International audience

Deep learning-driven digital twin system for pedestrian tracking and evacuation load assessment in public spaces

Enhanced maximum power point tracking for photovoltaic systems: A Modified African Vulture Optimization (MAVO) algorithm approach

Precision tracking of a piezoelectric nanopositioning system via cascaded serial ILC-based variable-parameter closed-loop inversion feedforward control

Finite-time trajectory tracking of multi-agent systems via higher-order dynamic networks with channel damage

GPS tracking reveals complex resource acquisition strategies of pastoralists in East Africa

Strengthening the health workforce is a central policy challenge in fragile and low-income settings, yet governments often lack basic information on what it costs to train health professionals. This study provides the first institutional level evidence on the structure and magnitude of medical education costs in Afghanistan using detailed administrative and financial records from Kabul Medical University. A retrospective costing approach combined with a step-down cost accounting framework is used to allocate all direct and shared expenditures across faculties and across years of study. The results reveal large differences in per student costs across programs, with Curative Medicine requiring substantially more resources than Stomatology, Nursing, and Public Health. Support services account for a large share of total expenditure, and resource use rises sharply in clinically intensive years. A sensitivity analysis that varies personnel costs, allocation rules, and price adjustments shows that these patterns remain stable. The findings indicate that fiscal pressures, gender imbalances in enrollment, and inefficient allocation of support costs limit the effectiveness of current training investments. The study provides a quantitative foundation for policies that seek to improve the financial sustainability of medical universities, align training capacity with national health needs, and strengthen resource tracking systems in fragile and low-income contexts.

The Intermediate Silicon Tracker (INTT), a two-layer barrel silicon strip tracker, is a key component of the tracking system for sPHENIX at the Relativistic Heavy Ion Collider (RHIC) at Brookhaven National Laboratory. The INTT is designed to enable the association of reconstructed tracks with individual RHIC bunch crossings. To evaluate the performance of preproduction INTT ladders and the readout chain, a beam test was conducted at the Research Center for Accelerator and Radioisotope Science, Tohoku University, Japan. This paper presents the performance of the INTT evaluated through studies of the signal-to-noise ratio, residual distribution, spatial resolution, hit-detection efficiency, and multiple track reconstruction.

AI-Driven Three-Dimensional Movement Analysis for Objective Assessment of Surgical Expertise in Suturing.

Adaptive helmet design for optically pumped magnetometry

Academic aspirations versus institutional realities: evaluating major selection and structural barriers for students with disabilities.

Focus on the experts: Co-designing an augmented reality eye-gaze tracking system with surgical trainees to improve endoscopic instruction

Understanding the neural basis of behavior requires imaging cellular activity in freely moving animals, which typically demands expensive, restrictive microscopy setups. To overcome these barriers, we developed Wormspy, a cost-effective, open-source epifluorescence microscopy system for high-magnification imaging and tracking of Caenorhabditis elegans. Wormspy enables the simultaneous recording of neuronal activity and behavioral dynamics without needing the animal to be restrained. We demonstrate its utility in imaging body wall muscles, sensory neurons, and subcellular calcium events within interneuron axons. Our platform reproduces known mutant phenotypes and uncovers, to the best of our knowledge, previously inaccessible sensorimotor correlations. We show that Wormspy provides a robust, modular framework that lowers technical barriers to high-resolution neural imaging, enabling flexible experimental designs for dissecting behavior in freely moving organisms.

In response to the issue of total interference suppression in space flexible manipulator systems, a state estimation and dynamic compensation cooperative sliding mode control strategy is proposed, based on a neural network-based disturbance observer and a sliding mode observer. This approach leverages the approximation capability of neural networks to estimate and compensate for the system's composite uncertainties in real-time. By combining a nonsingular terminal sliding mode control law, it ensures that the system's state converges to the equilibrium point within a finite time, significantly improving the system's tracking accuracy and robustness against disturbances. Considering the practical scenario where the system's states may not be fully measurable, a sliding mode observer is designed to accurately reconstruct the system's unknown states. Based on these estimated values, a cooperative sliding mode controller is designed. This cooperative control scheme effectively addresses the control challenges when the system's states are partially unmeasurable. Simulation results show that the system maintains excellent dynamic performance and stability, even in the presence of nonlinear factors such as friction.

Visual attention in complex environments is shaped by multiple stimulus characteristics, yet most studies have examined only one or two features at a time. Eye-tracking combined with a change blindness paradigm offers a sensitive way to evaluate how different visual features guide the detection of changes in naturalistic scenes. This study aimed to validate visual stimuli with varying characteristics and to determine which factors are most significantly associated with visual attention. Fifty university students viewed pairs of real-world scenes while their gaze behavior was recorded using an eye tracking system. Participants were required to detect a single changed item across image pairs. Four visual characteristics were manipulated: social versus nonsocial content, number of distractors, congruency, and level of interest (central vs. marginal). Response times varied significantly across all characteristics. Faster detection was observed for social stimuli, scenes with fewer distractors, congruent changes, and items of central interest. Linear mixed-effects models (LMMs) revealed that level of interest was the strongest predictor of response time, followed by the number of distractors, congruency, and social content. Eye-tracking measures, particularly response time to detect change, demonstrated that central interest and fewer distractors were the strongest predictors of visual attention. These findings accentuate the complexity of visual attention and highlight its potential impact on interpreting environmental stimuli and the broader effects on academic performance and daily functioning.

Advances in mosquito olfaction: Genetic, genomic, and behavioral approaches in Aedes aegypti.

The current digital recruitment landscape is heavily mediated by algorithmic parsing technologies, requiring job seekers to submit highly standardized, machine-readable professional documents. Traditional word processors frequently introduce formatting anomalies, while legacy web-based resume builders suffer from constrained data portability and complex backend rendering overhead. This report provides an exhaustive architectural evaluation and operational analysis of "Resume Craft," a contemporary Single Page Application engineered to resolve these inefficiencies. Constructed utilizing React 19, Vite, and Google Firebase v12, the platform facilitates real-time data binding, cloud-synchronized persistence, and purely client-side document compilation. By employing the html2canvas and jsPDF libraries, the system translates the Virtual DOM directly into a downloadable Portable Document Format payload within the local browser environment. A comprehensive system architecture analysis is presented herein, contrasting this serverless, client-side generation paradigm against traditional server-side rendering methodologies concerning computational resource allocation, network latency, and semantic data preservation. The findings indicate that while the Backend-as-a-Service approach affords exceptional horizontal scalability, rapid deployment, and high-fidelity authoring experiences, the reliance on HTML Canvas-based rasterization for document generation introduces a critical operational vulnerability. Specifically, the conversion of text into image matrices severely degrades Applicant Tracking System parsing efficacy, effectively rendering the candidate invisible to recruitment algorithms. The study concludes by proposing hybrid rendering architectures, server-side headless browser integrations, and advanced state management migrations to optimize both infrastructural efficiency and the end-user's employment prospects in highly competitive, artificially intelligent hiring ecosystems.

Abstract Fabry-Pérot (FP) etalon is an important instrument for high-precision wavelength calibration. Combined with white light, the FP etalon can generate dense and nearly equally spaced frequency peaks, having an advantage in short-term calibration over thorium-argon lamps. However, owing to variations in the cavity’s physical parameters, its resonant frequency will undergo slow drift, resulting in a significant decrease in long-term calibration precision, even strict environmental control cannot completely eliminate this effect. Therefore, equipping drift tracking systems for the FP etalon represents a valuable approach to improve long-term accuracy. FP cavities with meter-level lengths have been fabricated using Anti-Resonant Hollow Core Fiber (AR-HCF), which facilitates the realization of long cavity lengths while maintaining low temperature sensitivity. The drift tracking system is constructed with the External Cavity Diode Laser (ECDL), rubidium absorption cell, and AR-HCF FP cavity. The experimental results show that when the length of the AR-HCF FP cavity reaches above 1 m, it can effectively capture the local characteristics of laser frequency scan and achieve an accuracy of MHz-level.

The hospitality industry requires effective cost control to maintain operational efficiency, particularly in the Food and Beverage department. The pastry section is highly vulnerable to waste due to its use of high-cost and perishable raw materials. This study aims to analyze the implementation of food cost control techniques to reduce raw material waste in the Pastry Department at Padma Resort Legian, Bali. This research uses a qualitative descriptive approach with field research methods. Data were collected through observation, interviews, and documentation involving pastry staff, supervisors, and cost control personnel. The results show that food cost control techniques have been applied through standard recipes, portion control, inventory management using forecasting and FIFO methods, variance analysis, and waste monitoring. However, several obstacles were identified, including human error, inconsistent implementation of SOPs, inadequate procurement specifications, and unstable storage conditions. To overcome these challenges, the department implemented corrective actions such as waste tracking systems, yield optimization, improved coordination, and regular staff briefings. Overall, the implementation is effective but requires stronger consistency and supervision to achieve optimal efficiency.

Motion-induced artefacts are a common challenge in pediatric magnetic resonance imaging (MRI), often degrading image quality and leading to frequent use of sedation or general anesthesia (GA) to minimise movement and ensure image clarity. However, repeated exposure to GA raises both logistical challenges and potential neurodevelopmental concerns. This dataset includes cerebral MRI scans from 47 children aged 4 to 11 years, who were clinically referred for imaging under GA. The participants were trained and afterwards imaging was performed without GA. MRI scans were acquired with or without motion correction using a marker-less tracking system integrated into the standard clinical workflow. Each MRI sequence is accompanied by radiologist-based quality ratings, automated reference-free quality metrics, and motion traces. For standardisation, all image data are structured in compliance with the Brain Imaging Data Structure (BIDS) format. This dataset provides a valuable benchmark for testing image-based motion correction methods and developing motion-tolerant imaging strategies, addressing a critical gap in a publicly available clinical pediatric MRI data acquired under realistic motion conditions.