldwide, causing catastrophic losses, sometimes with the loss of lives. Therefore, some countries have created legislation to enforce mandatory fuel management within buffer zones in the vicinity of buildings and roads. The purpose of this study is to investigate whether inexpensive off-the-shelf drones equipped with standard RGB cameras could be used to detect the excess of trees and vegetation within those buffer zones. The methodology used in this study was the development and evaluation of a complete system, which uses AI to detect the contours of buildings and the services provided by the CHAMELEON bundles to detect trees and vegetation within buffer zones. The developed AI model is effective at detecting the building contours, with a mAP50 of 0.888. The article analyses the results obtained from two use cases: a road surrounded by dense forest and an isolated building with dense vegetation nearby. The main conclusion of this study is that off-the-shelf drones equipped with standard RGB cameras can be effective at detecting non-compliant vegetation and trees within buffer zones. This can be used to manage biomass within buffer zones, thus helping to reduce the risk of wildfire propagation in wildland–urban interfaces.
- ROBIN: Reference observatory of basins for international hydrological climate change detectionPublication . Turner, S.; Hannaford, J.; Barker, L. J.; Suman, G.; Killeen, A.; Armitage, R.; Chan, W.; et al.; Albuquerque, M.T.D.Human-induced warming is modifying the water cycle. Adaptation to posed threats requires an understanding of hydrological responses to climate variability. Whilst these can be computationally modelled, observed streamflow data is essential for constraining models, and understanding and quantifying emerging trends in the water cycle. To date, the identification of such trends at the global scale has been hindered by data limitations – in particular, the prevalence of direct human influences on streamflow which can obscure climate-driven variability. By removing these influences, trends in streamflow data can be more confidently attributed to climate variability. Here we describe the Reference Observatory of Basins for INternational hydrological climate change detection (ROBIN) – the first iteration of a global network of streamflow data from national reference hydrological networks (RHNs) – comprised of catchments which are near-natural or have limited human influences. This collaboration has established a freely available global RHN dataset of over 3,000 catchments and code libraries, which can be used to underpin new science endeavours and advance change detection studies to support international climate policy and adaptation.
- Evacuation of Lisbon’s Baixa-Chiado subway station in case of firePublication . Borralho, Tiago; Rodrigues, João Paulo; Calmeiro dos Santos, CristinaIt is essential to ensure that any building has conditions for a safe evacuation of its occupants. This aspect is essential in subway stations, where evacuation has to be carried out in an upward way, and usually correspond to large structures constituting a single fire compartment. Baixa-Chiado subway station, in Lisbon, Portugal, was selected for studying the evacuation in case of fire, due to its depth, high number of passengers that frequent the station and the existing of two intersecting train lines. A calculation of evacuation time was calculated and the way of evacuation studied, in different fire scenarios, number and location of occupants. The numerical simulations used Fire Dynamics Simulator and Pathfinder softwares, the first for fire spreading and the second for evacuation analysis. The importance of smoke control system, and its rapid activation in case of fire, was highlighted by the results obtained. In situations where this did not occur, there was a significant worsening in the evacuation of the occupants. It was estimated the incapacitation of a significant number of occupants, considering the levels registered for the fractional effective dose. The station’s architectural constraints proved to be a crucial factor in the results of the study. This article highlights important results applicable to subway stations around the world.
- In-bed posture classification using pressure data from a sensor sheet under the mattressPublication . Serra, André; Ribeiro, Fernando Reinaldo; Metrôlho, J.C.M.M.Monitoring and controlling the condition of bedridden individuals can help reduce health risks, as improper nocturnal habits or body positioning can exacerbate issues such as apnea, insomnia, sleep disorders, spinal problems, and pressure ulcers. Techniques using pressure maps from sensors placed on top of the mattress, along with machine learning (ML) algorithms to classify main postures (prone, supine, left side, right side), have achieved up to 99% accuracy. This study evaluated the feasibility of using a sensor sheet placed under the mattress to minimize patient discomfort. Experiments with ten commonly used ML algorithms achieved average accuracy values ranging from 79.14% to 98.93% using K-Fold cross-validation and from 80.03% to 97.14% using Leave-One-Group-Out (LOGO) for classifying the four main postures. The classification was extended to include 28 posture variations (7 variations for each of the 4 main postures), with the SVM algorithm achieving an accuracy of 65.18% in K-Fold validation, marking a significant improvement over previous studies, particularly regarding the number of postures considered. Comparisons with previous studies that used pressure sensors placed both under and on top of the mattress show that this approach achieves comparable accuracy to other methods, surpassing them with some algorithms and achieving the highest average accuracy. In conclusion, using sensors under the mattress is an effective and less invasive alternative for posture classification.