Search Results (2,200)

We present E-CAHORS—a compact mid-infrared open-path diode-laser spectrometer designed for the simultaneous measurement of carbon dioxide, methane, and water vapor concentrations in the near-surface atmospheric layer. These measurements, combined with simultaneous data from a three-dimensional anemometer, can be used to determine fluxes using the eddy-covariance method. The instrument utilizes two interband cascade lasers operating at 2.78 µm and 3.24 µm within a novel four-pass M-shaped optical cell, which provides high signal power and long-term field operation without requiring active air sampling. Two detection techniques—tunable diode laser absorption spectroscopy (TDLAS) and a simplified wavelength modulation spectroscopy (sWMS)—were implemented and evaluated. Laboratory calibration demonstrated linear responses for all gases (R² ≈ 0.999) and detection precisions at 10 Hz of 311 ppb for CO₂, 8.87 ppb for CH₄, and 788 ppb for H₂O. Field tests conducted at a grassland site near Moscow showed strong correlations (R = 0.91 for CO₂ and H₂O, R = 0.74 for CH₄) with commercial LI-COR LI-7200 and LI-7700 analyzers. The TDLAS mode demonstrated lower noise and greater stability under outdoor conditions, while sWMS provided baseline-free spectra but was more sensitive to power fluctuations. E-CAHORS combines high precision, multi-species sensing capability with low power consumption (10 W) and a compact design (4.2 kg). Full article

Time series forecasting plays a critical role in a wide range of applications, including energy load forecasting, traffic flow management, weather prediction, and vision-based state prediction for autonomous driving. In the context of autonomous vehicles, accurate forecasting of sequential visual information—such as traffic participant trajectories, road condition variations, and obstacle motion trends perceived by onboard sensors—is a fundamental prerequisite for safe and reliable decision-making. To overcome the limitations of existing long-term time series forecasting models, particularly their insufficient capability in temporal feature extraction, this paper proposes a Local–Global Adaptive Transformer (LGAT) for long-term time series forecasting. The proposed model incorporates three key innovations: (1) a period-aware positional encoding mechanism that embeds intrinsic periodic patterns of time series into positional representations and adaptively adjusts encoding parameters according to data-specific periodicity; (2) a temporal feature enhancement module based on gated convolution, which effectively suppresses noise in raw inputs while emphasizing discriminative temporal characteristics; and (3) a local–global adaptive attention layer that combines sliding window–based local attention with importance-aware global attention to simultaneously capture short-term local variations and long-term global dependencies. Experimental results on five public benchmark datasets demonstrate that LGAT consistently outperforms most baseline models, indicating its strong potential for time series forecasting applications in autonomous driving scenarios. Full article

Current management strategies for sorghum aphids heavily rely on indiscriminate chemical application, leading to severe environmental consequences and impacting food safety. While precision spraying offers a viable remediation for pesticide overuse, its effectiveness depends on accurately locating and classifying pests. To address the critical challenge of segmenting small, swarming aphids in complex field environments, we propose FESW-UNet, a dual-domain attention network that integrates Fourier-enhanced attention, spatial attention, and wavelet-based downsampling into a UNet backbone. We introduce an efficient multi-scale attention (EMA) module between the encoder and decoder to enhance global context perception, enabling the model to capture more accurate relationships between global and local features in the field. In the feature extraction stage, we embed a simple attention module (SimAM) to target key infestation regions while suppressing background noise, thereby enhancing pixel-level discrimination. Furthermore, we replace conventional downsampling with Haar wavelet downsampling (HWD) to reduce resolution while preserving structural edge details. Finally, a Fourier-enhanced attention module (FEAM) is added to the skip-connection layers. By using complex-valued weights to regulate frequency-domain features, FEAM effectively fuses global low-frequency structures with local high-frequency details, thereby enhancing feature representation diversity. Experiments on the Aphid Cluster Segmentation dataset demonstrate that FESW-UNet outperforms other models, achieving an mIoU of 68.76%, mPA of 78.19%, and mF1 of 79.01%. The model also demonstrated strong adaptability on the AphidSeg-Sorghum dataset, achieving an mIoU of 81.22%, mPA of 87.97%, and mF1 of 88.60%. The proposed method offers an efficient and feasible technical solution for monitoring and controlling sorghum aphids through image segmentation, demonstrating broad application potential. Full article

Diabetic retinopathy (DR) is a major cause of preventable vision loss, and its early detection is essential for timely clinical intervention. However, existing deep learning-based DR recognition methods still face two fundamental challenges: substantial lesion-scale variability and significant background noise in retinal fundus images. To address these issues, we propose a lightweight framework named Multi-Scale Soft-Thresholding Attention Network (MSA-Net). The model integrates three components: (1) parallel multi-scale convolutional branches to capture lesions of different spatial sizes; (2) a soft-thresholding attention module to suppress noise-dominated responses; and (3) hierarchical feature fusion to enhance cross-layer representation consistency. A squeeze-and-excitation module is further incorporated for channel recalibration. On the APTOS 2019 dataset, MSA-Net achieves 97.54% accuracy and 0.991 AUC-ROC for binary DR recognition. We further evaluate five-class DR grading on APTOS2019 with 5-fold stratified cross-validation, achieving 82.71 ± 1.25% accuracy and 0.8937 ± 0.0142 QWK, indicating stable performance for ordinal severity classification. With only 4.54 M parameters, MSA-Net remains lightweight and suitable for deployment in resource-constrained DR screening environments. Full article

Homomorphic encryption provides a powerful cryptographic solution for privacy-preserving deep neural network inference, enabling computation on encrypted data. However, the practical application of homomorphic encryption is fundamentally constrained by the noise budget, a core component of homomorphic encryption schemes. The substantial multiplicative depth of modern deep neural networks rapidly consumes this budget, necessitating frequent, computationally expensive bootstrapping operations to refresh the noise. This bootstrapping process has emerged as the primary performance bottleneck. Current noise management strategies are predominantly static, triggering bootstrapping at pre-defined, fixed intervals. This approach is sub-optimal for deep, complex architectures, leading to excessive computational overhead and potential accuracy degradation due to cumulative precision loss. To address this challenge, we propose a Deep Network-aware Adaptive Noise-budget Management mechanism, a novel mechanism that formulates noise budget allocation as a sequential decision problem optimized via reinforcement learning. The core of the proposed mechanism comprises two components. First, we construct a layer-aware noise consumption prediction model to accurately estimate the heterogeneous computational costs and noise accumulation across different network layers. Second, we design a Deep Q-Network-driven optimization algorithm. This Deep Q-Network agent is trained to derive a globally optimal policy, dynamically determining the optimal timing and network location for executing bootstrapping operations, based on the real-time output of the noise predictor and the current network state. This approach shifts from a static, pre-defined strategy to an adaptive, globally optimized one. Experimental validation on several typical deep neural network architectures demonstrates that the proposed mechanism significantly outperforms state-of-the-art fixed strategies, markedly reducing redundant bootstrapping overhead while maintaining model performance. Full article

This work presents a comparative study of voice activity detection in noise using simple acoustic features and relatively compact recurrent models within a controlled MATLAB-based framework. For each utterance, 9 baseline spectral-plus-periodicity features, MFCCs, and FBANKs are extracted and passed to several lightweight BiLSTM-based networks, either alone or preceded by a 1D CNN layer. The main experiments are carried out at a fixed SNR to separate the influence of the network structure and the feature type, and an additional series with four SNR levels is used to assess whether the same performance trends hold when the SNR varies. The results show that adding a compact CNN front-end before the BiLSTM consistently improves detection scores, that MFCCs generally outperform the baseline spectral–periodicity features and often give better recall/F1 than FBANKs for the considered lightweight models, and that $\mathrm{CNN}(3,32)$+BiLSTM with 13-dimensional MFCCs offers a favorable trade-off between accuracy, robustness across SNRs, and model size. Because all conditions share a single MATLAB implementation with fixed noise types, SNR values, and evaluation metrics, this work is positioned as a benchmark and practical guideline publication for noise-robust, resource-constrained VAD, rather than as a proposal of a completely new deep-learning architecture. Full article

Empirical debates about a “crisis of trust” highlight long-lived pockets of high trust and deep distrust in institutions, as well as abrupt, shock-induced shifts between the two. We propose a probabilistic model in which such phenomena emerge endogenously from social learning on hierarchical networks. Starting from a discrete model on a directed acyclic graph, where each agent makes a binary adoption decision about a single assertion, we derive an effective influence kernel that maps individual priors to stationary adoption probabilities. A continuum limit along hierarchical depth yields a degenerate, non-conservative logistic–diffusion equation for the adoption probability $u(x,t)$, in which diffusion is modulated by $(1-u)$ and increases the integral of u rather than preserving it. To account for micro-level uncertainty, we perturb these dynamics by multiplicative Stratonovich noise with amplitude proportional to $u(1-u)$, strongest in internally polarised layers and vanishing at consensus. At the level of a single depth layer, Stratonovich–Itô conversion and Fokker–Planck analysis show that the noise induces an effective double-well potential with two robust stochastic phases, $u\approx 0$ and $u\approx 1$, corresponding to persistent distrust and trust. Coupled along depth, this local bistability and degenerate diffusion generate extended domains of trust and distrust separated by fronts, as well as rare, Kramers-type transitions between them. We also formulate the associated stochastic partial differential equation in Martin–Siggia–Rose–Janssen–De Dominicis form, providing a field-theoretic basis for future large-deviation and data-informed analyses of trust landscapes in hierarchical societies. Full article

Detecting surface defects in forgings is crucial for ensuring the reliability of automotive components such as steering knuckles. In fluorescent magnetic particle inspection (FDMPI) images, normal forging surfaces generally exhibit locally symmetric texture patterns, whereas cracks and other flaws appear as locally asymmetric regions. Traditional FDMPI inspection relies on manual visual judgement, which is inefficient and error-prone. This paper introduces SDDNet, a symmetry-aware deep learning model for surface defect detection in FDMPI images. A dedicated FDMPI dataset is constructed and further expanded using a denoising diffusion probabilistic model (DDPM) to improve training robustness. To better separate symmetric background textures from asymmetric defect cues, SDDNet integrates a UPerNet-based segmentation layer for background suppression and a Scale-Variant Inception Module (SVIM) within an RTMDet framework for multi-scale feature extraction. Experiments show that SDDNet effectively suppresses background noise and significantly improves detection accuracy, achieving a mean average precision (mAP) of 45.5% on the FDMPI dataset, 19% higher than the baseline, and 71.5% mAP on the NEU-DET dataset, outperforming existing methods by up to 8.1%. Full article

The rapid growth of wireless devices, the expansion of the Internet of Things, and the aggregate demand for Ultra-Reliable Low-Latency communications (URLLC) are driving the improvement of next-generation wireless systems. One promising emerging technology in this area is cell-free massive Multiple Input Multiple Output (maMIMO) networks. The distributed nature of Access Points presents unique security challenges that must be addressed to unlock their full potential. This paper studies the key security concerns in Cell Free Massive MIMO (CFMM) networks, including eavesdropping, Denial-of-Service attacks, jamming, pilot contamination, and methods for enhancing Physical Layer Security (PLS). We also provide an overview of security solutions specifically designed for CFMM networks and introduce a case study of a Reconfigurable Intelligent Surface (RIS)-aided secure scheme that jointly optimizes the RIS phase shifts with the artificial noise (AN) covariance under power constraints. The non-convex optimization problem is solved via the block coordinate descent (BCD) alternating optimization scheme. The combined RIS, AN, and beamforming configuration achieves a balanced trade-off between security and energy performance, resulting in moderate improvements over the individual schemes. Full article

This work proposes an analysis of the capability of three deep learning models—the feedforward neural network (FFNN), long short-term memory (LSTM) network, and physics-informed neural network (PINN)—to identify the parameters of a flexible fixed-wing aircraft using in-flight data. These neural networks, composed of multiple hidden layers, are evaluated for their ability to perform system identification and to capture the nonlinear and dynamic behavior of the aircraft. The FNN and LSTM models are compared to assess the impact of temporal dependency learning on parameter estimation, while the PINN integrates prior knowledge of the system’s governing of ordinary differential equations (ODEs) to enhance physical consistency in the identification process. The objective is to exploit the generalization capability of neural network-based models while preserving the accurate estimation of the physical parameters that characterize the analyzed system. The neural networks are evaluated for their ability to perform system identification and capture the nonlinear behavior of the aircraft. The results show that the FFNN achieved the best overall performance, with average Theil’s inequality coefficient (TIC) values of 0.162 during training and 0.386 during testing, efficiently modeling the input-output relationships but tending to fit high-frequency measurement noise. The LSTM network demonstrated superior noise robustness due to its temporal filtering capability, producing smoother predictions with average TIC values of 0.398 (training) and 0.408 (testing), albeit with some amplitude underestimation. The PINN, while successfully integrating physical constraints through pretraining with target aerodynamic derivatives, showed more complex convergence, with average TIC values of 0.243 (training) and 0.475 (testing), and its estimated aerodynamic coefficients differed significantly from the conventional values. All three architectures effectively captured the coupled rigid-body and flexible dynamics when trained with distributed wing sensor data, demonstrating that neural network-based approaches can model aeroelastic phenomena without requiring explicit high-fidelity flexible-body models. This study provides a comparative framework for selecting appropriate neural network architectures based on the specific requirements of aircraft system identification tasks. Full article

The coronavirus is highly contagious, making rapid early diagnosis essential. Although deep learning-based diagnostic methods using CT or X-ray images have advanced significantly, they still face limitations in cost, processing time, and radiation exposure. In addition, for the possibility of real-time COVID-19 diagnosis, model lightweighting is required. This study proposes a lightweight deep learning model for COVID-19 diagnosis based on fluorescence images and demonstrates its applicability in embedded environments. To prevent data imbalance caused by noise and experimental variations, images were preprocessed using Gray Scale conversion, CLAHE, and Z-Score normalization to equalize brightness values. Among the tested architectures—VGG, ResNet, DenseNet, and EfficientNet—ResNet152 and VGG13 achieved the highest accuracies of 97.25% and 93.58%, respectively, and were selected for lightweighting. Layer-wise importance was calculated using an imprinting-based method, and less important layers were pruned. The pruned VGG13 maintained its accuracy while reducing model size by 18.9 MB and parameters by 4.2 M. ResNet152 (Prune 39) improved accuracy by 1% while reducing size by 161.5 MB and parameters by 40.22 M. The optimized model achieved 129.97 ms, corresponding to 7.69 frames per second (FPS) on an NPU(Furiosa AI Warboy), proving real-time COVID-19 diagnosis is feasible even on low-power edge devices. Full article

This study investigated the noise reduction performance of porous asphalt concrete (PAC) pavement under tire–pavement coupling conditions, addressing the limitations of field measurements and laboratory testing. First, tire excitation amplitude parameters were determined based on vibrational contact operational scenarios. Then, finite element simulations were conducted to systematically analyzing the tire–pavement coupling noise characteristics of PAC pavement. The results indicate that PAC pavement effectively reduces the air pumping noise due to its highly porous internal structure, leading to significant noise attenuation. Furthermore, the study examined the key factors influencing the tire–pavement coupling noise in PAC pavement. When maintaining constant vehicle parameters (300 kg load, 60 km/h speed), pavement thickness became the critical noise-control variable, achieving minimum vibration at 6 cm surface layer thickness. Additionally, tire tread depth (5–17 mm) and mold release angle (0–30°) had a more pronounced impact on the air pumping noise compared to groove width (20–60 mm). Increasing the mold release angle and reducing tread depth effectively mitigated the air pumping noise. However, the tire–pavement coupling noise in PAC pavement increased considerably with increasing vehicle speed and load. Particularly, as the speed increased from 30 km/h to 60 km/h, the growth of the air pumping noise was most pronounced, revealing an acoustic transition of tire–pavement coupling noise from vibration-dominated to air-pumping-dominated mechanisms. Full article

Near-infrared spectroscopy (NIRS) is widely used in food, pharmaceutical, and agricultural analyses but is highly susceptible to noise. To address this, we propose BiLSTM-FuseNet, a denoising framework that combines temporal modeling and explicit noise estimation. It uses stacked Bidirectional Long Short-Term Memory (BiLSTM) layers for global–local spectral learning and an MLP branch to predict and subtract noise. Evaluated on the Tablet and AnHui soil datasets with various synthetic noise types, the model outperformed the conventional methods, achieving an RMSE of 0.024 and R² of 0.68 under mixed noise. The downstream regression improved the tablet weight prediction R² from 0.079 to 0.218. These findings demonstrate the robustness of BiLSTM-FuseNet and its clear advantages for practical downstream NIR applications. Full article

Weak-signal fluorescence channels (e.g., 4′,6-diamidino-2-phenylindole (DAPI)) often fail to provide reliable structural details due to low signal-to-noise ratio (SNR) and insufficient high-frequency information, limiting the ability of single-channel super-resolution methods to restore edge continuity and texture. This study proposes a multi-channel guided super-resolution method based on SwinIR, utilizing the high-SNR fluorescein isothiocyanate (FITC) channel as a structural reference. Dual-channel adaptation is implemented at the model input layer, enabling the window attention mechanism to fuse cross-channel correlation information and enhance the structural recovery capability of weak-signal channels. To address the loss of high-frequency information in weak-signal imaging, we introduce a frequency-domain consistency loss: this mechanism constrains spectral consistency between the predicted and true images in the Fourier domain, improving the clarity of fine-structure reconstruction. Experimental results on the DAPI channel demonstrate significant improvements: PSNR increases from 27.05 dB to 44.98 dB, and SSIM rises from 0.763 to 0.960. Visual analysis indicates that this method restores more continuous nuclear edges and weak textural details while suppressing background noise; frequency-domain results reduce the minimum resolvable feature size from approximately 1.5 μm to 0.8 μm. In summary, multi-channel structural information provides an effective and physically interpretable deep learning approach for super-resolution reconstruction of weak-signal fluorescence images. Full article

This study tackles real-time state estimation for autonomous underwater vehicle (AUV)–mobile docking station (MDS) cooperation over low-bandwidth, high-latency, jitter-dominated acoustic links, with the goal of turning delayed/out-of-sequence measurements (OOSM) into consistent and informative constraints without sacrificing online operation. We propose an integrated scheme centered on a delay-compensated extended Kalman filter (DC-EKF): a ring buffer enables backward updates and forward replay so that OOSM are absorbed strictly at their physical timestamps; a data-driven delay threshold is learned from “effective information gain” combined with normalized estimation error squared (NEES) filtering; and dynamic confidence, derived from innovation statistics, delay, and signal-to-noise ratio (SNR) proxies, scales the measurement noise to adapt fusion weights. Simulations show the learned delay threshold converges to about 6.4 s (final 6.35 s), error spikes are suppressed, and the overall position root-mean-square error (RMSE) is 5.751 m; across the full data stream, 1067 station measurements were accepted and 30 rejected, and the fusion weights shifted smoothly from inertial measurement unit (IMU)-dominant to station-dominant (≈0.16/0.84) over time. On this basis, a cooperative augmented EKF (Co-Aug-EKF) is added as a lightweight upper layer for unified-frame cooperative estimation, further improving relative consistency. The results indicate that the framework reliably maps delayed acoustic measurements into closed-loop useful information, significantly enhancing estimator stability and docking readiness, while remaining practical to deploy and readily extensible. Full article