Search Results (569)

Unmanned platforms such as unmanned aerial vehicles, unmanned ground vehicles, and autonomous underwater vehicles often face challenges of data, device, and model heterogeneity when performing collaborative data processing tasks. Existing research does not simultaneously address issues from these three aspects. To address this issue, this study designs an unmanned platform cluster architecture inspired by the cloud-edge-end model. This architecture integrates federated learning for privacy protection, leverages the advantages of distributed model training, and utilizes edge computing’s near-source data processing capabilities. Additionally, this paper proposes a federated edge intelligence method (DSIA-FEI), which comprises two key components. Based on traditional federated learning, a data sharing mechanism is introduced, in which data is extracted from edge-side platforms and placed into a data sharing platform to form a public dataset. At the beginning of model training, random sampling is conducted from the public dataset and distributed to each unmanned platform, so as to mitigate the impact of data distribution heterogeneity and class imbalance during collaborative data processing in unmanned platforms. Moreover, an intelligent model aggregation strategy based on similarity measurement and loss gradient is developed. This strategy maps heterogeneous model parameters to a unified space via hierarchical parameter alignment, and evaluates the similarity between local and global models of edge devices in real-time, along with the loss gradient, to select the optimal model for global aggregation, reducing the influence of device and model heterogeneity on cooperative learning of unmanned platform swarms. This study carried out extensive validation on multiple datasets, and the experimental results showed that the accuracy of the DSIA-FEI proposed in this paper reaches 0.91, 0.91, 0.88, and 0.87 on the FEMNIST, FEAIR, EuroSAT, and RSSCN7 datasets, respectively, which is more than 10% higher than the baseline method. In addition, the number of communication rounds is reduced by more than 40%, which is better than the existing mainstream methods, and the effectiveness of the proposed method is verified. Full article

Malware evolution presents growing security threats for resource-constrained Internet of Medical Things (IoMT) devices. Conventional federated learning (FL) often suffers from slow convergence, high communication overhead, and fairness issues in dynamic IoMT environments. In this paper, we propose FedGCL, a secure and efficient FL framework integrating contrastive graph representation learning for enhanced feature discrimination, a Jain-index-based fairness-aware aggregation mechanism, an adaptive synchronization scheduler to optimize communication rounds, and secure aggregation via homomorphic encryption within a Trusted Execution Environment. We evaluate FedGCL on four benchmark malware datasets (Drebin, Malgenome, Kronodroid, and TUANDROMD) using 5 to 15 graph neural network clients over 20 communication rounds. Our experiments demonstrate that FedGCL achieves 96.3% global accuracy within three rounds and converges to 98.9% by round twenty—reducing required training rounds by 45% compared to FedAvg—while incurring only approximately 10% additional computational overhead. By preserving patient data privacy at the edge, FedGCL enhances system resilience without sacrificing model performance. These results indicate FedGCL’s promise as a secure, efficient, and fair federated malware detection solution for IoMT ecosystems. Full article

Federated semi-supervised learning (Fed-SSL) has emerged as a powerful framework that leverages both labeled and unlabeled data distributed across clients. To reduce communication overhead, real-world deployments often adopt partial client participation, where only a subset of clients is selected in each round. However, under non-i.i.d. data distributions, the choice of client sampling strategy becomes critical, as it significantly affects training stability and final model performance. To address this challenge, we propose a novel federated averaging semi-supervised learning algorithm, called FedAvg-SSL, that considers two sampling approaches, uniform random sampling (standard Monte Carlo) and a structured lattice-based sampling, inspired by quasi-Monte Carlo (QMC) techniques, which ensures more balanced client participation through structured deterministic selection. On the client side, each selected participant alternates between updating the global model and refining the pseudo-label model using local data. We provide a rigorous convergence analysis, showing that FedAvg-SSL achieves a sublinear convergence rate with linear speedup. Extensive experiments not only validate our theoretical findings but also demonstrate the advantages of lattice-based sampling in federated learning, offering insights into the interplay among algorithm performance, client participation rates, local update steps, and sampling strategies. Full article

In recent years, the number of fake news stories has significantly increased in the world of media, especially with the widespread use of social media. It has impacted several industries, including tourism. From a tourism point of view, the spread of fake news can contribute to the reduction of the popularity of a destination. It may influence travel decisions by discouraging tourists from visiting certain places and thus damage the reputation of the destination, contributing to economic loss. After a literature review on the communication aspect of fake news and a general introduction of fake news in the tourism and hospitality industry, we conducted a systematic literature review (SLR), a research methodology to collect, identify, and analyse available research studies through a systematic procedure. The current SLR is based on the Scopus, Web of Science, and Google Scholar databases of existing literature on the topic of fake news in the tourism and hospitality industry. The study identifies, lists, and examines existing papers and conference proceedings from a vast array of disciplines, in order to give a well-rounded view on the issue of fake news in the tourism and hospitality industry. After selecting a total of 54 previous studies from more than 20 thousand results for the keywords ‘fake news’ and ‘tourism,’ we have analysed 39 papers in total. The SLR aimed to highlight existing gaps in the literature and areas that may require further exploration in future primary research. We have found that there is relatively limited academic literature available on the subject of fake news affecting tourism destinations, compared to studies focused on hospitality services. Full article

Federated learning (FL) has emerged as a powerful approach for privacy-preserving model training in autonomous vehicle networks, where real-world deployments rely on multiple roadside units (RSUs) serving heterogeneous clients with intermittent connectivity. While most research focuses on single-server or hierarchical cloud-based FL, multi-server FL can alleviate the communication bottlenecks of traditional setups. To this end, we propose an edge-based, multi-server FL (MS-FL) framework that combines performance-driven aggregation at each server—including statistical weighting of peer updates and outlier mitigation—with an application layer handover protocol that preserves model updates when vehicles move between RSU coverage areas. We evaluate MS-FL on both MNIST and GTSRB benchmarks under shard- and Dirichlet-based non-IID splits, comparing it against single-server FL and a two-layer edge-plus-cloud baseline. Over multiple communication rounds, MS-FL with the Statistical Performance-Aware Aggregation method and Dynamic Weighted Averaging Aggregation achieved up to a 20-percentage-point improvement in accuracy and consistent gains in precision, recall, and F1-score (95% confidence), while matching the low latency of edge-only schemes and avoiding the extra model transfer delays of cloud-based aggregation. These results demonstrate that coordinated cooperation among servers based on model quality and seamless handovers can accelerate convergence, mitigate data heterogeneity, and deliver robust, privacy-aware learning in connected vehicle environments. Full article

Skin cancer is one of the most prevalent forms of cancer worldwide, and early and accurate diagnosis critically impacts patient outcomes. Given the sensitive nature of medical data and its fragmented distribution across institutions (data silos), privacy-preserving collaborative learning is essential to enable knowledge-sharing without compromising patient confidentiality. While federated learning (FL) offers a promising solution, existing methods struggle with heterogeneous and missing modalities across institutions, which reduce the diagnostic accuracy. To address these challenges, we propose an effective and flexible Personalized Multimodal Federated Learning framework (PMM-FL), which enables efficient cross-client knowledge transfer while maintaining personalized performance under heterogeneous and incomplete modality conditions. Our study contains three key contributions: (1) A hierarchical aggregation strategy that decouples multi-module aggregation from local deployment via global modular-separated aggregation and local client fine-tuning. Unlike conventional FL (which synchronizes all parameters in each round), our method adopts a frequency-adaptive synchronization mechanism, updating parameters based on their stability and functional roles. (2) A multimodal fusion approach based on multitask learning, integrating learnable modality imputation and attention-based feature fusion to handle missing modalities. (3) A custom dataset combining multi-year International Skin Imaging Collaboration(ISIC) challenge data (2018–2024) to ensure comprehensive coverage of diverse skin cancer types. We evaluate PMM-FL through diverse experiment settings, demonstrating its effectiveness in heterogeneous and incomplete modality federated learning settings, achieving 92.32% diagnostic accuracy with only a 2% drop in accuracy under 30% modality missingness, with a 32.9% communication overhead decline compared with baseline FL methods. Full article

Residential segregation remains a persistent challenge in European urban environments and is an increasing focal point in urban policy debates. This study investigates the changing geographies of ethnic diversity and residential segregation in Riga, the capital city of Latvia. The research addresses the complex dynamics of ethnic residential patterns within the distinctive context of post-socialist urban transformation, examining how historical legacies of ethnic diversity interact with contemporary migration flows to reshape neighborhood ethnic composition. Using geo-referenced data from 2000, 2011, and 2021 census rounds, we examined changes in the spatial distribution of five major ethnic groups. Our analysis employs the Dissimilarity Index to measure ethnic residential segregation and the Location Quotient to identify the residential concentration of ethnic groups across the city. The findings reveal that Riga’s ethnic landscape is undergoing a gradual yet impactful transformation. The spatial distribution of ethnic groups is shifting, with the increasing segregation of certain groups, particularly traditional ethnic minorities, coupled with a growing concentration of Europeans and non-Europeans in the inner city. The findings reveal distinctive patterns of ethnic diversification and demographic change, wherein long-term trends intersect with contemporary migration dynamics to produce unique trajectories of ethnic residential segregation, which differ from those observed in Western European contexts. However, the specific dynamics in Riga, particularly the persistence of traditional ethnic minority communities and the emergence of new ethnic groups, highlight the unique context of post-socialist urban landscapes. Full article

This study reports a successful selective waste collection initiative led by UFSC’s Araranguá campus in a municipality without a recycling system. The initiative, named “Recicla UFSC Ara”, was structured around three main components: (i) the installation of color-coded bins for recyclable waste (including paper, plastic, metals, and polystyrene) and non-recyclable waste in indoor and common areas; (ii) the establishment of a Voluntary Delivery Point (PEV) to gather specific recyclable materials, such as glass, electronics waste, plastic bottles, writing instruments, and bottle caps; and (iii) the execution of periodic educational community-focused campaigns aimed at encouraging participation from both the university and the broader local community. Recyclables were manually sorted and weighed during regular collection rounds, and contamination rates were calculated. Quantitative data collected from 2022 to 2025 were analyzed using descriptive statistics and one-way ANOVA to assess waste generation and contamination trends. Gathered recyclables were directed to appropriate partner institutions, including local “Ecoponto”, non-profit organizations, and corporate recycling programs. The study also conducted a literature review of similar university-led waste management programs to identify standard practices and regional specificities, providing a comparative analysis that highlights both shared elements and distinctive contributions of the UFSC model. Results demonstrate a significant volume of waste diverted from landfills and a gradual improvement in waste disposal practices among the university community. Targeted communication and operational changes mitigated key challenges, improper disposal, and logistical issues. This case underscores the role of universities as agents of environmental education and local sustainable development. Full article

With the hierarchical evolution of distribution network control architectures, distributed state estimation has become a focal point of research. To address communication delays arising from inter-level data exchanges, this paper proposes a multi-level, asynchronous, robust state estimation algorithm that accounts for such delays. First, a multi-level state estimation model is formulated based on the concept of a maximum normal measurement rate, and a hierarchical decoupling modeling approach is developed. Then, an event-driven broadcast transmission strategy is designed to unify boundary information exchanged between levels during iteration. A multi-threaded parallel framework is constructed to decouple receiving, computation, and transmission tasks, thereby enhancing asynchronous scheduling capabilities across threads. Additionally, a round-based synchronization mechanism is proposed to enforce fully synchronized iterations in the initial stages, thereby improving the overall process of asynchronous state estimation. Case study results demonstrate that the proposed algorithm achieves high estimation accuracy and strong robustness, while reducing the average number of iterations by nearly 40% and shortening the runtime by approximately 35% compared to conventional asynchronous methods, exhibiting superior estimation performance and computational efficiency under communication delay conditions. Full article

This study aimed to investigate phytoplankton dynamics and water quality at three drinking water intakes (Duchang, Hukou, and Xingzi) in Poyang Lake through monthly monitoring from May 2023 to April 2024. The results showed that a total of 168 species of phytoplankton were identified in nine phyla, and there were significant spatial and temporal differences in the abundance of phytoplankton at the three waterworks intakes, with a spatial trend of annual mean values of Duchang > Xingzi > Hukou and a seasonal trend of summer and autumn > spring and winter. The dominant species of phytoplankton in the waterworks intakes of the three waterworks also showed obvious spatial and temporal differences. Cyanobacteria (particularly Pseudanabaena sp. and Microcystis sp.) dominated the phytoplankton communities during summer and autumn, demonstrating significant water degradation potential. In contrast, Cyclotella sp. prevailed in winter and spring assemblages. Based on water quality assessments at the three intake sites, the Duchang County intake exhibited year-round mild eutrophication with persistent mild cyanobacterial blooms (June–October), while the other two sites maintained no obvious bloom conditions. Further analyzing the toxic/odor-producing algal strains, the numbers of dominant species of Pseudanabaena sp. and Microcystis sp. in June–October in Duchang County both exceeded 1.0 × 10⁷ cells·L⁻¹. It is necessary to focus on their release of ATX-a (ichthyotoxin-a), 2MIB (2-Methylisoborneol), MCs (microcystins), etc., to ensure the safety of the water supply at the intake. Building upon these findings, we propose a generalized algal monitoring framework, encompassing three operational pillars: (1) key monitoring area identification, (2) high-risk period determination, and (3) harmful algal warnings. Each of these is substantiated by our empirical observations in Poyang Lake. Full article

Water insecurity, intensified by climate change, presents a significant challenge globally, especially in arid and semi-arid regions of Africa. In northern Ghana, where agriculture heavily depends on seasonal rainfall, prolonged dry seasons exacerbate water and food insecurity. Despite efforts to improve water access, there is limited understanding of how climate change preparedness affects water insecurity risk in rural contexts. This study investigates the relationship between climate preparedness and water insecurity in semi-arid northwestern Ghana. Grounded in the Sustainable Livelihoods Framework, data was collected through a cross-sectional survey of 517 smallholder households. Nested ordered logistic regression was used to analyze how preparedness measures and related socio-environmental factors influence severe water insecurity. The findings reveal that higher levels of climate change preparedness significantly reduce water insecurity risk at individual [odds ratio (OR) = 0.35, p < 0.001], household (OR = 0.037, p < 0.001), and community (OR = 0.103, p < 0.01) levels. In contrast, longer round-trip water-fetching times (OR = 1.036, p < 0.001), water-fetching injuries (OR = 1.054, p < 0.01), reliance on water borrowing (OR = 1.310, p < 0.01), untreated water use (OR = 2.919, p < 0.001), and exposure to climatic stressors like droughts (OR = 1.086, p < 0.001) and floods (OR = 1.196, p < 0.01) significantly increase insecurity. Community interventions, such as early warning systems (OR = 0.218, p < 0.001) and access to climate knowledge (OR = 0.228, p < 0.001), and long-term residency further reduce water insecurity risk. These results underscore the importance of integrating climate preparedness into rural water management strategies to enhance resilience in climate-vulnerable regions. Full article

Objectives: Identification of monosodium urate (MSU) and calcium pyrophosphate (CPP) crystals in synovial fluid should ideally be performed within 24 h to ensure optimal diagnostic accuracy for gout and CPP arthritis. However, crystal identification is often delayed in community-based healthcare facilities due to limited access to specialists or necessary equipment. This study aimed to determine whether MSU and CPP crystals remain detectable in synovial fluid after two weeks of storage at 4 °C and −20 °C. Methods: Anonymized synovial fluid samples were obtained from Thammasat University Hospital between February and March 2024. All samples underwent an initial round of crystal identification using compensated polarized light microscopy, conducted by two experienced examiners blinded to the clinical diagnosis. Following the initial analysis, each sample was divided into two equal portions and placed in ethylenediaminetetraacetic acid (EDTA)-coated tubes. One portion was stored at 4 °C, while the other was frozen at −20 °C. After two weeks, all samples underwent a second round of crystal identification. Results: Forty-nine samples were included for the first evaluation; MSU and CPP crystals were identified in 14 and 6 samples, respectively. On the second examination, MSU crystals were detectable in 13/14 (92.8%) samples stored at 4 °C and 12/14 (85.7%) samples stored at −20 °C. However, CPP crystals were detectable in 2/6 (33.3%) samples stored at both temperatures. No new crystal formation in initially negative samples was observed. Conclusion: MSU crystals remain detectable in synovial fluid for up to two weeks when stored in a standard refrigerator or freezer. However, the identification rate of CPP crystals tends to decline over this period. These findings may help inform best practices for handling synovial fluid samples in cases where immediate access to a specialist or necessary equipment is unavailable. Full article

Background/Objectives: A reasonable assessment of the self-perception of aging (SPA) is of great significance to the health outcomes of older adults. This study aimed to develop the Modified Aging Perception Questionnaire (M-APQ) and to verify its psychometric properties. Methods: A multi-method study was conducted. In phase I, a qualitative study was conducted to supplement items to form the draft M-APQ. In phase II, three rounds of cognitive interviews were conducted to revise ambiguous items and form the prefinal M-APQ. In phase III, items were selected using Classical Test Theory (CTT) and Item Response Theory (IRT) to form the final M-APQ. In phase IV, the psychometric properties of the final version of M-APQ were validated. Results: Three items were added in Phase I. Six items were revised in Phase II. Eleven items were removed in phase III, leaving twenty-four items in the final version of M-APQ. In phase IV, the M-APQ showed good construct validity and convergent validity. The known-group validity analysis indicated significant differences in the M-APQ dimension scores on different self-rated health statuses. The Cronbach’s α for M-APQ and each dimension ranged from 0.798 to 0.888, and the intraclass correlation coefficients ranged from 0.704 to 0.883. The IRT analysis showed that item discrimination parameters ranged from 1.746 to 3.630, and difficulty parameters increased sequentially. Conclusions: The 24-item M-APQ includes seven dimensions and is a valid tool for assessing the self-perception of aging (SPA) among community-dwelling older adults. Full article

Wildfires are complex natural disasters that significantly impact ecosystems and human communities. The early detection and prediction of forest fire risk are necessary for effective forest management and resource protection. This paper proposes an innovative early detection system based on a wireless sensor network (WSN) composed of interconnected Arduino nodes arranged in a hybrid circular/star topology. This configuration reduces the number of required nodes by 53–55% compared to conventional Mesh 2D topologies while enhancing data collection efficiency. Each node integrates temperature/humidity sensors and uses ZigBee communication for the real-time monitoring of wildfire risk conditions. This optimized topology ensures 41–81% lower latency and 50–60% fewer hops than conventional Mesh 2D topologies. The system also integrates artificial intelligence (AI) algorithms (multiclass logistic regression) to process sensor data and predict fire risk levels with 99.97% accuracy, enabling proactive wildfire mitigation. Simulations for a 300 m radius area show the non-dense hybrid topology is the most energy-efficient, outperforming dense and Mesh 2D topologies. Additionally, the dense topology achieves the lowest packet loss rate (PLR), reducing losses by up to 80.4% compared to Mesh 2D. Adaptive routing, dynamic round-robin arbitration, vertical tier jumps, and GSM connectivity ensure reliable communication in remote areas, providing a cost-effective solution for wildfire mitigation and broader environmental monitoring. Full article

Background/Objectives: The increased availability of naloxone underscores the urgent need for scalable, effective training interventions. While current training modalities show promise, critical challenges persist, particularly regarding the development of interactive, self-efficacious platforms that mitigate anxiety in real-world overdose response, especially among laypersons. Therefore, this study aimed to develop and evaluate the usability and acceptability of a novel, self-paced mixed reality-based training tool (ReviveXR). Methods: ReviveXR was designed using the Apple Vision Pro spatial computing headset and Unity platform, employing mixed reality technology to facilitate interaction with virtual overdose scenarios while maintaining awareness of the physical environment. The intervention included a simulated tutorial and interactive modules on overdose response, rescue breathing, and chest compressions. Field testing was conducted in two rounds across various settings with a heterogeneous sample (N = 25), including individuals who use drugs, bystanders, first responders, and technology specialists. Data collection involved pre- and post-intervention surveys and qualitative interviews. Results: Participants demonstrated significant improvements in knowledge related to overdose recognition, naloxone administration, rescue breathing, and chest compressions. ReviveXR increased participants’ confidence and intent to help overdose victims while reducing uncertainty during overdose reversal. Participants were predominantly from rural areas and primarily identified as White and male. Qualitative feedback emphasized the platform’s heightened engagement, realism, patient responsiveness, and capacity to enhance knowledge acquisition and behavioral preparedness compared with conventional training approaches. Conclusions: ReviveXR offers a scalable, cost-effective, engaging alternative to traditional naloxone training programs, demonstrating strong feasibility across diverse environments and participants. ReviveXR holds considerable promise for expanding and enhancing community overdose response capacities and training healthcare professionals and first responders. Full article