Advanced Electromagnetics

Observation-Aware Calibration of Agent-Based Epidemic Simulations from Public Aggregate Reports: A COVID-19 City-Report Benchmark Study

2026-08-19T18:12:24+02:00

City reports record confirmation, recovery or discharge, and death by publication date; agent-based simulations track latent disease events. We calibrated Covasim through an observation layer that converts cumulative reports to daily increments and aligns them with simulation outputs. A two-parameter Optuna-TPE search estimated transmission and initial infections. Shenzhen was reproduced under fixed stochastic settings, followed by independent recalibration in 12 non-Hubei first-wave cities. The Shenzhen fit reproduced the reported peak within one day (RMSE 7.2072, MAE 5.5848) but underestimated total reported cases by 20.66%. Gompertz fitted the same short curve more closely (RMSE 3.6874). Across the city panel, median RMSE was 3.3049 and peak-normalized RMSE was 0.1933. The framework links report-curve calibration to stochastic, auxiliary, and scenario outputs within a historical COVID-19 benchmark. Policy evaluation lies outside this benchmark.

Research on the Construction and Application Innovation of Smart Sports Training Scenarios Driven by Digital Intelligent Technology

2026-08-19T18:12:24+02:00

Against the backdrop of deep integration between digital intelligent technologies and the sports industry, traditional sports training models are plagued by prominent drawbacks including homogenization, insufficient precision and weak data support. Smart sports training has evolved into a core trend for the scientific development of sports. Centered on the upgrading of sports training scenarios empowered by digital intelligent technologies, this paper analyzes the current development status and existing problems of smart sports training scenarios, constructs a multi-layer, multi-sub-scenario smart sports training system, explores its innovative application paths, analyzes application effects through empirical research, and puts forward targeted optimization strategies. The research demonstrates that digital intelligent technologies including the Internet of Things (IoT), big data, artificial intelligence (AI) and digital twins can effectively resolve pain points of traditional sports training and realize the upgrading of personalized training, dynamic monitoring and integrated training-competition systems. Nevertheless, prevailing challenges such as homogenized scenarios, incomplete data systems and insufficient deep integration of technologies still remain. In the future, standardized construction, in-depth technological empowerment and industry-university-research collaborative development will promote the large-scale and refined development of smart sports training scenarios, providing theoretical references and practical support for the digital transformation of sports training.

Animation of Historical Images: A Video Diffusion Generation Method Guided by Depth Estimation

2026-08-19T18:12:24+02:00

The use of artificial intelligence and smart algorithms for the restoration, enhancement and creation of videos from historical images continues to grow. To tackle the problems of spatial structure drift, inter-frame flickering, and motion discontinuity in the animation of historical images, we propose a video diffusion generation model based on depth estimation. Based on the video diffusion model, the model includes the historical image preprocessing module, the single-frame depth estimation module, the spatial structure encoding module, and the temporal motion compensation module, using the depth feature as a conditional constraint in the reverse denoising generation of the model. Experimental results show that the proposed method achieves a PSNR of 28.76, an improvement of 1.28 over conventional video diffusion models; SSIM increases to 0.883, LPIPS decreases to 0.108, and FVD decreases to 219.54, with both generation quality and temporal stability outperforming the comparison methods.

Research on the Driving Factors of Impulse Buying Behavior among Consumers in Live Streaming E-Commerce

2026-08-19T18:12:24+02:00

Live streaming e-commerce compresses real-time display, interactive Q&A, community gathering, and instant payment into the same scene, significantly reducing the time for consumers to complete payment from product contact and increasing the possibility of unplanned purchases. This paper does not fabricate survey data, but builds an explanation framework of "platform and marketing stimulation - cognitive and emotional state - purchase impulse - actual behavior" based on the stimulus organism response model, integrating research on impulse purchase, availability of information technology, social presence and flow experience, and combining the public data of China Internet Network Information Center. Research suggests that limited time discounts, scarce prompts, anchor credibility, interaction quality, product matching, algorithm recommendations, and group leads do not work in isolation, but rather drive decision-making through perceived value, trust, arousal, and immersion; Price level, product involvement, availability of time and funds, impulse buying tendency, and consumer self-control constitute important boundaries. Based on the above mechanism, the article proposes a governance plan with true disclosure, moderate promotion, algorithm moderation, and calm decision support as the core, providing reference for platform optimization of business quality, merchants to reduce returns and complaints, and consumers to form rational purchasing ability.

Composite Design and Friction-Wear Characteristics of High-Efficiency Transmission Gear Materials

2026-08-19T18:12:24+02:00

To meet the high-speed and heavy-load service requirements of high-efficiency transmission gears, Fe-2Ni iron-based powder metallurgy was used as the matrix, and a composite system reinforced by TiC hard phase and MoS2 solid lubricant was designed. The gear composites were prepared by hot pressing sintering. The effects of reinforcement ratio and sintering parameters on microstructure and mechanical properties were studied, and the friction and wear behaviors under different loads, sliding speeds and lubrication conditions were investigated. The results show that the optimal sintering process is 1150 ◦C, 60 min and 30 MPa, with a relative density of 98.6% and microhardness of HV 620. The composite with 10 wt% TiC and 3 wt% MoS2 exhibits the best comprehensive performance, with a friction coefficient as low as 0.08 and a wear rate reduced by 62% compared with the matrix. The wear mechanism transforms from abrasive wear to oxidative wear and slight delamination wear. The composite presents excellent friction reduction and wear resistance, which can provide a quantitative reference for the design of high-performance transmission gears.

Business Environment Optimization and Audit Quality in the STAR Market: Inverted U-Shaped Nonlinear Effect and ESG Risk Moderation Mechanism

2026-08-19T18:12:24+02:00

The business environment constitutes a fundamental institutional arrangement shaping the behavioral incentives of capital market participants. However, whether its relationship with audit quality follows a monotonic linear pattern or exhibits more complex nonlinear dynamics remains an open empirical question. Using a sample of 4,502 firm-year observations from the STAR Market and ChiNext Board during 2020–2024, this paper employs the modified Jones model to measure audit quality and constructs a comprehensive provincial business environment index encompassing four dimensions — government efficiency, legal environment, financial development, and market maturity — using the entropy-weight method. Through individual fixed-effects panel models, we systematically examine the impact of business environment on audit quality. The findings reveal three key insights. First, a significant inverted U-shaped nonlinear relationship exists between business environment and audit quality, with an inflection point at approximately 0.544. Second, ESG disclosure serves as a risk-moderating barrier rather than a mediating channel. Third, significant heterogeneity is observed across regions, industries, and firm life-cycle stages. This paper breaks the conventional assumption of a linear institutional-audit quality relationship and provides new empirical evidence and policy implications for audit regulation under the registration-based IPO system.

Data-driven Personalized Family Music Therapy: a Reinforcement Learning Recommendation Model Based on Family Member Behavior and Feedback

2026-08-19T18:12:24+02:00

This study proposes a data-driven personalized family music therapy recommendation framework to address the limitations of conventional recommendation systems in modeling emotional interactions among multiple family members. The framework employs a reinforcement learning approach tailored to family environments, where emotional states evolve through continuous interpersonal interactions and feedback. A relationship-aware family state representation is developed by integrating behavioral observations and physiological responses, enabling the modeling of both individual emotional states and emotional contagion among family members. To support adaptive intervention, the action space incorporates music selection, spatial delivery strategies, and intentional silence decisions, allowing the system to balance active intervention with natural emotional self-regulation. In addition, a multi-channel reward mechanism is designed by combining physiological recovery, behavioral relaxation, and environmental stability indicators, while a temporal credit assignment strategy addresses delayed therapeutic effects. Experimental evaluation on a multimodal dataset covering four representative family stress scenarios demonstrates that the proposed framework achieves 89.7% accuracy in family emotion representation, 86.3% consistency with expert intervention judgments, and an 82.3% positive emotion conversion rate after continuous intervention. These findings suggest that reinforcement learning can provide an effective foundation for transforming music recommendation systems from individual entertainment tools into proactive family emotional support systems.

Research on the Network Architecture of Secure Power Wireless Local Area Network Based on WAPI and SDN

2026-08-19T18:11:58+02:00

In smart-grid WLANs, a terminal is not fully secured once it has merely passed wireless access authentication. The more difficult part is how the authentication result is carried forward into service isolation, forwarding control, bandwidth allocation, and security operation. In many conventional WLAN deployments, these functions are configured in separate systems, so terminals for distribution automation, electricity information collection, inspection robots, and other power services may still depend on similar access and management policies. This separation also makes burst traffic and O&M fault tracing harder to handle. To deal with this coupling problem, this study designs a secure power WLAN architecture that combines Wireless Local Area Network Authentication and Privacy Infrastructure (WAPI) with Software-Defined Networking (SDN). WAPI provides public-key-based two-way authentication, while SDN supplies centralized control through control-forwarding separation. In the architecture, the SDN controller manages WAPI servers, AP rules, and network resources together. Security domains and service domains are isolated by VLANs, and OpenFlow-supported WAPI-APs are used in the data layer to keep authenticated access, encrypted transmission, and flexible forwarding under the same control logic. A security-resource linkage module maps WAPI authentication results to SDN resource policies, so trusted power terminals can obtain service-specific access permissions and bandwidth guarantees. For typical power scenarios, differentiated WAPI policies and a “detection-alarm-disposal” monitoring loop are also introduced. The experimental results show end-to-end latency below 5 ms, throughput above 93%, and packet loss below 7%; unauthorized access alarms are reduced by 94.8% after WAPI-based protection is enabled. These results suggest that using WAPI authentication semantics as an input to SDN policy control can improve both access security and operational manageability in power WLANs.

FedCAD: Federated Contrastive-Aware Decoupling for Heterogeneous Power System API Security Detection

2026-08-19T18:11:58+02:00

Digital power grids increasingly rely on API-based interaction among dispatch automation systems, distribution management systems, advanced metering infrastructure, marketing platforms, energy management systems, distributed energy resource access platforms, and edge IoT gateways. These interfaces support remote monitoring, operation command delivery, load forecasting, meter data acquisition, fault isolation, and cross-domain business collaboration, but they also introduce new attack surfaces for unauthorized command invocation, abnormal parameter tampering, replay access, excessive service calling, and sensitive grid-data leakage. Building an accurate API security detection model for such scenarios is challenging because power-grid data cannot be freely centralized: dispatch logs, equipment identifiers, customer electricity information, station topology, and operation records are distributed across regional grids, voltage levels, business departments, and terminal types, and are subject to strict privacy and operational-security constraints. However, power-system API security data usually exhibits multiple forms of heterogeneity. Label distributions vary because different regions and business systems face different proportions of normal operation, abnormal metering access, dispatch-command abuse, distribution-terminal intrusion, and new-energy access anomalies. Feature distributions vary because API traffic is strongly coupled with grid operation modes, voltage levels, device types, communication protocols, seasonal load patterns, and local gateway configurations. In addition, missing timestamps, clock drift, incomplete gateway logs, noisy alarm labels, and irregular field-device communication may corrupt the observed data. These factors lead to inconsistent power API behavior representations, shifted risk decision boundaries, and local overfitting in traditional distributed learning methods. To address these challenges, this paper proposes FedCAD (Federated Contrastive-Aware Decoupling), a federated framework for heterogeneous power system API security detection. FedCAD decouples representation learning from classifier adaptation. A hierarchical contrastive learning module aligns power API behavior representations at both instance and risk-category levels, while preserving region-specific and business-specific operational characteristics. An adaptive classifier module further adjusts decision boundaries according to local power API risk distributions through logit adjustment and local fine-tuning. The proposed framework enables collaborative security modeling across dispatch, distribution, metering, marketing, and new-energy systems without sharing raw power-grid data, improving robustness and generalization under multi-source heterogeneous conditions.

Optimization and Accuracy Improvement of Power Forecasting Models for Wind Farms Under Transient Weather Conditions

2026-08-19T18:11:58+02:00

Given the problems of non-stationary power time series, response lag, and amplified prediction errors due to sudden changes in wind direction under transitional meteorological conditions, this study proposes a wind power forecasting model that integrates multiscale time-series features, transition-aware attention mechanisms, physical constraints on turbine operation, and dynamic residual correction. To improve the model's ability to jointly characterize different scales of meteorological evolution and power lag characteristics, a feature extraction network with short-, medium-, and long-term branches based on TCN (Temporal Convolutional Network) is built, which incorporates bidirectional GRU and Transformer architectures; additionally, the feature weights at each scale are dynamically adjusted according to the severity of inflection points, and the prediction output is constrained by air density, power curves, operational status and ramping limits. Residual caching and time-delay gating are employed to correct for lag errors in the range of 1 to 6 sampling steps. 50,842 valid data sets from a wind farm were used for validation. The general MAE and RMSE of the model were 0.258 MW and 0.386 MW, respectively, and these were lower than those of the baseline TCN by 21.58% and 20.90%. In the transition period, the MAE and RMSE were 0.305 MW and 0.448 MW; these had been reduced by 22.19% and 23.29%. Therefore, the developed model can reduce peak deviations caused by abrupt changes in weather and improve the accuracy and stability of wind power forecasting under all operating conditions.