Search Results (14,119)

Wi-Fi Channel State Information (CSI)-based indoor localization enables high-precision positioning, but its deployment across multiple environments faces two major challenges: privacy concerns from centralizing CSI data, and severe statistical heterogeneity (non-IID) arising from the strong environment-dependency of CSI. This heterogeneity creates a stability–plasticity trade-off in federated learning—maintaining precision in known environments (stability) while adapting to unseen domains (plasticity). To address this trade-off, we propose AdaFed-LDR, which combines server-side Confidence-Weighted Adaptive Aggregation with client-side Layerwise Dynamics Regularization (LDR). The aggregation recalibrates client contributions based on feature covariance changes, while LDR imposes depth-dependent constraints—stronger constraints on shallow layers to preserve environment-agnostic features and weaker constraints on deeper layers to allow environment-specific adaptation. Evaluated across 8 indoor environments using Leave-One-Out Cross-Validation and 5 random seeds, AdaFed-LDR achieved a mean localization error (MLE) of 0.41 cm in known environments, corresponding to an 88.2% reduction compared with FedAvg. In domain generalization to unseen environments, AdaFed-LDR achieved an MLE of $218.2\pm 2.8$ cm, demonstrating an improvement over FedPos ($257.6\pm 14.04$ cm). With one adaptation sample per reference point, MLE improved to 21 cm. Ablation experiments confirmed that combining the two proposed components achieved the highest improvement (83.9%) compared with applying them individually, supporting AdaFed-LDR as a reproducible approach to the stability–plasticity trade-off in federated CSI-based localization. Full article

Sustainable management of peatlands is one of the key global strategies for mitigating climate change. The balance between carbon (C) sequestration and emission in peatlands reflects environmental conditions over time and can provide insight into long-term ecosystem dynamics. However, current methods for estimating greenhouse gas (GHG) fluxes are often labor-intensive, costly, and site-specific. In this study, we propose a simplified and cost-efficient method to estimate long-term carbon balance in peatlands based on the inorganic (mineral) content of drill core samples. The approach uses exponential decay equations to approximate peat accumulation and decomposition processes over time. A conceptual model is applied that accounts for both anaerobic transformation of organic matter of varying molecular complexity and enhanced aerobic decomposition resulting from anthropogenic drainage during the last century. The model was applied to more than 100 drill cores from four peatland systems in Estonia. The resulting trends were compared qualitatively with known climatic fluctuations of the last millennium, including periods associated with the Little Ice Age. The results suggest that, in many cases, carbon losses from decomposition in deeper peat layers may exceed carbon accumulation in upper layers, even in peatlands that appear to be well preserved. The proposed method provides a rapid, low-cost, first-order approximation of peatland carbon dynamics and may serve as a complementary tool for large-scale assessments where detailed process-based models are not feasible. Full article

Fusion-based hyperspectral image super-resolution (HSI-SR) on diffusion models exhibits promising performance in generating high-quality, realistic features. However, existing methods are confronted with two limitations: (1) static conditional guidance is discordant with the dynamic denoising process, and (2) modality conflicts are inadequately addressed by concatenation. To address these challenges, we propose a novel Modulated Diffusion Framework with Spatial–Spectral Disentangled Guidance (SSDG). Specifically, it introduces a Dynamic Modulated Residual Network (DMRN), which leverages a time-aware mechanism to dynamically adjust conditional feature injection, ensuring adaptive guidance throughout all denoising stages. Furthermore, we design a training-free SSDG strategy to explicitly decouple spatial and spectral guidance during sampling, allowing for flexible control over the fusion process to mitigate modality conflicts. Extensive experiments on three public datasets demonstrate that the proposed method achieves state-of-the-art performance, exhibiting superior robustness, particularly in challenging noisy scenarios. Full article

Securing IoT networks presents fundamental challenges rooted in hardware constraints: firmware is often non-upgradeable and every security boundary is fixed at manufacture. Machine learning-based intrusion detection offers a scalable response, yet nearly all published systems assume clean training data and clean inference conditions. Production IoT environments satisfy neither assumption. Sensors degrade, packets drop, and adversaries deliberately corrupt telemetry streams to evade detection. The framework described here is built around that reality. The proposed framework is distinguished from prior work by four design decisions. First, three encoding branches, a residual DNN, a 1D-CNN, and a BiLSTM, are run in parallel and are fused by concatenation, each capturing structural patterns in tabular traffic data that the others miss. Second, a dual-view consistency loss trains the model under simultaneous feature masking and Gaussian noise, penalizing prediction divergence between two independently corrupted views of the same sample. Third, we introduce entropy-weighted attention: rather than fixed learned weights, per-feature importance is adjusted dynamically from information entropy measured across training batches, giving higher-entropy features stronger influence because they carry more discriminative variation. Fourth, branch-dropout regularization randomly silences entire branches during training, forcing each to develop independently useful representations instead of co-adapting. Class imbalance is handled through severity-aware loss weighting which scales contributions by the operational cost of missing each attack category, not purely by inverse frequency. On UNSW-NB15, the full model achieves 99.99% accuracy, 100% precision, 99.97% recall, and a false-negative rate of 2.65 × 10⁻⁴—the lowest across all compared architectures. Full article

Plant mitochondrial genomes are difficult to analyze because of their structural dynamism and frequent annotation errors. To address these challenges, we first constructed a high-confidence mitochondrial reference library for Rhodiola by integrating transcriptomic evidence, public sequence resources, and experimental validation. This curated resource defined 30 mitochondrial protein-coding genes (PCGs), including corrected exon–intron boundaries and validated 5′-terminal variants in ccmC, ccmFn, and nad9. Leveraging this curated dataset, we developed the RhoMitoAnnotator, which integrates three novel algorithms, EBAnno, REAnno, and NCAnno, to accurately annotate trans-splicing, RNA editing, and non-canonical start/stop codons. Using long-read sequencing guided by the RhoMitoAnnotator, we completed the mitogenomes of R. rosea, R. crenulata, and R. sacra, systematically re-annotated seven publicly available mitogenomes, revealing cross-chromosomal gene arrangement, and widespread structural misannotations. To enable scalable analysis with short-read data, we built Polypods, an integrated pipeline that successfully assembled mitochondrial PCGs from 108 samples across 39 Rhodiola species, and identified variant genes, stop codon-lacking regions in nad6, and internal stop codons in rpl16. Phylogenetic analyses based on mitochondrial and chloroplast PCGs showed a lineage pattern consistent with the hypothesis of an evolutionary transition from hermaphroditism to dioecy in Rhodiola, and consistently supported six species as monophyletic lineages. Overall, this study provides a curated mitochondrial gene atlas for Rhodiola and a reference-guided analytical framework for mitochondrial PCG annotation and recovery in this genus, with potential adaptability to other plant lineages after lineage-specific database construction and parameter optimization. Full article

Series arc faults on the DC side of photovoltaic (PV) systems are a critical hazard that can trigger system fires. Conventional contact-based detection methods suffer from cumbersome installation and high retrofit cost, whereas existing non-contact approaches mostly rely on megahertz-level high-frequency sampling and therefore require expensive radio-frequency instrumentation or high-performance computing platforms. As a result, it remains difficult to simultaneously achieve strong interference immunity and real-time performance on low-cost embedded devices with limited resources. To address this engineering paradox between high-frequency sampling and constrained computational capability, this paper proposes a fully embedded, non-contact arc fault detection system based on a 12–80 kHz low-frequency sub-band selection strategy. By exploiting the physical characteristic of broadband energy elevation induced by arc faults, the proposed strategy avoids dependence on high-bandwidth hardware. Guided by this strategy, a Moebius-topology coaxial shielded loop antenna is employed as the near-field sensor, while an ultra-simplified passive analog front end is constructed directly by using the on-chip programmable gain amplifier and analog-to-digital converter of the microcontroller unit, enabling efficient signal acquisition and fast Fourier transform processing within the target sub-band. To cope with complex background noise in the low-frequency range, an environment-adaptive baseline mechanism based on exponential moving average and exponential absolute deviation is developed for dynamic decoupling. In addition, a lightweight INT8-quantized multilayer perceptron is introduced as a nonlinear auxiliary module, thereby forming a robust hybrid decision architecture with complementary rule-based and artificial intelligence components. Experimental results show that, under the tested household, laboratory, and PV-site conditions, the proposed system achieved an overall detection rate of 97%, while the remaining 3% mainly corresponded to failed ignition or non-sustained arc attempts rather than persistent false triggering during normal monitoring. Full article

Chronic kidney disease (CKD) is a progressive and irreversible disorder affecting over 850 million individuals globally and is associated with significant morbidity, mortality, and healthcare burden. Conventional diagnostic approaches rely on intermittent laboratory measurements, including serum creatinine, estimated glomerular filtration rate (eGFR), and urinary albumin, which provide limited temporal resolution and fail to capture dynamic physiological changes. Recent advances in wearable biosensing technologies offer new opportunities for continuous, non-invasive monitoring of biochemical and physiological markers relevant to renal function. This review provides a comprehensive analysis of wearable biosensors for CKD monitoring, focusing on sensing mechanisms (electrochemical, optical, and field-effect transistor), biofluid interfaces (sweat, interstitial fluid, and saliva), and materials engineering strategies enabling flexible, high-performance devices. Emphasis is placed on biofluid transport dynamics, analytical performance across sampling matrices, and system-level integration with wireless communication and digital health platforms. Key challenges limiting clinical translation, including biofouling, enzymatic instability, and variability in biofluid composition, are examined—alongside emerging solutions such as antifouling interfaces, synthetic recognition elements, and multimodal sensing architectures. Finally, regulatory pathways and the role of artificial intelligence in digital nephrology are discussed. This review highlights the potential of wearable biosensors to transform CKD management through continuous monitoring, early detection, and personalized therapeutic intervention. Full article

Ultrafast physical random bit generators (PRBGs) are essential components for modern applications in secure communication, quantum cryptography, encrypted optical fiber sensing and artificial intelligence. While optical chaos-based PRBGs offer high-speed capabilities, conventional systems often rely on discrete components that suffer from system complexity and environmental instability. This paper proposes and experimentally demonstrates a robust, integrated solution using a two-section mutual injection DFB laser. The device was fabricated using the reconstruction equivalent chirp (REC) technique, which provides precise control over grating phase variation while utilizing low-cost, high-volume fabrication methods. The laser sections, each measuring 450 μm in length, were designed with a free-running wavelength difference of 0.3 nm to ensure a flat optical spectrum and enhanced chaotic dynamics. By optimizing the bias currents, we achieved a chaos RF bandwidth of 20.1 GHz. Notably, the resulting chaotic signal lacks time-delayed signatures, which simplifies the randomness extraction process. To generate random bits, the chaotic waveform was sampled by an 8-bit analog-to-digital converter at 100 GSa/s. Following post-processing through delay-subtracting and the extraction of the four least significant bits (4-LSBs), we realized a total physical random bit rate of 400 Gb/s. The randomness of the generated sequence was successfully verified using the NIST SP 800-22 statistical test suite. This approach offers a compact, energy-efficient, and high-performance integrated chaotic source suitable for secure communication and high-performance computation. Full article

Low-cost IoT-based air quality sensors enable dense monitoring networks but suffer from significant measurement noise and instability particularly in dynamic environments. Conventional fixed-window smoothing reduces noise but introduces a trade-off between signal stability and temporal responsiveness, often attenuating short-term pollution events. This paper proposes an adaptive filtering algorithm that dynamically adjusts the averaging window size based on short-term signal variability. The method relies on real-time variance estimation to balance noise suppression and sensitivity to rapid changes without increasing computational complexity. The approach is implemented within an IoT-based monitoring framework and evaluated using parallel measurements with a certified reference device. Comparative analysis against a certified reference device demonstrates strong agreement, with Pearson correlation coefficients reaching r = 0.88 for PM_2.5 and r = 0.86 for PM₁₀, and low error levels (RMSE ≈ 2.1–2.2 µg/m³). The proposed adaptive filtering approach preserves temporal dynamics while improving signal stability and robustness compared to raw and fixed-window filtering. In addition, this method improves event detection stability, achieving low false alarm rates and near real-time response (latency < 1 sampling interval), supporting RPA-based workflow triggering. The results show that the proposed adaptive filtering provides an efficient and lightweight solution for real-time signal processing on resource-constrained devices, making it suitable for large-scale deployment in environmental monitoring systems. Full article

A thorough understanding of the resistivity response characteristics of ethanol-contaminated soil is of great significance for the development of non-destructive geophysical detection techniques and for supporting contaminated site investigation and assessment. This experimental study aims to systematically investigate the resistivity behavior of ethanol-contaminated sandy soils, with a focus on the coupled mechanisms of multiple factors, including water content, ethanol concentration, particle size distribution, and contamination time. It is hypothesized that water content serves as the dominant factor controlling resistivity, whereas ethanol concentration and contamination time regulate resistivity by altering the physicochemical properties of the pore fluid. Under laboratory conditions, silt, fine sand, and medium sand were selected as the test materials. Resistivity was systematically measured using a Miller Soil Box with increasing water content, Wenner array configuration across varying water contents (3–24%), ethanol concentrations (40–98%), and contamination durations (0–144 h). The experimental results indicate the following: (1) Regardless of the presence of ethanol contamination, the resistivity of sandy soil decreases with increasing water content following a power-law relationship. The decrease is most pronounced at low water contents (3–9%), and gradually stabilizes at higher water contents. The results show that, at a constant water content, resistivity systematically and consistently follows the order: silt > medium sand > fine sand. (2) The influence of ethanol concentration on resistivity is constrained by water content levels, and the overall increase in resistivity is primarily attributed to ion dilution and the obstruction of conductive pathways. (3) Over time, resistivity exhibits a two-stage increasing trend, associated with ethanol volatilization and water loss. Resistivity changes in fine sand samples contaminated with ethanol at concentrations ranging from 75% to 95% follow a two-stage pattern. The initial phase of growth is characterized by a gradual increase over a period of 0–48 h, followed by a more rapid increase during the subsequent phase, which extends from 48 to 144 h. The results show that higher initial ethanol concentrations enhance the sensitivity of resistivity to temporal changes. Comprehensive analysis indicates that the resistivity variation mechanism under multi-factor coupling conditions can be summarized as follows: the water content is the dominant factor in the regulation of the conductive pathways; the particle size distribution determines pore structure and the characteristics of the particle interface; ethanol concentration and contamination time dynamically alter pore fluid properties, collectively regulating the resistivity response. Although the experiments were conducted under controlled laboratory conditions and the results have certain limitations, they provide a preliminary reference for interpreting resistivity responses in relatively homogeneous sandy contaminated sites and offer theoretical support for the application of resistivity methods in contamination identification and dynamic monitoring. Full article

Forward-looking sonar (FLS) is an important sensing modality for autonomous underwater vehicles and other marine robotic systems operating in turbid, low-visibility, and acoustically cluttered environments. Reliable target detection in FLS imagery remains challenging because target echoes are often weak, compact targets can be obscured by background clutter, and embedded processors impose strict limits on model size, latency, and computation. To address these issues, this study presents a lightweight FLS sensing framework for embedded target detection in resource-constrained underwater systems. The framework combines a compact detection architecture, difficulty-aware supervision, and teacher–student knowledge transfer. Specifically, FPN-Mix is developed as a lightweight backbone with a Conv-Mix module to improve contextual aggregation under limited computational budgets. A target-aware dynamic weighting loss is introduced to increase the supervision weight of difficult acoustic samples associated with weak echoes, ambiguous boundaries, and clutter interference. A multi-level knowledge distillation strategy is then adopted to transfer feature-level and prediction-level knowledge from an enhanced teacher model to the compact student detector. Experiments on the public UATD benchmark and the independently collected Zhanjiang Bay No.1 field dataset show that the proposed method achieves a favorable balance between detection accuracy and efficiency and remains competitive in a real marine aquaculture environment. The proposed model contains only 2.83 M parameters and requires 6.68 GFLOPs. After ONNX export and TensorRT FP16 acceleration, the model reaches 72.23 frames per second (FPS) on an NVIDIA Jetson Orin NX platform, supporting its practical use in embedded FLS sensing systems. Full article

Accurate paper fiber identification is essential for cultural heritage conservation. Traditional staining methods are destructive, while macroscopic AI models often lack physicochemical interpretability. This study explores the feasibility of a non-destructive analytical approach using micro-hyperspectral imaging (Micro-HSI) to overcome both limitations. Three traditional Japanese bast fibers, Kozo, Mitsumata, and Gampi, were analyzed as standard reference samples. Relative reflectance spectra were extracted from microscopic fiber regions using Micro-HSI. Dynamic normalization and Savitzky–Golay first-derivative filtering were applied to suppress scattering effects and baseline drift. Principal component analysis (PCA) and linear discriminant analysis (LDA) were applied in parallel for dimensionality reduction and supervised classification, respectively. The results indicated that unsupervised PCA exhibited substantial inter-class overlap because of the shared cellulose matrix among the fiber types. In contrast, supervised LDA amplified subtle chemical differences and achieved clear separation among the three fibers. Feature-loading analysis indicated that the classification was mainly associated with visible range reflectance characteristics, lignin π→π* absorption bands in the 400–450 nm region, and near-infrared O−H and C−H overtone vibrations near 835 nm. Leave-One-Specimen-Out Cross-Validation yielded an overall accuracy of 77.8%, with error-free classification of Kozo (F1 = 1.00) and misclassification limited to the chemically similar Gampi and Mitsumata pair. This proof-of-concept study demonstrates that combining Micro-HSI with chemometric analysis enables non-destructive fiber discrimination while retaining physicochemically interpretable spectral features. The findings also establish a microscopic spectral reference framework for future non-destructive analysis of historical paper materials. Full article

Recent evidence highlights the therapeutic potential of Astragali radix-Huaier fermentation products for hyperuricemia treatment, although the dynamics of the fermentation process remain poorly understood. This study employed high-resolution mass spectrometry and untargeted metabolomics for real-time monitoring of chemical components and hypouricemic activity throughout fermentation. The results revealed significant alterations in the chemical composition, with distinct sample separations observed on days 7, 14, 21, and 28. A total of 33 differential components were identified, including 20 flavonoids and 13 saponins, eight of which showed notable changes. Polysaccharides and saponins were found to correlate positively with the uric acid-lowering effect. On day 21, the levels of total polysaccharides and cycloastragenol-6-glucoside, a saponin derivative, peaked, coinciding with the highest hypouricemic activity of the Astragalus fungal fermentation products. This study provides the first evidence of dynamic changes in the chemical profile and pharmacological activity of Astragali radix-Huaier during fermentation, paving the way for optimizing fermentation processes and developing medicinal and dietary products based on Astragali radix. Full article

As a key indicator of low-carbon economic transformation, the influencing factors of carbon productivity (CP) have attracted considerable academic attention. However, the study of the role of artificial intelligence (AI) technology innovation is comparatively confined. Using China’s prefecture-level-and-above cities as the sample, this study measures regional AI technology innovation based on AI patent stocks and empirically examines its impact on carbon productivity. The principal findings of this paper are as follows: (1) AI technology innovation boosts urban carbon productivity through three channels: enhancing green innovation, reducing transaction costs, and increasing AI attention. (2) The regional heterogeneity analysis shows that this positive impact of AI technology innovation on carbon productivity exerts a stronger facilitating effect on eastern regions, resource-dependent cities, and central cities. The heterogeneity analysis at the technological level further provides evidence of the effect of AI technology innovation on carbon productivity varying along different tiers of technological development, innovation mode, and innovation role. (3) The analysis identifies the energy structure as a pivotal threshold variable governing the efficacy of AI innovation in bolstering carbon productivity. Notably, crossing the threshold of clean energy penetration triggers an escalating positive feedback loop between AI innovation and carbon productivity. (4) Estimation of temporal effect dynamics via non-parametric panel model shows that the impact of AI technology innovation on CP exhibits phased characteristics. The coefficient became significantly positive in 2010 and peaked in 2015, after which its effect gradually weakened. This study provides comprehensive empirical evidence for understanding the relationship between AI technology innovation and CP and provides policy references for the use of AI technology to promote the coordinated achievement of economic growth and carbon reduction. Full article

Ants are widely used as bioindicators because of their sensitivity to environmental change and their functional roles in ecosystems. This study presents the first comparative analysis of ant communities in two habitats, an agricultural system and a semi-natural forest, within the Natural Park of Montesinho (northeastern Portugal). From May to October 2022, four plots were sampled per habitat: (i) semi-natural oak forest and (ii) chestnut orchard under human management, using five pitfall traps in each plot. A total of 1969 ants were captured, representing 32 species and 15 genera. Traditional chestnut orchards supported more exclusive species and greater functional diversity, dominated by generalist and thermophilic taxa. In contrast, oak forests hosted more specialist and cold-adapted species, which may reflect a higher structural stability. Seasonal variation was more pronounced in chestnut orchards, consistent with disturbance-driven dynamics. The functional composition also differed: chestnut orchards favoured granivores and scavengers, while oak forests supported predators and mutualists. These findings highlight the value of ant communities as sensitive indicators of land use and ecosystem condition in Mediterranean mountain systems. Full article