Publications Repository of the Polytechnic Institute of Bragança
Not a member yet
30168 research outputs found
Sort by
Comparing RL policies for robotic pusher
No full textReinforcement learning (RL) has been consolidated as a promising approach to optimizing robotic tasks, allowing the improvement of performance and energy efficiency. This study investigates the effectiveness of five RL algorithms in the Pusher environment. Advantage Actor-Critic (A2C), Proximal Policy Optimization (PPO), Deep Deterministic Policy Gradient (DDPG), Soft Actor-Critic (SAC) and Twin Delayed Deep Deterministic Policy Gradient (TD3). We evaluated training time, computational efficiency, and reward values to identify the most balanced solution between accuracy and energy consumption. The results indicate that the PPO offers the best compromise between performance and efficiency, with reduced training time and stability in learning. SAC achieves the best rewards but requires more training time, while A2C faces difficulties in continuous spaces. DDPG and TD3, despite t he good results, have high computational consumption, which limits their viability for real-time industrial applications. These findings highlight the importance of considering energy efficiency when choosing RL algorithms for robotic applications. As a future direction, we propose the implementation of these algorithms in a real-world environment, as well as the exploration of hybrid approaches that combine different strategies to improve accuracy and minimize energy consumption.This work was supported by national funds through FCT/MCTES (PIDDAC): CeDRI, UIDB/05757/2020 (DOI: 10.54499/UIDB/05757/2020) and UIDP/05757/2020 (DOI: 10.54499/UIDP/05757/2020); and SusTEC, LA/P/0007/2020 (DOI: 10.54499/LA/P/0007/2020)
Optimization, Learning Algorithms and Applications: 5th International Conference, OL2A 2025 - Part 1
No full textThe volumes CCIS 2617 and 2618 contain the refereed proceedings of the V International Conference on Optimization, Learning Algorithms and Applications (OL2A 2025), a
hybrid event held on April 28–30.
OL2A provided a space for the research community on optimization and learning to get together and share the latest developments, trends and techniques as well as develop new paths and collaborations. OL2A had the participation of more than two hundred participants in an online and face-to-face environment throughout three days, discussing topics associated with areas such as optimization and learning and state-of-the-art applications related to multi-objective optimization, optimization for machine learning, robotics, health informatics, data analysis, optimization and learning under uncertainty and the 4th industrial revolution.
Four special sessions were organized under the topics Artificial Intelligence in Healthcare and Medicine, Optimization in the SDG Context, Optimization in Control Systems Design, and Machine Learning and Artificial Intelligence in Robotics. The event had 38 accepted papers. All papers were carefully reviewed and selected from the 92 received submissions, with each paper receiving three double-blind reviews on average.
All the reviews were carefully carried out by a scientific committee of 115 researchers from twenty-one countries
Schizophrenia diagnosis support with spectral and cepstral features of speech
No full textSchizophrenia is a severe mental illness affecting over 20 million people worldwide, significantly impairing quality of life and daily functioning. Current diagnostic methods rely heavily on subjective assessments and interactions between doctors and patients, leaving room for potential misdiagnoses. Recent advancements in technology have introduced non-invasive, fast, and user-friendly approaches, such as machine learning, to support psychiatric diagnosis. In this study, spectral features extracted from speech samples of individuals with and without schizophrenia were analyzed. Using an ensemble bagged tree model, we achieved an accuracy of 96.3%, a sensitivity of 94.6%, and an F1-score of 95.4%. These results highlight the potential of speech-based machine learning models as effective tools for aiding schizophrenia diagnosis.This work was partially funded by FCT - Fundação para a Ciência e a Tecnologia (FCT) I.P., through national funds, within the scope of the UIDB/00127/2020 project (IEETA/UA, http://www.ieeta.pt/) The authors are grateful for the European Regional Development Fund (ERDF) via the Regional Operational Program North 2020; Foundation for Science and Technology (FCT, Portugal) support from national funds FCT/MCTES (PIDDAC) to CeDRI (UIDB/05757/2020 and UIDP/05757/2020) and SusTEC (LA/P/0007/2021). The authors are grateful for the IN_GENIOS_DUERO_DOURO, Cooperación para la puesta en valor de la cultura industrial del Duero-Douro a través del turismo, 0247_IN_GENIOS_DUERO_DOURO_2_E, financed by FEDER, through the POCTEP. The authors are also grateful to the Foundation for Science and Technology (FCT, Portugal) for financial support through national funds FCT/MCTES (PIDDAC): UNIAG, UIDB/04752/2020 (DOI 10.54499/UIDB/04752/2020) and UIDP/04752/2020 (DOI 10.54499/UIDP/04752/2020). The authors would also like to thank Adriana Santos for her collaboration throughout the work
Food Consumption Practices and Implications for Mycotoxin Exposure Risk: A Cross-sectional Study in Inhambane and Gaza Provinces, Mozambique
Full text linkDiet is considered the main route of exposure to the risk of mycotoxin poisoning. Consequently, given the prevalence of mycotoxins in Mozambique, there is a need to develop studies on the population's consumption habits to implement control and mitigation strategies in a sustainable way. This quantitative and cross-sectional study aimed to report the dietary habits of the population in two provinces of Mozambique and to investigate whether province, place of residence, gender, and age were differentiating factors in dietary habits. Primary data were collected using a food frequency questionnaire adapted from a validated version for Portugal, which was modified to include foods commonly consumed in Mozambique. The data, collected by agricultural technicians between October and November 2022, refer to 300 individuals from 55 households. According to the results, 58.4% of the products consumed by families were of local origin, 40.2% were of national origin, and only 1.4% were imported. Foods such as milk, oil, fresh fish, eggs, rice, bread, sugar, onions, water, mango, and bananas are part of the daily diet of these households. In addition, it was found that the province, environment, and gender of the participants acted as differentiators of eating habits. That is, a greater quantity of foods from all food groups was consumed in the urban environment compared to the rural one. Considering the province, the consumption of fats, bread, and cereals is higher in Gaza, while, in Inhambane, dairy products, fats, fruits, and beverages are consumed in greater quantities. Finally, considering gender, it was found that women consume a greater quantity of dairy products. These results are crucial for the development of sampling strategies that assess fungal and mycotoxin contamination and estimate the risk of exposure to the studied population.This research was funded by the Foundation for Science and Technology (FCT, Portugal) and the Aga Khan Development Network (AKDN), through the project Ref. FCT AGA-KHAN/541590696/2019 “MYCOTOX-PALOP—Multi-actor partnership for the risk assessment of mycotoxins along the food chain in African Portuguese-speaking countries (PALOP)”. We also thank FCT for their financial support through national funds FCT/MCTES (PIDDAC) to CIMO (UIDB/00690/2020 and UIDP/00690/2020), SusTEC (LA/P/0007/2020), and LABBELS (LA/P/0029/2020)
Development of microparticles incorporating chondroitin sulphate for application as Pickering stabilisers for topical applications
No full textMestrado de dupla diplomação com a UTFPR - Universidade Tecnológica Federal do ParanáPickering emulsions, stabilised by solid particles, have been gaining prominence as an alternative to traditional surfactants and stabilisers, giving products of greater stability and unique characteristics. In this study, chitosan-based particles incorporating chondroitin sulphate were developed to produce Pickering emulsions for topical applications, followed by characterising the stabilising particles and the emulsions. For particle dispersion, size and distribution (DLS (dynamic light scattering)), stability (zeta potential (ζ)), and wettability (contact angle (θ)) were characterised. The emulsions’ type, droplet size and distribution (DLS),
stability (zeta potential and creaming index (CI)), colour, rheological properties, antioxidant activity, toxicity, cytoprotection and skin corrosion were assessed. In preliminary tests, relatively large and uniform particles were obtained, identifying the most promising formulations and the most compatible oil. The particles showed good stability, demonstrated by zeta potential values above 30 mV (ζ = 32.30 ± 1.64 mV), indicating good electrostatic repulsion and colloidal stability. In addition, they exhibited relative hydrophobicity, evidenced
by a contact angle greater than 90° (θ = 97.10 ± 2.91°), characterising a moderately hydrophobic surface. Two emulsions were prepared: a base emulsion and another with the addition of vitamin E. Both were identified as O/W (oil/water) emulsions, characterised by the aqueous phase being continuous. These emulsions remained stable for 30 days, without phase separation, showing controlled droplet growth and pseudoplastic behaviour in the rheological tests, properties of a gel-like material in the amplitude and frequency analysis. In addition, the vitamin E emulsion showed high antioxidant activity throughout the study period. The emulsions had a light green colour characteristic of olive oil, with no toxicity or cytoprotective
properties.As emulsões de Pickering, estabilizadas por partículas sólidas, têm vindo a ganhar destaque como alternativa aos tradicionais tensioactivos e estabilizantes, conferindo aos produtos maior estabilidade e caraterísticas únicas. Neste estudo foram desenvolvidas partículas à base de quitosano incorporando sulfato de condroitina para a produção de emulsões Pickering destinadas a aplicações tópicas, seguindo-se a caraterização das partículas estabilizadoras e das emulsões. Para a dispersão das partículas, foram caracterizados o tamanho e respetiva distribuição (DLS (dynamic light scattering)), a estabilidade (potencial zeta (ζ)), a molhabilidade (ângulo de contacto (θ)). No caso das emulsões, foram avaliados o tipo de emulsão, o tamanho e a distribuição de tamanho das gotículas (DLS), a estabilidade (potencial zeta e creaming index (CI)), a cor, as propriedades reológicas, a atividade antioxidante, a toxicidade, a citoprotecção e a corrosão cutânea. Em testes preliminares, foram obtidas partículas relativamente grandes e uniformes, identificando as formulações mais promissoras e o óleo mais compatível. As partículas apresentaram boa estabilidade, demonstrada por valores de potencial zeta acima de 30 mV (ζ = 32,30 ± 1,64 mV), indicando uma boa repulsão eletrostática e estabilidade coloidal. Além disso, exibiram hidrofobicidade relativa, evidenciada por um ângulo de contato superior a 90° (θ = 97,10 ± 2,91°), caracterizando uma superfície moderadamente hidrofóbica. Foram preparadas duas emulsões: uma emulsão base e outra com
a adição de vitamina E. Ambas foram identificadas como emulsões do tipo O/W (óleo/água),
caracterizadas pela fase aquosa como contínua. Estas emulsões mantiveram-se estáveis durante
30 dias, sem separação de fases, apresentando um crescimento controlado das gotículas, e um
comportamento pseudoplástico nos testes reológicos, propriedades de um material tipo gel na
análise de amplitude e frequência. Adicionalmente a emulsão com vitamina E apresentou uma
elevada atividade antioxidante durante todo o período de estudo. As emulsões apresentavam
uma cor verde claro caraterística do azeite, sem toxicidade nem propriedades citoprotectoras
Weather and seasonal effects in behavioural patterns for grazing cattle
No full textWeather conditions influence grazing cattle behaviour, affecting activities like grazing, ruminating, and resting. Understanding these behavioural responses to temperature, humidity, wind, and solar radiation is essential for managing livestock welfare and productivity amid changing climate patterns. The study monitored 20 Minhota breed cattle year-round, from January 1 to December 31, 2024 using GPS-accelerometer collars in Northern Portugal. Behaviour and weather data were analysed via canonical correlation and Kruskal-Wallis tests to assess seasonal impacts on activity patterns and environmental influences. Significant differences (p < 0.001) were found between all the analysed behaviours between the seasons. Grazing and walking peaked in spring and summer, following greater pasture availability. Conversely, resting and rumination increased in autumn and winter, reflecting lower energy demands and greater reliance on roughage. All recorded behaviours were influenced by the temperature indexes and solar radiation (p < 0.001) and the temperature-humidity-sun-wind index correlated most strongly with resting time reductions (-0.62). Surprisingly, moderate heat stress did not suppress grazing and other behaviours like eating or playing. Rainfall negatively affected grazing while increasing rumination and resting (p < 0.001). Canonical correlation revealed strong links between behaviour and environment, and the two canonical covariate pairs (CCP) explain 79.9 % of the variation between the two sets of variables. Seasonal and heat-related factors shaped distinct activity-rest patterns, highlighting cattle's adaptive strategies and the importance of climate-aware livestock management. These findings highlight the need for climate-adaptive livestock management, including strategic feeding, shelter provision, and flexible grazing practices to mitigate climate change impacts.The authors sincerely thank NaturalMinho for supporting the field work and Innogando for providing the GPS collars. Centre for Research and Development in Agrifood Systems and Sustainability is supported by Fundação para a Ciência e a Tecnologia (FCT, Portugal) (UIDB/05937/2020 DOI:10.54499/UIDB/05937/2020 and UIDP/05937/2020 DOI:10.54499/UIDP/05937/2020).info:eu-repo/semantics/publishedVersio
Impact of physical exercise interventions on functional fitness in older adults
No full textAging is associated with a decline in functional fitness, compromising the autonomy of older adults. This systematic review evaluated recent randomized clinical trials (2015–2024) on the effects of physical exercise in individuals aged ≥65 years. Ninety-five studies were included, showing that multicomponent programs, resistance training, and supervised protocols improved muscular strength, mobility, balance, gait speed, and body composition. Interventions lasting more than 12 weeks with high adherence yielded more favorable outcomes. Effects on cognitive function were inconsistent. Structured exercises tailored to the capacities of older adults are effective strategies to promote functional independence and healthy aging, although study heterogeneity limits some conclusions.The author(s) declared that financial support was received for this work and/or its publication. This research was supported by the CXTD2023PY07 project of Mianyang Normal University
Combined nandrolone and resistance training induced cardiac remodelling and oxidative stress despite enhanced cardiomyocyte contractility
No full textNandrolone decanoate (ND) is widely used by individuals engaged in resistance training (RT), yet their combined effects on cardiac function remain unclear. Objective To investigate the effects of RT and ND on cardiac structure and function, cellular contractility, Ca-2 (+) -handling protein expression, and redox balance in rats. Methods Thirty-two male Wistar rats were assigned to four groups: control (C), trained (C-T), ND (N), and trained ND (N-T). Animals received ND or saline for eight weeks, and RT was performed 3 x /week. Cardiac function was assessed by echocardiography, and isolated cardiomyocytes from the left and right ventricles (LV and RV) were evaluated for contractile function. Protein expression of Ca-2 (+) -handling regulators and oxidative stress markers was quantified. Results N-T increased LV and RV diameters by 25% and 33%, septal thickness by 41.7%, and reduced ejection fraction by 12.3% compared to N (p < 0.05). LV cardiomyocytes increased width by 29% and volume by 23% in N-T versus C-T (p <= 0.04). In LV, N-T showed greater contraction amplitude and velocity at 5 Hz (p = 0.04) and increased relaxation velocity at 1, 3, and 5 Hz (p < 0.01), with shorter time to peak contraction and 50% relaxation at 1 Hz versus N (p = 0.04). In RV, N-T increased contraction amplitude and velocity at 1 and 5 Hz, reduced time to peak at 1 and 3 Hz, shortened relaxation time at 1 Hz, and showed higher relaxation velocity than N and C-T (p < 0.02). Molecular analyses revealed reduced RyR2 expression (25% in LV and 9% in RV) and a 41% decrease in LV phospholamban in N-T versus C-T (p < 0.05). N-T also exhibited higher LV malondialdehyde compared to C-T (p = 0.03). Conclusion ND combined with RT induced adverse cardiac remodeling and impaired ventricular function, despite enhanced cardiomyocyte contractility, and intensified molecular and oxidative disturbances, indicating a maladaptive cardiac response. This is the first study to demonstrate regional differences in contractility and redox balance of isolated ventricular cardiomyocytes under ND plus RT
First long-term air quality assessment in Luanda, Angola: Performance evaluation of a low-cost monitoring station against reference equipment
No full textLow-cost air quality monitoring stations (LCMS), which integrate sensors for gases and particulate matter (PM), offer an economical solution for expanding monitoring networks. However, their reliability requires validation, particularly in rapidly urbanising regions with limited infrastructure. This study presents the first long-term, continuous, multi-pollutant air quality assessment in Luanda, Angola - where no monitoring stations currently exist - by evaluating the performance of an LCMS against reference-grade equipment. Daily averages, correlation metrics (R-2, RMSE), and a hybrid Bland-Altman/regression analyses were used to evaluate the agreement. Results indicated strong correlation for CO (R-2 = 0.96; RMSE = 0.24 ppm) and good for NO2 (R-2 = 0.81; RMSE = 6.35 ppb), although limitations near detection limits were noted. Significant challenges were identified in O-3 measurements (R-2 = 0.77, RMSE = 7.13 ppb), primarily due to strong cross-sensitivity to high ambient NO2 levels and potential sensor ageing. For PM10 and PM2.5, although good linear correlations (R-2 similar to 0.82) were observed with reference methods, the LCMS exhibited considerable systematic bias (RMSE over 46 mu g/m(-3)) and consistently underestimate concentrations. The study also registered frequent and severe exceedances of WHO AQG and EU standards for PM10, PM2.5, and NO2, underscoring significant public health risks. Despite limitations, particularly for O-3 measurements and biases in PM data, the LCMS demonstrates potential as a cost-effective tool to complement reference networks, enhance spatial monitoring coverage, identify pollution hotspots, and support air quality management in resource-constrained settings, since continuous calibration and validation procedures are implemented to mitigate measurement uncertainties.This work was supported by the Portuguese Foundation for Science and Technology (FCT) through the PhD fellowships 2023.02059.BD, 2023.04826.BD, SFRH/BD/04992/2021 and SRFH/BD/08461/2020. The research was performed in the frame of the project APAM, financially supported by national funds (OE), through FCT/MCTES (DOI: 10.54499/2022.04240.PTDC). This work was also financed by national funds through FCT under the project/grant UID/50006 + LA/P/0094/2020 (doi.org/10.54499/LA/P/0094/2020), and through FCT/MCTES (PIDDAC): CIMO, UIDB/00690/2020 (DOI: 10.54499/UIDB/00690/2020); and SusTEC, LA/P/0007/2020 (DOI: 10.54499/LA/P/0007/2020). We would also like to sincerely thank d·nota® and bettair® for providing the equipment used in our tests. Additionally, we would like to thank the Faculty of Engineering of the Agostinho Neto University for providing the logistical conditions for carrying out the monitoring campaign in Luanda
Chitosan-xanthan gum-based hydrogels loaded with essential oil distillation by-products of Aloysia citrodora Paláu for antimicrobial systems
No full textHydrogels, 3D hydrophilic networks formed by oppositely charged biopolymers like chitosan and xanthan gum, offer a safe, non-toxic, and biocompatible option for delivery applications. Essential oil (EO) by-products, such as hydrosols and wastewater, are sources of antioxidant and antimicrobial compounds, but their high dilution can limit direct applications. In this context, this work focused on the development of hydrogels via electrostatic complexation incorporating hydrosol and wastewater by-products from the steam distillation of Aloysia citrodora Palau, using a two-stage approach: (a) initial loading during hydrogel formation and (b) subsequent reloading of the hydrogels to further enhance the concentration of bioactive compounds. The effect of pH (4, 7, and 11) on polymer complexation was evaluated, as it influences polymer-polymer and polymer-bioactive compound interactions by modifying the protonation and deprotonation states of their functional groups. This effect was evident in swelling, release kinetics, morphology, and rheological properties. Fourier-transform infrared (FTIR) analysis confirmed the successful formation of the polymer complex. Neutral pH hydrogels showed the highest hydrosol entrapment (70.3%) and were selected as the most promising systems. Biological characterisation showed that the reloading process enhanced bioactivity. Wastewater-load-reload improved antioxidant capacity, driven by the high phenolic content. Moreover, hydrosol-loaded-reload systems exhibited antimicrobial activity, with bactericidal effects against Staphylococcus aureus and Escherichia coli, outperforming both unloaded and loaded systems. These findings highlight the potential of loading and reloading steps to valorise EO by-products, producing sustainable, functional hydrogels with high bioactivity, suitable for food, pharmaceutical, medical, and biotechnological applications.Financial support through national funds FCT/MCTES (PIDDAC): CIMO, UID/00690/2025 (DOI: 10.54499/UID/00690/2025) and UID/PRR/00690/2025 (DOI: 10.54499/UID/PRR/00690/2025); SusTEC, LA/P/0007/2020 (DOI: 10.54499/LA/P/0007/2020); LSRE-LCM, UID/50020/2025 (DOI: 10.54499/UID/50020/2025); and ALiCE, LA/P/0045/2020 (DOI: 10.54499/LA/P/0045/2020). National funding by FCT, Foundation for Science and Technology, through the institutional program contract for scientific employment of A. Santamaria-Echart, and the PhD grant of Heloísa Helena Scorsato de Almeida (DOI: 10.54499/SFRH/BD/148124/2019). Deifil Technology Lda (www.deifil.pt) for supplying the studied plants