Complex scene segmentation aims to segment objects with intricate details or those concealed within the background. Despite significant advancements, a persistent challenge remains: accurately identifying object edges in backgrounds with high inherent similarity and complex structures. To address this, we identify the prevalent spectral bias in image segmentation, where networks preferentially learn low-frequency information, as a key impediment to recognizing and learning object edges, which are rich in high-frequency details. To mitigate this bias, we propose MCNet, a segmentation framework designed to promote balanced frequency learning. MCNet comprises two primary components: multi-frequency perception (MP), which independently captures high-frequency details and low-frequency structural components of objects, and complementary fusion (CF), which intelligently fuses these distinct frequency features through learnable, adaptive mechanisms. Crucially, MCNet employs a novel frequency-aware consistency adversarial loss to explicitly guide the learning across different frequency bands. MCNet effectively integrates MP and CF, enhancing the detection of high-frequency details and low-frequency structures, thereby alleviating challenges posed by spectral bias. We evaluate the proposed method on complex scene segmentation tasks, including camouflaged object detection and dichotomous image segmentation. Through extensive comparisons with 31 existing methods across 8 benchmark datasets, we demonstrate the superiority of the proposed method.

As public indoor activities become increasingly diversified, the demand for accurate indoor positioning has grown significantly. To address the limitations of GPS in indoor environments, this article proposes a lightweight pedestrian dead reckoning (PDR) method based on inertial sensors embedded in mobile devices. The approach integrates three key components: 1) a device placement pattern recognition algorithm that classifies stationary, stable, and swing modes using acceleration and angular velocity variance; 2) an adaptive step detection algorithm based on alternating peak-valley rules and temporal thresholds, enabling robust step identification across various usage modes; and 3) a placement-aware stride length estimation model derived from an improved inverted pendulum framework, which dynamically adjusts stride estimates based on walking frequency and device mode. Experimental results demonstrate that the proposed method achieves accurate step detection and positioning performance comparable to civilian GPS systems, while requiring no external infrastructure or prior training. This work offers a practical solution for real-time pedestrian tracking in indoor environments using mobile devices.

With the rapid increase in the number of unmanned aerial vehicles (UAVs), the air traffic environment has become increasingly complex. The ability of UAVs to detect other airborne objects and estimate their azimuth is critical for preventing mid-air collisions and ensuring flight safety. This study proposes a method for estimating the azimuth of airborne objects by small fixed-wing UAVs based on an onboard electrostatic sensing model, and verifies its effectiveness through experiments. An onboard electrostatic sensing model based on electrostatic field theory is established, based on which an azimuth estimation method is developed and validated through simulations. A four-sensor electrostatic sensing platform with a compact electrode layout is designed and implemented, featuring an inter-electrode spacing of approximately 1 mm and a nominal power consumption of approximately 0.8 W. Indoor experiments are conducted to evaluate the method in terms of estimation accuracy, stability, applicability, and robustness. The azimuth estimation error remains within +/- 10 degrees under various experimental conditions, without significant degradation under temperature and humidity variations, visibility changes, or electromagnetic interference. Equivalent experiments further demonstrate reliable azimuth estimation for a small quadcopter UAV at distances of up to approximately 25 m. The method exhibits low power consumption and stable performance under environmental interference. It eliminates the reliance on large electrode spacing and is suitable for airborne platforms with stringent size and payload constraints. This approach offers new insights into UAV sensing technology and contributes to enhanced flight safety.

Accurate fall risk prediction is crucial for early intervention and prevention, effectively reducing the incidence of falls and the associated harm. This paper proposes a non-contact gait detection and fall risk prediction method based on the human electrostatic field and Stacking ensemble learning algorithm. A theoretical model for gait detection based on the human electrostatic field is established, and an experimental scheme is designed. The electrostatic gait measurement system is used to collect electrostatic gait signals from healthy young individuals, healthy elderly individuals, and elderly individuals with a history of falls. Gait features, including 28-dimensional quantifiable characteristics, are proposed for evaluating human balance and motor abilities, covering four aspects: gait time parameters, gait symmetry based on ratios and signal similarity, gait stability based on the maximum Lyapunov exponent and entropy information, and gait time parameter variability. A hybrid feature reduction method based on Particle Swarm Optimization (PSO) is used to obtain the optimal feature subset. Fall risk prediction models based on single classifiers (DT, SVM, KNN, and NB) are constructed using both the original feature set and the optimal feature subset. The single classifier based on the optimal feature subset achieves better classification performance. Furthermore, a Stacking ensemble learning model using LightGBM as the meta-learner is developed, achieving an accuracy of 97.78%. This study provides a novel approach for fall risk prediction that can predict the likelihood of falls and reduce the probability of their occurrence.

Gas-solid friction can induce interfacial charge transfer, and the harvested triboelectric charges can, in principle, power distributed sensor networks on aircraft surfaces. However, the charging behavior of gas-solid triboelectrification under varying flight conditions remains insufficiently investigated. In this study, a theoretical estimation model for gas-solid triboelectric charging on moving surfaces is developed, and its key influencing factors are analyzed. Based on the Maxwell-Boltzmann velocity distribution, the model demonstrates that, under constant ambient temperature, the generated charge is jointly governed by the collision frequency between gas molecules and the solid surface and by the effective frictional area. The collision frequency, in turn, depends on the object's velocity and the ambient pressure. A fluid-solid coupled simulation conducted in ANSYS Workbench provides preliminary verification of the model's validity. Additionally, first-principles calculations confirmed that aluminum tends to lose electrons, exhibiting a positive charge during gas-solid triboelectric charging. Furthermore, a non-contact experimental system employing double through-type Faraday cups is designed to measure the charge-transfer efficiency predicted by the Maxwell-Boltzmann-based estimation under controlled variations in motion velocity, effective friction area, and ambient pressure. The experimental results confirm the correctness of the proposed estimation model. This work provides a theoretical foundation for analyzing the feasibility of powering distributed sensors on aircraft skins through gas-solid triboelectric charging.

The interaction between the vehicle's electrostatic accumulation and the plasma in hypersonic flight environments has received little attention. To investigate this mechanism, an inductively coupled plasma (ICP) system was employed to generate plasmas with plasma densities up to 1017m-3 and electron temperatures of several eV, replicating key characteristics of hypersonic plasmas. Using this platform, the interactions of metallic and dielectric targets with the plasma were systematically examined. The results show that positively charged metallic targets significantly enhance local plasma density by up to 180% via electrostatic attraction. However, negatively charged metallic targets reduced plasma density due to electric field repulsion, followed by a subsequent increase attributed to secondary electron emission from positive ion bombardment. The motion of charged particles induced measurable currents at the milliampere level in the circuit. In contrast, dielectric targets accumulate surface charge that effectively suppresses electric-field penetration and produces negligible influence on the adjacent plasma. These observations illustrate how electrostatic potentials modify the dynamics of charged-particle near different materials and provide useful guidance for electrostatic protection and flight safety considerations in hypersonic vehicles.

Recently, millimeter-wave radar has been increasingly applied to Human Activity Recognition (HAR) due to its inherent privacy protection and resistance to low-light interference. However, methods relying solely on Time-Doppler Map (TDM) suffer from severe performance degradation as the subject's direction deviates from the radar's line-of-sight. To address this, we propose DIHARNet, a dual-stream framework designed for direction-insensitive recognition, which concurrently exploits the intrinsic micro-motion dynamics of TDM and the geometric robustness of Point Cloud Data (PCD). Specifically, we circumvent the prohibitive costs of dense volumetric learning by factorizing high-dimensional spatial data into orthogonal 2D projections. Furthermore, to rectify directional ambiguity, we introduce a geometry-guided recalibration mechanism that leverages direction-invariant spatial priors to dynamically calibrate Doppler features, ensuring discriminative representations across representative directions. To fill the critical gap of dual-domain benchmarks, we construct MDHA, a novel dataset for direction-insensitive radar-based HAR. Extensive experiments demonstrate that DIHARNet achieves superior performance, reaching an overall accuracy of 93.46% while maintaining stable recognition across representative aspect-angle variations, highlighting its effectiveness for robust direction insensitive radar-based HAR.

Lightning is a long-distance gas discharge process that occurs in atmosphere, and the direct force and thermal load impact accompanying lightning strike aircraft is one of the main causes of many lightning strike accidents. As the development of aircraft shell from metal material to lightweight composite material, lightning strikes can cause extra damage to lightweight material aircraft shell. The laboratory lightning strike test is essential for assessing the effectiveness of lighshows that after the gap breakdowntning strike protection. This paper carried out a study on three methods for measuring the mechanical load of laboratory lightning strikes. In the experiment, a pulsed-power source and a 6-μF pulse capacitor provide a high-power and microsecond timescale pulsed current to the load. The range of charging voltages involved in the experiment ranged from 9 kV to 25.8 kV. Two different load structures are established, which are essentially rod-plate loads, and three test methods of mechanical effects are designed under 40-kA-level pulsed current. Experimental results indicate that three measurement methods can effectively characterize the mechanical effects of laboratory lightning strikes. In these methods, the maximum impulse measured by the metal sheets upward is 3.08 kg·m/s, the force measured by the piezoelectric sensor can reach 850 N, and the single pendulum motion obtained the maximum impulse which is 0.0145 kg·m/s.

The charge quantity is a fundamental physical parameter that reflects the electrical state of an object. Accurately estimating the charge of an object facilitates the assessment of electrostatic discharge risks and aids in preventing accidents. Measuring the charge of a moving object has long posed a technical challenge in this field. This paper proposes a non-contact method for estimating the charge of a moving object by utilizing the electrostatic signals generated by the object’s motion and its motion data. First, a non-contact charge measurement model based on a mutual capacitance matrix was developed using the image charge method in electrostatics. The accuracy of the model was verified through simulations of the charge on the sensing electrode. Next, a correction method for charge calculation was further proposed to reduce measurement errors caused by parasitic capacitance from the experimental setup. Finally, a verification experiment was conducted, wherein an electrometer measured the charge of the object in a stationary state, providing a reference to validate the proposed method. The experimental results demonstrated a strong correlation (r > 0.96) and consistency (within the 95% confidence interval) between the measured and reference values across various conditions. The absolute error of the measurements was within ±1 nC (mean ± standard deviation: -0.04 ± 0.4 nC), with a relative error of approximately ±10%. This study contributes to the prevention of electrostatic discharge accidents involving moving objects and presents novel insights and technological approaches for electrostatic detection.

The ability to synthesize nanoparticles (NPs) with desirable structural/compositional properties has been the driving goal of functional material applications. As a "one-step" evaporation-condensation method for NPs production, the electrical explosion can achieve ultrafast heating/quenching rates (dT/dt similar to 10(10) K/s) of current-carrying metals. This study proposes a spatial filter to control explosion energy release and product quenching procedure. A supersonic plasma jet (>2 Mach) erupts through the small hole of the filter, forming a high-pressure homogenizer structure, enhancing the diffusion and cooling down of explosion products. Microscopic characterization indicates NPs size decreases from 46.3 +/- 16.7 to 27.7 +/- 7.7 nm. Structural study has yielded two assembly routes of NPs: the first involves the top-down approach (Cu particles >100 nm), while the second concerns the bottom-up way (Cu and CuxO NPs <100 nm). The improved performance is attributed to the multi-stage control of explosion products, resulting in less instability development and a more reasonable evaporation-condensation process.

This paper explores the aerodynamic characteristics of the He plasma jet under nanosecond pulse needle-ring discharge excitation (dielectric-barrier), and the influence of typical pulse parameters on the flow field is preliminarily ascertained. Schlieren photography is employed to visualize the flow field of the jet, and dynamic processes from plasma generation to post-discharge evolution are recorded in detail by a high-speed camera. The visualized images reveal a transition from a laminar to a turbulent region in the flow field outside the tube following discharge excitation. The turning point (turbulent vortex) propagates at a speed of m s-1. The results show that a higher driving voltage leads to an earlier onset of the turbulent vortex and further shortens the laminar zone length, with a maximum reduction of 36%. The turbulent vortex dynamics gradually diminish as the frequency increases, resulting in a maximum 30% reduction in the laminar zone length after the cumulative effect of several cycles. Also, the impact of the pulse width parameter on the flow field is minimal. By observing the changes in the distribution of gas temperature under different discharge parameters, it is concluded that the thermal effect is the main factor influencing the kinetic process of the flow field.

Electrical explosion is a physical process driven by electric current that comes with a conductor to be heated up, undergo phase transitions in form of explosion. It can provide extreme conditions of instantaneous high temperature (>10000 K) and pressure (GPa) in a very short period (<20 μs), accompanied by intense shock waves and light radiation. These characteristics make it a unique but effective methodology for designing and synthesizing composite materials. In this work, polyethylene (PE) was introduced into the electrical explosion process. Therein, a 20-kA current pulse was applied to drive electrical explosion, resulting in a special composites with metal nanoparticles (NPs) adhering to thin PE sheets. The morphology of the material mainly shows small sized metal nanoparticles attached to the large sized polymer surface. A small amount of core-shell structured composites were also observed in the SEM images. Statistical analysis revealed that the shock wave produced in free space during the explosion propagated at a speed of 775 m/s, which enhanced the mixing of different components in the explosion products. This method enabled the rapid modification of polyethylene materials, leading to the preparation of polymer materials attaching with copper and aluminum nanoparticles. Compared with different kinds of raw material, copper foil and copper wire, the former one forms a tighter structure of nanoparticles, and a more uniform distribution. Some of the nanoparticles are even able to be embedded into carbon material. These composites are expected to exhibit diverse properties, offering potential for advanced applications.

[Purposes] To systematically explore the evolutionary patterns of research on specific technical movements in the field of sports biomechanics, a bibliometric approach combined with CiteSpace visualization analysis technology is employed to conduct a knowledge graph analysis of relevant literature from the CNKI database. [Methods] CiteSpace software was utilized to analyze the temporal evolution, spatial distribution, and topic clustering of literature published between 2000 and 2024. [Results] The results indicate that: (1) The temporal dimension exhibites a three-stage characteristic: a slow growth phase from 2000 to 2009 (with an average of fewer than 10 papers published annually), a fluctuating growth phase from 2010 to 2016, and a stable development phase since 2017; (2) The institutional distribution demonstrates significant clustering features, primarily concentrates in sports-specialized universities and economically developed regions; (3) In terms of topic evolution, three major research clusters are identified through keyword co-occurrence and clustering analyses: a fundamental theory cluster (kinematic parameters, dynamic modeling), a methodology cluster (three-dimensional motion analysis, EMG signal processing), and an applied research cluster (technique optimization, injury prevention); (4) The time-zone map reveals a shift in the research paradigm from single-discipline validation to multidisciplinary integration, entering a technology-driven innovation stage after 2021. [Conclusions] According to this study, notable issues such as a low density of inter-institutional collaboration networks and an imbalance in regional research capabilities are found. It is suggested that future efforts should strengthen the cross-integration of biomechanics and artificial intelligence and establish a standard system for multimodal motion analyses. This study provides a quantitative basis for grasping the development trends in this field and optimizing the allocation of scientific research resources.

Plasma instability and discharge randomness is the main challenge in achieving precision nanoparticle synthesis via electrical explosion. This study presents a split-wire array load configuration strategy to enhance the product uniformity. Through controlled experiments with 500 J discharge energy on 25 mg metal conductors, synchronized high-speed imaging and electrophysical diagnostics reveal that increased specific surface area (achieved via wire splitting) effectively suppresses electro-thermal/plasma instabilities and secondary breakdown, compared to conventional single-wire configuration. Systematic analysis demonstrates at least a 60 % reduction in nanoparticle mean diameter (from 55.10 nm to 20.60 nm) and improved polydispersity index (similar to 45.36 % reduction) with optimized load topology. The spatial-temporal evolution of phase transitions and ionization processes is quantitatively correlated with load geometry, establishing surface-area-dominated plasma dynamics as a critical mechanism for nanoparticle quality control. These findings provide a transformative approach to overcoming intrinsic limitations in electrical explosion synthesis through innovative load design.

During flight operations, unmanned aerial vehicles (UAVs) accumulate significant electrostatic charges, which may lead to electrostatic discharge events, potentially damaging onboard electronic systems and compromising flight safety. Accurate measurement of these accumulated charges is crucial for optimizing anti-static protection measures. First, a refined physical model for UAV charge measurement is proposed, which incorporates the influence of parasitic capacitance. The model utilizes the electrode's spatial sensitivity to establish a quantitative relationship between the induced charge on the electrode and the UAV electrostatic charge. Then, a noncontact method for measuring UAV charge is proposed, utilizing the spatial sensitivity at the UAV's position and the induced charge on the electrode to estimate the UAV charge. Further, experimental validation was conducted using electrometer measurements as the reference values. The results demonstrate strong correlation and consistency between measured and reference values, with measurement error maintained within +/- 10 %. Finally, the proposed method was applied to measure the UAV charge, and its experiment result ranges from 9 to 18 nC. The proposed method accurately measures the UAV charge during flight and can be used to evaluate electrostatic discharge risk, providing a foundation for the anti-static design of UAVs and enhancing flight safety.

Falls are one of the most serious health risks faced by older adults worldwide, and they can have a significant impact on their physical and mental well-being as well as their quality of life. Detecting falls promptly and accurately and providing assistance can effectively reduce the harm caused by falls to older adults. This paper proposed a noncontact fall detection method based on the human electrostatic field and a VMD-ECANet framework. An electrostatic measurement system was used to measure the electrostatic signals of four types of falling postures and five types of daily actions. The signals were randomly divided in proportion and by individuals to construct a training set and test set. A fall detection model based on the VMD-ECA network was proposed that decomposes electrostatic signals into modal component signals using the variational mode decomposition (VMD) technique. These signals were then fed into a multichannel convolutional neural network for feature extraction. Information fusion was achieved through the efficient channel attention network (ECANet) module. Finally, the extracted features were input into a classifier to obtain the output results. The constructed model achieved an accuracy of 96.44%. The proposed fall detection solution has several advantages, including being noncontact, cost-effective, and privacy friendly. It is suitable for detecting indoor falls by older individuals living alone and helps to reduce the harm caused by falls.

Gait stability is an important indicator of human health and physical ability, and it is of great significance for early detection, diagnosis, and rehabilitation of diseases. This article proposes the first comprehensive and in-depth method for quantitatively evaluating gait stability using electrostatic gait signals (EGSs). A quantitative evaluation of gait stability was conducted on the EGSs of ten healthy subjects (HSs) and ten hemiplegic patients (HPs) from three perspectives: variability of gait phase temporal parameters, nonlinear local dynamic stability, and energy frequency band distribution differences. The coefficient of variation [CV( x )] reflects the degree of variation of each temporal parameter. Local dynamic stability is characterized by using the short-term largest Lyapunov exponent ( lambda & lowast;S ) to reflect the sensitivity to local perturbations and to characterize gait stability. Power spectral entropy (PSE) is used to quantify the complexity and instability of the signal in the frequency domain to characterize gait stability. The results showed that the CV( x ), lambda(& lowast;)(S) , and PSE were all significantly greater in HPs than in HSs. The Mann-Whitney test was used to perform significance testing on each indicator between the HS and HP groups. The results showed that, except for the variability of the affected side support phase [CV( T-LHI )], the variability of the other 13 gait temporal parameters, short-term largest Lyapunov exponent ( lambda(& lowast;)(S) ), and PSE differed significantly between the groups ( p<0.05 ). This research explores a feasible technical approach to using EGSs for gait analysis and quantitatively assessing the gait stability of subjects.

Pulsed discharge in the vicinity of a multi-phase interface, where a discontinuity of physical properties exists, can be a joint problem of both electro- and thermo-physics. This study shows a comprehensive analysis of electric breakdown across an air-water interface and its successive multi-physical effects. The scenario is constructed via a pair of pin electrodes positioned on both sides of the interface, and the transient discharge is analyzed using high-speed backlight photography synchronized with electrical and optical diagnostics. It is observed that the corona/streamer develops from either side of the pin electrode. Electrostatic instability causes the interface to fluctuate and a water column to form above the interface. By increasing the applied voltage, discharge evolves from "dielectric barrier" mode (pin to interface) to "through breakdown" mode (pin to pin). Once the conductive channel bridges two electrodes, electric power of similar to 40 kW peak and deposited energy of 100 mJ will be injected into the channel and promote the "streamer-spark" transition, resulting in a crown-like splash (100 m s(-1)) near the interface and cavity formation. As the quenching of diffused plasmas, the over-expanded splash (5 mm in diameter) would be re-compressed by the ambient air. Particularly, the shrinkage of the thin water film of the splash can reach a 20 mm jet near the axis and develop Rayleigh-Taylor instability, along with the formation of micro-jets eruption during the convergent collision. More sophisticated interactions will appear at higher repetitive frequency (>100 Hz), where the perturbation caused by one pulse will influence the next, namely the "memory" effect. Furthermore, periodic loading on the interface effectively changes the cavity characteristics, showing an attractive prospect in fluid control applications.