搜索结果：Workspaces

Shared immersive environment sports venues, virtual classrooms, and collaborative workspaces require multiple users to stream 360° videos simultaneously over the same edge network, yet every existing adaptive bitrate system optimises each viewer in isolation. This self-interested behaviour triggers a bandwidth auction that chronically starves the most uncertain viewers: Jain's Fairness Index for ten independently optimised agents routinely falls below 0.85. We present FairEdge360, a hierarchical multi-agent reinforcement learning framework that reformulates multi-user 360° streaming as a Decentralised Partially Observable Markov Decision Process (Dec-POMDP) and proves, formally, that fairness and quality are complementary rather than competing objectives. Three tightly coupled innovations make this possible. First, a Lightweight Uncertainty Estimator (LUE) a compact 8385-parameter four-layer MLP evaluates per-device viewport prediction confidence cti=σ(w4⊤h3) in under approximately 2.1 ms on commodity smartphones (95th percentile, iPhone 12 A14 Bionic), enabling selective edge offloading that reduces device energy consumption by 38.9%. Second, a variational Graph Neural Network compresses each agent's 256-dimensional GRU state into a 32-byte INT8 latent, transmitted over a dynamic RTT-gated neighbourhood graph at 96 bytes per agent per 500 ms 75% less overhead than competing approaches. Third, the edge coordinator maximises the Nash social welfare objective NSW=(∏i=1NQi)1/N, whose gradient ∂NSW/∂Qi∝1/Qi automatically prioritises the most disadvantaged viewer; a formal proof guarantees that every Pareto-optimal policy satisfies Qi/∑jQj≥1/N. Counterfactual advantage estimation correctly attributes each agent's marginal contribution to the global reward, eliminating the credit-assignment ambiguity inherent in standard multi-agent baselines. Evaluated on 284 users, 52 omnidirectional videos, and 10,000 real network traces spanning 4G LTE, 5G mmWave, HSDPA, and campus WiFi, FairEdge360 raises Jain's Fairness Index from 0.934 to 0.976 (+4.5%), improves worst-case user quality-of-experience from MOS 2.54 to MOS 3.21 (+26.4%), and halves rebuffering rate from 2.1% to 1.1%, all within a 20 ms motion-to-photon budget on a commodity smartphone.

Shared resources enhance productivity, yet at the same time open pathways for biological and digital contamination, turning physical or digital hygiene into a cooperation dilemma prone to free-riding. Here, we introduce a game of sequential sharing of common resources, an empirically parameterized evolutionary model of population dynamics in sequential-use settings such as gyms or shared workspaces. The success of the strategies implemented in the model, which involve equipment cleaning before or after use, is based on the trade-offs between cleaning costs, contamination risks and social incentives to mitigate disease transmission. We find that cooperative hygiene can be achieved by lowering the effective costs of cleaning, strengthening pro-social incentives and monitoring population-level noncompliance. Remarkably, the stability of fully altruistic populations is primarily affected by the cleaning costs. In contrast, increasing effective infection costs, for example, through punishment, appears less important in this case. The model's evolutionary dynamics exhibit multi-stability, hysteresis and abrupt shifts in strategy composition, broadly consistent with empirical observations from shared-use facilities. Our framework offers testable predictions and is amenable to quantitative calibration with behavioural and environmental data. Our predictions can be used to inform the design of cost-effective public health and digital security policies. A bird that flies away leaves no trace (Japanese proverb).

The NIH Human BioMolecular Atlas Program (HuBMAP) Data Portal (https://portal.hubmapconsortium.org/) serves as a comprehensive repository for multimodal, multi-scale spatial and single-cell data from healthy human tissues. As of June 2026, the portal hosts 9,232 public datasets from 25 data types spanning 29 organ classes across 498 donors. Portal infrastructure and user interfaces support data search and discovery, visualization, and analysis directly in web browsers. These capabilities include metadata- and data-driven search, collaborative Workspaces with access to high-performance compute, and interactive Vitessce visualizations across non-spatial, 2D, and 3D spatial datasets. Data-type-specific uniform processing pipelines and rigorous quality control processes ensure comparability of results across laboratories, organs, and donors, while externally processed community-contributed datasets provide complementary perspectives. Here we describe portal functionality, infrastructure, and design, and highlight its role as a platform for large-scale spatial single-cell research across diverse data types, organs, and scales.

The determination of hazardous gases in confined environments, such as underground mines, is essential for ensuring suitable occupational health conditions. Although gas chromatography coupled with mass spectrometry (GC-MS) is the gold-standard analytical technique for gas analysis, its application is not feasible in this context due to the bulkiness of the instrumentation and the inability to provide real-time responses. Alternative low-cost sensing techniques, such as metal oxide semiconductors (MOXs), are limited by low molecular selectivity, susceptibility to humidity interference, and long-term instability. Herein, we propose an integrated and compact analytical platform that combines multiple quantum-cascade lasers (QCLs) with substrate-integrated hollow waveguides (iHWGs) acting as miniaturized gas cells for the real-time quantification of hazardous gases, including SO2, H2S, CO2, CH4, NO, and NO2, in confined workspaces such as underground mines. Each QCL was selected and tuned to match the absorbance profile of a particular target analyte, with the exception of H2S, which was converted to SO2 via UV irradiation prior to quantification. An Arduino-based system facilitates signal acquisition and processing, enabling rapid data interpretation and wireless communication with external devices. At optimized conditions, the system enables the quantification of all target gases near their permissible exposure limits (PELs) in underground mines, with limits of detection of 0.3, 7, 100, 100, 5, and 3 ppmv for SO2, H2S, CO2, CH4, NO, and NO2, respectively. The entire system has a footprint of 470 × 320 × 100 mm, which is compatible with real-world deployment scenarios in underground mines. This advanced sensing platform provides real-time monitoring of toxic gases in confined environments with high molecular selectivity and suitable sensitivity. Designed for harsh conditions, it can operate reliably at high humidity, dust, vibrations, and electromagnetic interferences.

Assistive robotics are increasingly adopted in manufacturing to support human work. In automotive assembly, confined spaces and takt-time constraints mean technological support affects biomechanical, cognitive, psychosocial, and organisational dimensions of work. This paper presents a multi-dimensional evaluation of an upper-limb exoskeleton and a proximity-aware cobot through two studies: a focus group with automotive practitioners addressing safety, trust, acceptance, and implementation barriers, and an experimental study with workers performing assembly tasks. The technologies were assessed independently and in combination, combining objective posture and performance indicators with subjective measures of workload, usability, comfort, trust, and acceptance. While each technology reduced biomechanical load alone, their combined use yielded no additive benefits and introduced cognitive and coordination costs that offset physical gains. Findings emphasise the importance of multi-dimensional evaluation when integrating assistive technologies in automotive assembly. This study combines practitioner insights and experimental evaluation of exoskeleton and cobot use in automotive assembly. Each technology reduced biomechanical load alone, but combining them in confined workspaces added cognitive and coordination demands that offset physical benefits.

This paper introduces an advanced Internet of Things (IoT)-driven smart furniture system designed to dynamically adapt to individual users by integrating deep reinforcement learning with federated meta-learning. Personalization is formulated as a Markov decision process, enabling the system to make optimized, sequential adjustments tailored to each user's behavior. To estimate hidden ergonomic preferences in real time, an adaptive Kalman filter is applied, while a sparse autoencoder reduces raw sensor signals by 82 %, preserving key temporal features essential for accurate modeling. In a comprehensive user study involving 48 participants and more than 160,000 time-series sensor samples, the framework significantly reduced cumulative user dissatisfaction by 43 % and cut energy consumption by 21 %, compared with conventional rule-based control systems. Real-time adaptations occur with an average latency of 280 ms, and constraints for ergonomics are upheld in 95 % of use cases, confirming the system operates swiftly and safely. Federated learning (FL) enables privacy-preserving collaboration across distributed furniture units. Training converges to 87 % of global performance within 30 global iterations, without any raw data exchange, reinforcing both scalability and data privacy. These empirical results strongly support the framework's suitability for deployment in health-aware workspaces, smart homes, and eldercare environments, delivering a robust, responsive, and interpretable solution for enriching human-furniture interaction.

Network meta-analysis (NMA) plays an important role in comparative effectiveness research, particularly in fields such as traditional and complementary medicine, where multiple interventions often need to be assessed within a single evidence framework. However, conducting NMA remains labor-intensive, methodologically demanding, and difficult to complete efficiently using conventional workflows. Although recent advances in large language models have created new opportunities for supporting evidence synthesis, their routine use in NMA is still constrained by limited transparency, prompt dependency, and difficulty integrating with structured analytical procedures. In this study, we introduce SmartEBM, a web-based human-AI collaborative platform designed to support the full workflow of NMA. SmartEBM provides an integrated environment for title and abstract screening, full-text screening, data extraction, risk of bias assessment, statistical analysis, and certainty of evidence assessment. The platform is organized around a human-in-the-loop model, in which AI-assisted functions support repeated and labor-intensive tasks while researchers retain oversight of methodological judgement, verification, and final decision-making. Through its six functional modules, SmartEBM offers low-code interfaces, structured outputs, and verification-oriented workspaces that help connect major steps of evidence synthesis within one platform. Rather than functioning as a stand-alone automation tool, SmartEBM is intended as a practical platform for end-to-end NMA support. This platform-oriented approach may help make evidence synthesis more manageable, traceable, and accessible in routine research practice, especially in complex review settings.

The E-Research Institutional Cloud Architecture (ERICA) is a code-driven orchestration framework that automates the configuration and management of Amazon Web Services (AWS) resources to provide trusted research environments (TREs) for sensitive data. Independent ERICA TREs are now operational in universities and government agencies. The framework was developed by the University of New South Wales, Australia, with support from the Australian Research Data Commons. ERICA was designed to overcome the limitations of traditional on-premise TREs by providing secure, scalable, and flexible cloud environments that protect privacy while enabling advanced, data-intensive research. Using an infrastructure-as-code model, ERICA delivers consistent, reliable, and error-free setup of Project Spaces. It integrates robust security features, including encryption-at-rest and in transit, and multi-factor authentication. Hosted in AWS onshore data centres, ERICA ensures data sovereignty while supporting diverse operating systems and high-performance computing configurations. The architecture also allows rapid deployment of new AWS services, including generative AI tools, within research workspaces. ERICA implements the Five Safes framework-covering safe projects, people, data, settings, and outputs-to ensure compliance and secure research. Its modular architecture enables multiple independent TREs, each governed by host-institution policies and capable of supporting hundreds of Project Spaces. This flexibility allows replication across any jurisdiction with AWS public cloud infrastructure. However, reliance on AWS introduces challenges, including charges in US dollars and delayed rollout of new services in smaller regions. ERICA represents a step change in providing privacy-by-design cloud infrastructure for sensitive, data-intensive research. By combining strong governance with the scalability of AWS, it enables researchers to work securely with large, complex datasets while rapidly adopting cutting-edge analytical tools. ERICA TREs offer a replicable, future-proof model for supporting secure research at scale.

Articulated endoscopic instruments can enhance dexterity within constrained workspaces; however, their complexity and cost limit accessibility for training, prototyping, and early-stage evaluation. Additive manufacturing offers a practical approach for developing mechanically functional, low-cost articulated devices for engineering exploration and simulation-oriented applications. This study describes the mechanical design, bench-top characterization, and preliminary proof-of-concept assessment of two additively manufactured, cable-driven endoscopic prototypes: a flexible endoscope with an integrated working channel and bidirectional distal articulation, and a 4-mm steerable surgical forceps with independent jaw actuation. Both devices were fabricated using fused deposition modeling. Bench-top experiments quantified distal deflection, motion repeatability, force-related behavior, simulated task execution, and user perception. The flexible endoscope achieved effective distal deflections of 178° ± 2° vertically and 171° ± 3° horizontally while maintaining compatibility with standard 2.8-mm endoscopic instruments. During cyclic testing, it demonstrated sub-centimeter positional consistency after 1000 articulation cycles, with 95th percentile errors below 8 mm in both directions. The steerable forceps exhibited a geometric articulation range of 180°, an effective distal deflection of 85° ± 2°, and repeatable tip motion over 500 cycles, with a mean positional error of 0.26 mm. Simulated tasks were completed with high user perception scores for both prototypes. Fabrication costs were USD 166.49 for the endoscope and USD 27.75 for the forceps. These findings suggest that cable-driven architectures combined with additive manufacturing can reproduce key kinematic principles of articulated endoscopic instruments using low-cost fabrication methods. The prototypes should be interpreted as early-stage engineering demonstrators rather than clinically deployable devices.

The rapid shift to remote work following the COVID-19 pandemic transformed private residential spaces into permanent workplaces, often without professional ergonomic oversight. While occupational lighting is strictly regulated by standards such as EN 12,464-1, these norms are legally unenforceable in private homes due to privacy regulations. This study addresses the "Privacy-Safety Paradox" by conducting a comparative legal analysis of employer responsibilities versus privacy rights in the EU (including France, Germany, Italy, Portugal, and Poland) and the USA. It presents a case series analysis of five distinct home-office scenarios, evaluating illuminance levels (Eavg, Emin) and uniformity (U0) against occupational safety requirements. Results reveal a 100% non-compliance rate, with critical deficiencies identified in makeshift workspaces. The most severe case exhibited an average illuminance of only ~ 188 lx and a minimum of ~ 22 lx, which is drastically below the biologically required thresholds for visual health and circadian regulation. The study identifies a systemic "compliance gap" where legal protections of the home, such as the German Selbstauskunft model or the Italian "right to disconnect", limit the employer's ability to perform physical audits. To mitigate this risk, the authors propose a simplified self-assessment checklist designed to allow remote workers to identify and correct visual hazards without specialized photometric equipment.

An augmented virtuality framework is presented in which the physical scene is admitted only through developer-defined geometric apertures, while all other pixels are rendered virtually. The scene initialises as a black field and uses surface-projected, overlay-style compositing so that the camera stream is visible only on circular or square meshes registered as projection surfaces. Tracking origin is configured to minimise unintended recentring, and the same method can operate either as a minimal black scene or within an optional three-dimensional environment. Main features of the framework are as follows: deterministic compositing that binds real imagery to user-defined meshes and prevents leakage outside apertures. stable alignment between virtual geometry and the physical workspace during head motion achieved through standard XR configuration. a flexible scene recipe that supports circular or square apertures and optional contextual environments without altering projection logic. The framework is intended for sensory and consumer studies that require real product interaction under controlled context and generalises to training, human factors, and rehabilitation scenarios requiring constrained visibility of the real world.

Rapid sequence intubation (RSI) in helicopter emergency medical services (HEMS) is conventionally performed at the scene before transport, potentially delaying time to definitive care. The feasibility of performing RSI during flight in civilian HEMS operations has not been established in the United Kingdom. We evaluate the feasibility, safety, and temporal efficiency of performing simulated in-flight RSI in an AW169 helicopter under operational flight conditions. A prospective proof-of-concept study using high-fidelity simulation was conducted across 2 phases (May 2023, January 2025) at the Air Ambulance Charity Kent Surrey Sussex. Eight simulations were completed by 4 distinct clinical teams comprising operationally experienced HEMS physicians and paramedics. Scenarios replicated a standardized traumatic brain injury scenario requiring RSI during the return transit phase. The primary outcome was time from RSI checklist initiation to confirmed intubation. Secondary outcomes included overall mission times, safety events, and crew-perceived feasibility assessed via post-scenario questionnaires. All simulations (8/8, 100%) achieved successful first-pass intubation. Median time from checklist initiation to confirmed intubation was 5 minutes (interquartile range [IQR]: 5-7). Median total mission time from base departure to RSI completion was 42 minutes (IQR 40-44). No safety events, procedural complications, or communication failures occurred. Crew questionnaires (93% response rate) confirmed unanimous perceived feasibility, with participants identifying adequate workspace, effective communication, and manageable equipment accessibility. High-fidelity simulation demonstrates that in-flight RSI is technically feasible in an AW169 helicopter, with consistent procedural times and no safety events. These findings may support further evaluation of in-flight RSI as a complementary strategy for time-critical patients where scene-based airway management may delay definitive care.

Background/Objectives: Efficient patient flow in urgent urology services is critical to timely care delivery, yet workflow bottlenecks in specialty clinics remain underexplored. This study aimed to identify workflow bottlenecks, evaluate patient flow and staff attitudes, and explore clinician readiness for digital decision-support in a high-volume NHS emergency urology walk-in clinic. Methods: A two-week observational study was conducted at an emergency urology service in London. Time-stamped pathway data were collected for 80 patient journeys to identify total clinic duration. Differences associated with investigation ordering and senior escalation were analyzed using t-tests. Clinicians (n = 34) completed a questionnaire assessing perceptions of AI, and nursing staff provided qualitative feedback on operational pressures. Results: Mean total clinic journey time was 2 h 42 min, with the post-assessment phase accounting for 64% of total duration. Investigation ordering was the principal source of delay: patients undergoing investigations remained significantly longer in clinic than those who did not (3 h 17 min vs. 2 h 15 min, p < 0.05), and doctor-to-discharge time more than doubled (2 h 20 min vs. 1 h 2 min, p < 0.005). Senior escalation did not significantly prolong patient flow. Staff surveys demonstrated moderate trust in and comfort with AI as a decision-support tool. Nursing feedback highlighted inappropriate attendances, limited staffing, and workspace constraints as key stressors. Discussion: Delays were primarily driven by investigation ordering rather than senior review, identifying investigation timing as a potential target for future pathway optimisation. Conclusions: Investigation-related delays were the dominant workflow bottleneck. While no AI system was deployed in this study, these findings provide empirical groundwork to inform the design and prospective evaluation of AI-supported triage in specialty acute care settings.

Cable-driven planar robots have attracted increasing attention due to their simple mechanical structure and flexible workspace, while achieving accurate trajectory tracking in discrete-time settings remains challenging because of nonlinear kinematics and real-time computational requirements. In this paper, a discrete-time neural dynamics (DTND)-based online quadratic programming (QP) framework is developed for discrete-time trajectory tracking of a cable-driven planar robot with a structure consistent with typical wall-drawing (V-plotter) systems. The inverse kinematics problem of the robot is first formulated as an online QP by transforming the desired end-effector trajectory into a discrete-time velocity tracking objective. Inspired by DTND for time-varying quadratic programming, a DTND-based iterative solver is employed to compute the cable-length rates at each sampling instant, enabling real-time trajectory tracking without explicit offline optimization. The proposed method operates entirely in discrete-time form and is suitable for digital implementation. Numerical simulations based on a cable-driven planar robot demonstrate that the proposed DTND-based online QP approach achieves stable error convergence and accurate trajectory tracking. Furthermore, physical experiments on a cable-driven planar robot are conducted to validate the practical feasibility of the proposed DTND-based framework and demonstrate its applicability in real-world engineering scenarios.

Identifying the neural correlates of conscious visual perception remains a major challenge in neuroscience, requiring theories that bridge between subjective experience and measurable neural correlates. However, theoretical interpretation of empirical evidence is often post hoc and susceptible to confirmation bias. Building upon the adversarial collaboration mediated by the COGITATE Consortium, we present a generalizable approach for the data-driven identification, evaluation, and theoretical modeling of connectivity-based neural correlates of conscious visual perception. Using the same magnetoencephalography (MEG) dataset and accompanying pre-registered hypotheses from the COGITATE Consortium, we systematically compared 246 functional connectivity (FC) measures between regions predicted to underlie conscious vision by Integrated Information Theory (IIT) and/or Global Neuronal Workspace Theory (GNWT). We identified a family of FC measures based on the barycenter-tracking the 'center of mass' between two signals-as the top-performing stimulus decoding measures that generalize across regions central to predictions of both IIT and GNWT. To interpret these findings within a theoretical framework, we developed neural mass models that recapitulate the neural dynamics hypothesized to underlie conscious perception by each theory. Comparing simulated barycenter values from these models against empirically measured MEG data revealed that both the GNWT-based model, featuring delayed ignition dynamics, and the IIT-based model, which relied on synchronous sensory dynamics, captured the observed connectivity patterns. These results lend tentative support to GNWT, as the presence of ignition dynamics independent of task-demand conditions contradicts the predictions of IIT. Beyond dataset-specific conclusions and limitations, we introduce a framework for systematically identifying and testing candidate neural correlates of conscious visual perception in an unbiased and interpretable manner.

Understanding whether prefrontal cortex is necessary for conscious experience remains central to ongoing debates between global workspace and recurrent processing theories of consciousness. To dissociate neural processes underlying perceptual awareness from those related to report and access, we combined report and no-report visual masking paradigms with magnetoencephalography (MEG) to characterize the spatiotemporal and network dynamics of visual awareness. Across both report and no-report conditions, early and sustained responses emerged in occipital and temporal cortices, beginning at &#x223c;60-70&#xa0;ms post-stimulus. Posterior sensory regions consistently encoded stimulus presence and category information irrespective of reporting demands, and exhibited enhanced &#x3b1;-band recurrent coupling for visible stimuli, suggesting stable perceptual representations supported by local recurrent interactions. In contrast, prefrontal activity emerged only when explicit report was required, beginning at &#x223c;100&#xa0;ms and persisting thereafter. Time-resolved decoding revealed that anterior regions encoded stimulus information exclusively under report conditions and failed to represent categorical content when reporting was absent. Moreover, long-range fronto-posterior connectivity selectively increased during report, indicating task-dependent integration beyond posterior sensory processing. These findings demonstrate that posterior cortical dynamics are sufficient to support perceptual awareness, whereas prefrontal cortex contributes to report-related access and global integration. By disentangling awareness from reporting, this work provides evidence favoring posterior-centered accounts of phenomenal consciousness and clarifies the functional role of anterior cortex in conscious processing.

This paper presents SPINDER, a compact 18-DoF hexapod robot designed for reproducible bio-inspired locomotion research. The platform integrates a time-of-flight depth camera, six-axis inertial measurement unit, and per-foot contact switches, with computation split between an ARM-based single-board computer and a real-time microcontroller driving Dynamixel actuators. We derive a closed-form inverse kinematics solver for the yaw-pitch-pitch leg configuration that enforces workspace feasibility and handles kinematic singularities, achieving real time computations per leg. The locomotion controller combines a coupled-oscillator central pattern generator for inter-leg coordination with local motion controllers that regulate torso motion and generate cycloid-based swing trajectories with stance compensation, before mapping foot targets to joint commands via analytic IK. Hardware experiments demonstrate stable tripod, ripple, and wave locomotion, including omnidirectional commands, smooth online gait transitions that settle within approximately two gait cycles, and successful traversal of structured uneven terrain (height variations up to 18&#xa0;mm) with 100% completion rate and mean body-pitch deviation below 5[Formula: see text]. All mechanical designs, kinematic models, and control software will be released to support replication and extension.

The global dissemination of carbapenem resistance is predominantly facilitated by plasmid-mediated carbapenemase genes, notably blaOXA-48-like genes. A comprehensive understanding of their evolutionary relationships and structural conservation is essential for monitoring their spread and informing therapeutic strategies. This study aimed to investigate the phylogenetic relationships and structural conservation of blaOXA-48-like carbapenemase genes in multiple Gram-negative bacterial species. We analysed blaOXA-48-like carbapenemase sequences obtained from a hospital in Pakistan and compared them with globally reported variants retrieved from GenBank. Carbapenemase gene sequences (blaOXA-48-like, blaNDM, and blaVIM) were analyzed using maximum-likelihood phylogenetics (MEGA11, Tamura-Nei model, 1000 bootstrap replicates). Comparative global sequences were retrieved from GenBank. Structural modeling of blaOXA-48-like genes was performed using SWISS-MODEL Workspace with the template PDB 3HBR, followed by validation using GMQE, QMEANDisCo, and Ramachandran plot analyses. Phylogenetic analysis revealed a tight clustering of blaOXA-48-like genes across A. baumannii, K. pneumoniae, and E. meningoseptica, showing high similarity to globally distributed plasmid-associated sequences. Structural modeling demonstrated strong conservation of the enzyme, with preserved catalytic residues (Ser70, Lys73, Ser118, Trp157, and Tyr211) and minimal structural deviation (RMSD < 0.3 &#xc5;). blaOXA-48-like carbapenemases exhibit strong phylogenetic conservation and structural stability across species and regions, consistent with the horizontal dissemination of blaOXA-48-like genes across bacterial hosts. These findings indicate that blaOXA-48-like carbapenemases have high evolutionary stability.

暂无摘要（点击查看详情）