Search Results (2,358)

Background: Acute sore throat (pharyngitis) accounts for 10–30% of ambulatory visits annually. Bacterial etiology, primarily Group A Streptococcus, is confirmed in only 20–34% of cases, and inappropriate antibiotic administration remains frequent, contributing to antimicrobial resistance. Community pharmacists play a key role in symptomatic management and stewardship, but real-world practices remain underexplored. Methods: A 16-item online questionnaire surveyed 629 Italian community pharmacists via professional networks, covering demographics, diagnostic tools, symptom assessment, treatment preferences, and follow-up. Responses were analyzed using descriptive statistics and interpreted within a multidisciplinary expert framework. Results: Italian community pharmacists (n = 629), predominantly mid-career and experienced individuals, prioritize targeted symptom questioning (flu-like signs (59.96%), fever (48.73%), pain intensity (49.64%)) and medical referral over formal diagnostic tools like the Centor criteria (unused in 53.66% of cases). A total of 46.74% report performing a Strep A test in fewer than 20% of patients. Topical sprays (83.33%) and lozenges (44.69%), especially containing flurbiprofen (94.19%), dominate recommendations, followed by systemic analgesics (45.79%). For mild cases, anti-inflammatory/analgesic sprays/lozenges (35.42%) and antiseptic lozenges (35.61%) dominate, while in severe cases, drugs with anti-inflammatory or analgesic effects are often preferred (72.47%). Most pharmacists request a follow-up (73.22%). Conclusions: Italian community pharmacists report a predominantly topical-first approach focused on symptomatic treatment and referral when appropriate. Diagnostic gaps and inconsistent follow-up represent actionable targets. The findings inform training outputs prioritizing simplified triage, expanded symptom checklists, spray/lozenge optimization, and 3–5-day call-backs. This work aims to promote pharmacists’ antimicrobial stewardship to reduce antibiotic use while enhancing patient-centred sore throat care in Italy’s pharmacy network. Full article

Accurate network traffic prediction is important for the autonomy, resilience and resource orchestration of 6G and AI-native communication infrastructures, while also supporting green networking and digital twin network applications. However, existing data-driven prediction models face several limitations: over-reliance on massive labeled data, physically implausible predictions, black-box non-interpretability and over-parameterization that impairs edge deployment. To address these issues, this paper proposes a Physics-Informed Network Traffic Prediction (PINTP) framework for graph topology network traffic prediction, which formalizes network traffic evolution as Graph-based Advection–Diffusion–Reaction (ADR) equations and embeds physical regularization into the neural architecture. The framework adopts a hybrid differentiation paradigm unifying automatic differentiation for temporal dynamics and spectral graph theory-derived operators for discrete spatial topologies, and designs a physics-constrained composite loss function with data-driven collocation to balance data fidelity and physical consistency. Experiments are conducted in two complementary settings: a 100-node synthetic random-graph benchmark that evaluates the full graph-topological formulation, and a topology-unavailable real-world telemetry proxy based on Alibaba Cluster Trace v2018 for evaluating sparse-label physics-informed temporal regularization. Comparative analysis with mainstream baselines, including Multilayer Perceptron (MLP), Spatio-Temporal Graph Convolutional Network (STGCN), Graph WaveNet, Transformer, Temporal Convolutional Network (TCN), and XGBoost, shows that the proposed PINTP/PINN implementation achieves a test $R^2$ of 0.898 and MSE of 0.000723 on the 100-node synthetic graph benchmark, close to the strongest Transformer result ($R^2=0.900$, MSE = 0.000710), while using substantially fewer trainable parameters. PINTP/PINN also outperforms Graph WaveNet, STGCN and TCN in this setting, indicating that physics-informed regularization can remain competitive as graph size increases. On the Alibaba proxy task, PINTP/PINN achieves the strongest result among the evaluated models with a test $R^2$ of 0.963. In an independent Alibaba ablation protocol, physical regularization (e.g., $\lambda =10.0$) reduces the mean squared error by 89.15% compared with pure data-driven models and helps mitigate overfitting. This work presents a systematic PINTP framework for graph topology network traffic prediction, achieving competitive prediction accuracy with high parameter efficiency and a degree of physical interpretability. It helps address several limitations of traditional data-driven models, indicates potential for future deployment-oriented studies on real-time network management and resource-constrained edge analytics, and provides an interpretable modeling route for physics-informed network analytics in next-generation communication systems. Full article

This paper presents the design, construction, and experimental evaluation of a low-cost, portable solar tracking system for monitoring ultraviolet solar radiation in real time. It integrates a Raspberry Pi Zero 2W as the embedded control unit, an AS7331 spectral sensor to measure UVA, UVB, and UVC irradiance, two 270° servomotors to position the system toward the sun, an NEO-6M GPS module to geolocate the system, and a DS3231 real-time clock to synchronize the time. To enable autonomous outdoor operation, a multistage power supply architecture based on a solar panel, a rechargeable battery, and LM2596 and MP1584EN DC-DC regulators was implemented. The tracking algorithm uses astronomical equations to estimate the solar azimuth and elevation and updates the sensor orientation during daylight hours. This allows the UV sensor to remain approximately normal to the incoming solar radiation. Experimental tests were conducted in Arequipa, Peru. The recorded data included UVA, UVB, and UVC irradiance; sensor temperature; geographic coordinates; time; and solar angles. The measured UV profiles exhibited the anticipated diurnal behavior: maximum values around solar noon, higher UVA levels than UVB levels, and minimal UVC levels due to atmospheric absorption. We compared the radiometric response with reference information from EarthKit, PVGIS 5.3, SAMPA, and a Davis Vantage Pro 2 weather station. We evaluated the solar positioning performance against Stellarium, NOAA, and the NREL Solar Position Algorithm. Across the complete five-day validation at three daily evaluation times, the maximum percentage errors were 0.0584% for azimuth and 0.5059% for elevation relative to the NREL SPA, NOAA, and Stellarium reference calculations. The results demonstrate that the proposed system constitutes an embedded, portable, autonomous, and low-cost platform for in situ monitoring of solar ultraviolet radiation. Due to its modular architecture, georeferencing capability, time synchronization, and independent power supply, the prototype can be used as a mobile measurement unit or as part of a distributed network of UV stations at various locations in Arequipa. In this regard, the system enables multipoint measurement campaigns, complements fixed weather stations, validates solar models, and generates local experimental data for the spatial and temporal assessment of the solar UV resource under real-world field conditions. Full article

Public building retrofit is an important pathway for reducing operational carbon emissions, but evidence-based technology prioritisation remains limited under real-world multi-technology retrofit conditions. This study develops an interpretable data-driven framework to identify priority technologies and technology co-adoption patterns for public buildings in China’s Hot Summer and Cold Winter (HSCW) region. Based on 218 completed retrofit projects, the carbon reduction rate (CRR) was used as the performance indicator, and 15 retrofit technologies were analysed using FDR-adjusted Mann–Whitney U tests, repeated-validation XGBoost–SHAP analysis, and Apriori association-rule mining. The projects showed substantial variation in CRR, with a mean of 24.98% and a median of 21.70%. Across five repeated 5-fold cross-validations, the predictive XGBoost model achieved a mean R² of 0.417 and a mean RMSE of 0.127. Roof insulation, external wall insulation, and ventilation system retrofit showed the strongest combined evidence and were classified as core technologies. Apriori analysis further revealed three empirical co-adoption patterns: integrated passive-envelope retrofit, solar-control and renewable-energy integration, and operational-management improvement. The findings suggest that retrofit planning in the HSCW region should prioritise envelope insulation and ventilation performance, while selecting shading, system, renewable-energy, and operational-control measures according to project-specific conditions. Full article

Meat quality and safety are critical aspects of global food security. However, traditional evaluation techniques, including sensory analysis and chemical and instrumental tests, are constrained by subjectivity, high time consumption, their destructive character, and susceptibility to bias. With rapid advances in sensor technologies and computational methods, there is a growing demand for the nondestructive, accurate, and fast measurement of meat quality and safety attributes. In recent years, artificial intelligence (AI) integrated with nondestructive sensing has emerged as a transformative paradigm, offering unparalleled capabilities for extracting quality information from complex datasets generated by various nondestructive sensing technologies. This review provides a comprehensive analysis of AI-driven nondestructive technologies for meat quality and safety assessment, focusing on the integration of machine learning and deep learning with various sensing techniques. Additionally, the review evaluates state-of-the-art algorithms and their performances and identifies deployment barriers, particularly calibration transfer, environmental sensitivity, reproducibility issues, sensor fouling, and generalization challenges across batches and processing plants. Furthermore, economic and regulatory constraints, including high sensor costs, small and medium enterprise (SME) adoption challenges, and alignment with HACCP/ISO frameworks that further limit commercial scalability, are discussed. Unlike previous reviews that primarily focus on individual sensing techniques, this review emphasizes the practical challenges associated with industrial implementation and the development of scalable solutions for real-world deployment. Finally, strategic research priorities and recommendations are highlighted to accelerate the industrial adoption of intelligent meat quality monitoring systems across the global meat industry. Full article

Machine learning techniques for network-based intrusion detection systems (NIDS) have advanced considerably over the past decade. Still, improvements are inhibited by handcrafted feature pipelines, isolated public benchmark data, and evaluation procedures that do not reflect real-life deployment. AutoML, a branch of ML automating model selection, automated architecture search, and the creation of model pipelines, may help overcome these shortcomings. While numerous NIDS applications employing automated ML techniques have been proposed, and recent surveys have mapped the AutoML framework landscape for network intrusion detection, no existing review critically audits the evaluation practice of this literature: the quality of its benchmark datasets, the reproducibility of its reported results, and the realism of its deployment assumptions. This paper critically reviews 26 research works published between January 2023 and June 2026, collected via a two-phase structured search: a documented keyword search across five databases (Scopus, IEEE Xplore, Web of Science, ACM Digital Library, and Google Scholar), followed by full-text eligibility screening, citation chaining, and expert evaluation. Findings drawn from this collection capture trends observed among the selected studies, rather than reflecting the broader state of the field. Analysis of the corpus reveals that 88% of dataset-verified studies evaluate exclusively or partly on the legacy benchmark family (KDD-derived, CICIDS, UNSW-NB15, CIDDS), 21% evaluate on a single dataset only, and among attribute-verified studies only 32% release source code, 40% report statistical significance testing, and 36% include variance analysis, findings that collectively motivate the four contributions of this study. First, a recommended evaluation framework is proposed, addressing baseline parity, transparent search-space and budget reporting, nested cross-validation for selection-bias control, and stability reporting across multiple random seeds. Second, a dataset quality scoring framework is introduced, assessing five dimensions: overlap rate, duplication rate, label correctness, attack-type representativeness, and coverage of benign, IoT, and IIoT traffic. Third, a cross-domain justification is provided for neural architecture search (NAS) and meta-learning in NIDS, grounded in advances in federated NAS, out-of-distribution robustness, edge-constrained search cost reduction, and few-shot adaptation. Fourth, a structured research roadmap is outlined, targeting real-world validation, standardized benchmarks, curated datasets, resource-aware AutoML, and privacy-preserving federated NAS. In contrast to prior surveys of AutoML for network intrusion detection, which map frameworks and computational paradigms, this review contributes a formalized evaluation checklist, an explicit and partially empirically validated dataset quality scoring scheme, and evidence-based methodological guidance grounded in a transparent, fully enumerated study corpus. Full article

Hardware-in-the-loop (HIL) technology combines physical hardware with virtual models to achieve efficient closed-loop simulation of automotive systems, demonstrating significant advantages in the development of automotive emissions and energy consumption. This paper reviews the current applications of three HIL technologies, including powertrain-in-the-Loop (PIL), engine-in-the-Loop (EIL), and Virtual Test Bed (VTB). It also explores their role in addressing the increasingly stringent regulations on emissions and energy consumption. Research indicates that PIL technology can significantly improve the efficiency with which hybrid powertrain control strategies are verified by integrating real powertrains with virtual environments. EIL technology enables the high-precision simulation of real-world emissions at low hardware cost. VTB technology, meanwhile, significantly reduces the calibration period by leveraging high-precision models and intelligent algorithms. These three technologies form a comprehensive development and verification chain, covering everything from components to vehicles. However, HIL technology still faces challenges relating to model accuracy, system complexity and cost. Therefore, the most suitable technology should be selected based on development objectives, timeframe, and budget. Full article

Small unmanned aerial vehicles (UAVs) often occupy only a limited number of pixels in an image and can be easily confused with surrounding objects in cluttered scenes, which makes reliable detection difficult. To address this challenge, we develop PCCS-YOLOv8, an enhanced object detector tailored to small UAV targets. A P2 prediction branch is added to retain fine spatial information associated with tiny objects. The cross-stage partial pyramid convolution (CSPPC) module is introduced to offset the additional computational burden caused by the detection branch with high resolution, while the spatial pyramid pooling with efficient layer aggregation network (SPPELAN) combines multiscale pooling with efficient feature aggregation. The convolutional block attention module (CBAM) is further integrated to emphasize features related to targets and reduce interference from complex backgrounds. Experiments were conducted on a UAV dataset containing 7785 images collected from TIB-UAV, Anti-UAV, and self-collected sources. PCCS-YOLOv8 achieved an mAP@0.5 of 94.0% and an mAP@0.5:0.95 of 50.6%, outperforming the YOLOv8 baseline by 2.9 and 2.2 percentage points, respectively. After training, the model was exported, converted to RKNN format, and then deployed on an Orange Pi 5 Pro development board. In real-world detection tests, the embedded system achieved an average frame rate of 27.7 FPS and an average runtime of 46.3 ms per frame. These results demonstrate the potential of the proposed method for real-time UAV detection on edge devices. Full article

Ultra-wideband (UWB) and inertial measurement unit (IMU) fusion is an effective scheme for train and rail-guided target localization because UWB provides absolute ranging results and the IMU provides high-rate motion prediction. In practical rail transportation systems, UWB anchors are usually deployed along the track with large longitudinal intervals and limited lateral separation to reduce installation and maintenance costs. This anchor deployment leads to a large condition number of the observation matrix, so slight ranging errors caused by non-line-of-sight (NLoS) propagation may be amplified into large localization errors. To mitigate LoS/NLoS interference, this article proposes a propagation reliability estimation method for track-constrained UWB/IMU fusion localization. First, rail transportation localization along a narrow path is formulated as a one-dimensional track-constrained problem, and each UWB ranging result is converted into a candidate longitudinal coordinate on the known track centerline. Second, the reliability of each anchor–target propagation is estimated in a sliding window by comparing the motion increments solved by UWB observations with the motion prediction by the IMU. Third, the estimated reliability is incorporated into a reliability-weighted track-domain update before a closed-loop position–velocity Kalman correction. The simulation results show that, under the mixed LoS/NLoS scenario, the proposed method achieves an NLoS-interval RMSE of 0.0090 m. Compared with Track-EKF, Track-Gauss-AUKF, Track-Adaptive KF, and Track-SW-FGO, the proposed method reduces the NLoS-interval RMSE by 92.9%, 62.1%, 92.8%, and 92.6%. A supplementary real-data stress test on the public STAR-loc dataset demonstrates an average longitudinal RMSE of 0.0555 m under a strict online calibrated-range protocol, supporting the algorithm’s practical applicability against real-world lateral sway and asynchronous sensor noise. Full article

Accurate object localization is essential for enabling autonomous operation of indoor robotic systems. As a low-cost, compact, and flexibly deployable solution, monocular visual object localization (MVOL) is highly applicable to lightweight embedded robotic platforms. However, conventional data-driven MVOL methods suffer from inherent limitations of 2D-to-3D ill-posed mapping, which is caused by the incapability of constraining the spatial physical logic of real scenes, resulting in severe depth ambiguity, inaccurate scale estimation, and physically unreasonable predictions. To address these issues, this paper proposes a novel physics-guided monocular visual localization framework termed PC-IMVL for indoor scenarios. The PC-IMVL integrates deep visual perception with embedded physical modeling, which explicitly introduces spatial physical constraints into the network optimization process and builds a physical consistency-aware loss function to regularize 3D position and pose estimation. Combined with a lightweight tailored architecture, the framework enables efficient and reliable embedded deployment. Offline experiments and real-world online tests validate the effectiveness of the proposed method. PC-IMVL yields average absolute errors (AE) of 0.095–0.333 m, reducing the localization error of early fusion methods by more than 50%. Within a working distance of 3–4 m, it achieves a relative error (RE) of 2.4% and a horizontal viewing angle error (VAE) below 2°, outperforming existing state-of-the-art MVOL methods. The effectiveness of the physical guidance mechanism is verified. This work provides a practical high-precision localization solution for embedded indoor robotic systems. Full article

Artificial intelligence (AI) is transforming oncologic imaging by enabling automated, accurate, and scalable diagnostic decision support. However, successful clinical translation requires not only strong predictive performance but also interpretability, transparency, and clinician confidence. We present an attention-guided deep learning framework for automated slice-level classification of parathyroid adenoma (PTA) from arterial-phase computed tomography (CT) images. The framework incorporates lightweight hierarchical attention mechanisms within residual neural networks to enhance feature representation and contextual understanding while maintaining computational efficiency for real-world deployment. Three novel architectures were developed: the Residual Block Light Attention Network (Res-BLANet), Residual Stage Light Attention Network (Res-SLANet), and Residual Layer Light Attention Network (Res-LLANet). Models were trained and evaluated on a rigorously curated, expert-annotated dataset of 63 patients, including 39 pathologically confirmed PTA cases, with 350–450 slices per patient. Training employed Hounsfield unit normalization, extensive data augmentation, and 10-fold cross-validation to ensure robust performance assessment. A key innovation is the integration of hierarchical attention modules that generate attention maps at multiple network levels, enabling qualitative visualization of diagnostically relevant regions and providing exploratory insight into the model’s decision-making process. Experimental evaluation showed strong diagnostic performance. Res-SLANet achieved the highest overall accuracy (87.37%), precision (92.56%), and F1-score (86.69%), while Res-LLANet attained the highest sensitivity (96.0%) on independent testing. These results should be interpreted as preliminary, given the limited size of the independent patient-level test cohort (n = 3). Res-BLANet delivered substantially faster inference with minimal computational overhead. These findings demonstrate that hierarchical attention-guided deep learning can achieve accurate and computationally efficient PTA detection from CT imaging while providing qualitative visual insights into the model’s decision-making process. The proposed framework represents a promising approach toward more interpretable AI-assisted diagnostic systems for oncologic imaging. Full article

Background: Cryptococcal infections remain life-threatening fungal infections affecting both people living with HIV (PLWH) and non-HIV populations. Data from real-world cohorts in middle-income settings, particularly those including both groups, remain limited. This study aimed to evaluate the clinical characteristics and identify predictors of one-year mortality in patients with cryptococcal infections. Methods: In this retrospective multicenter cohort study, adult patients diagnosed with cryptococcal infection between 2013 and 2025 were included and followed for one year after diagnosis or until death. Demographic, clinical, laboratory, and treatment data were collected. Survival was analyzed using Kaplan–Meier methods and compared with the log-rank test. Cox proportional hazards regression was used to identify factors independently associated with mortality. Subgroup analyses were performed according to HIV status. Results: A total of 92 patients with an average age of 46 ± 16 years were included; 56.5% were PLWH, and 87% were immunocompromised. Central nervous system involvement was the most common presentation (64.1%). Overall one-year mortality was 46.7%. In the multivariable Cox regression analysis, altered consciousness (HR 3.110, 95% CI 1.624–5.956, p < 0.001) and disseminated infection (HR 2.155, 95% CI 1.156–4.016, p = 0.016) were independently associated with increased mortality, whereas higher serum albumin levels were independently associated with improved survival (HR 0.532, 95% CI 0.328–0.861, p = 0.010). Age remained independently associated with mortality (HR 1.025 per year, 95% CI 1.006–1.045, p = 0.010). Kaplan–Meier analysis showed significantly lower one-year survival among patients with altered consciousness and disseminated infection. Although CNS involvement was more frequent among PLWH (78.8% vs. 45%, p < 0.001), mortality rates were comparable between PLWH and non-HIV patients. Conclusions: Cryptococcal infections are associated with high mortality regardless of HIV status. Altered consciousness and disseminated infection were independently associated with one-year mortality, whereas higher albumin levels were associated with improved survival. Simple clinical parameters may contribute to early clinical risk assessment, particularly in resource-limited settings. Full article

Background: Diabetic kidney disease (DKD) is the leading cause of chronic kidney disease and end-stage renal disease worldwide. Despite robust evidence-based guidelines, real-world adherence to recommended preventive care remains inadequate. Most prior evaluations have focused on individual care components rather than comprehensive composite care delivery. Methods: A retrospective cross-sectional study of 1010 adult patients with diabetes attending outpatient clinics at King Abdulaziz Medical City, Ministry of National Guard Health Affairs (MNGHA), Jeddah, Saudi Arabia. Full guideline-concordant care (GCC) was defined as composite completion of HbA1c monitoring, albuminuria screening using the urine albumin-to-creatinine ratio (ACR), and renin–angiotensin–aldosterone system inhibitor (RAASi) use among patients with detected albuminuria. Multivariable logistic regression (n = 1007) identified independent predictors of full GCC. Results: Only 22.7% (95% CI: 20.1–25.3%) of patients achieved full GCC. Partial care was the most common pattern (46.2%), followed by no care (20.7%) and near-complete care (10.4%). HbA1c testing was performed in 55.0% and albuminuria screening in 49.3% of patients. Among those with detected albuminuria, RAASi therapy was prescribed in 78.9%. Primary care management (OR 4.79, 95% CI 3.45–6.66; p < 0.001) and hypertension (OR 2.00, 95% CI 1.33–3.00; p = 0.001) were independently associated with full GCC. Age, gender, kidney function, and cardiac disease were not significant predictors. Conclusions: Substantial gaps exist in guideline-recommended DKD care, driven primarily by screening deficiencies rather than treatment failures. Care delivery is influenced more by healthcare setting than by patient characteristics, highlighting the need for system-level interventions targeting detection infrastructure across all care settings. Full article

To enhance unmanned ground vehicle (UGV) intelligence in smart cities, disaster rescue, and infrastructure inspection, this paper investigates the collaborative optimization of multimodal fusion end-to-end architectures. Through dynamic alignment of heterogeneous features, a multi-head distillation attention mechanism, and parallel decision-planning, a high-precision, low-latency closed-loop autonomous navigation framework is constructed. A Multi-Head Distillation Attention-based Trajectory Prediction (MDA-TP) method is proposed, combining a BEVFormer-based multimodal fusion perception model with a two-stage progressive knowledge distillation framework. On NuScenes, the method achieves an ADE of 0.88 m, an FDE of 1.32 m (4.34% and 10.81% reductions), and a collision rate of 15.2%, with 42.6 M parameters and 46 ms latency. Ablation shows removing attention distillation increases FDE by 13.6%. For system validation, a multi-sensor UGV platform is built. Through NuScenes online testing and real-world closed-loop validation, the Euclidean deviation remains within 0.5 m. Compared with traditional distillation, speed prediction MSE is reduced by 51.5%, wheel angle RMSE by 58.4%, and route completion improves from 60.99% to 97.26%. The results provide practical support for autonomous driving in smart cities, disaster rescue, and infrastructure inspection. Full article

The increasing integration of photovoltaic systems into modern power grids requires forecasting models that not only provide accurate predictions but also reliable uncertainty quantification under evolving operating conditions. In this paper, we propose an Out-of-distribution-aware time series conformal prediction framework with adaptive retraining, designed to address key limitations of standard conformal prediction methods in temporally dependent and dynamically changing environments. The framework is built upon the Ensemble batch prediction intervals method, which enables distribution-free uncertainty quantification without relying on a fixed calibration set, making it particularly suitable for time series applications. To ensure robustness to distribution shifts, a conformal out-of-distribution detection module is incorporated, where out-of-distribution detection is formulated as a hypothesis testing problem and enhanced through calibration-conditional p-values obtained via the Simes correction, providing conservative false-positive control intended to limit unnecessary model retraining. The proposed framework demonstrates superior performance compared to state-of-the-art approaches in uncertainty quantification, while conformal out-of-distribution detection reduces false positives and the adaptive retraining mechanism ensures effective adaptation to evolving data distributions in real-world scenarios. Full article