Search Results (1,189)

Driver fatigue is a primary factor in traffic accidents and poses a serious threat to road safety. To address this issue, this paper proposes a multi-feature fusion fatigue detection method based on an improved YOLOv8 model. First, the method uses an enhanced YOLOv8 model to achieve high-precision face detection. Then, it crops the detected face regions. Next, the lightweight PFLD (Practical Facial Landmark Detector) model performs keypoint detection on the cropped images, extracting 68 facial feature points and calculating key indicators related to fatigue status. These indicators include the eye aspect ratio (EAR), eyelid closure percentage (PERCLOS), mouth aspect ratio (MAR), and head posture ratio (HPR). To mitigate the impact of individual differences on detection accuracy, the paper introduces a novel sliding window model that combines a dynamic threshold adjustment strategy with an exponential weighted moving average (EWMA) algorithm. Based on this framework, blink frequency (BF), yawn frequency (YF), and nod frequency (NF) are calculated to extract time-series behavioral features related to fatigue. Finally, the driver’s fatigue state is determined using a comprehensive fatigue assessment algorithm. Experimental results on the WIDER FACE and YAWDD datasets demonstrate this method’s significant advantages in improving detection accuracy and computational efficiency. By striking a better balance between real-time performance and accuracy, the proposed method shows promise for real-world driving applications. Full article

[Objectives]: By performing genome assembly, annotation, comparative characterization, and phylogenetic analysis on the complete chloroplast genomes of E. tubulosa and E. japonica—two medicinal plants belonging to the genus Euchresta—this study aims to identify their differential genes, thereby providing fundamental research for screening candidate genes as DNA barcodes for species identification and facilitating the conservation of these endangered species. [Methods]: Illumina PE150 sequencing was performed. Chloroplast genomes (plastomes) were assembled and annotated with GetOrganelle/SPAdes. Comparative analyses assessed gene content, IR/LSC/SSC structure, repeat profiles, and codon-usage bias. Using related Fabaceae as references, we conducted mVISTA alignments and sliding-window nucleotide diversity (Pi) analyses to identify candidate DNA barcodes. Phylogenies from whole-plastome sequences were inferred with Maximum Likelihood, Bayesian Inference, and Maximum Parsimony. [Results]: The plastomes measured 153,960 bp (E. japonica) and 150,146 bp (E. tubulosa), with GC contents of 36.30% and 36.20%, respectively, each exhibiting a typical quadripartite structure. IR/SC boundaries were highly conserved without evident expansion or contraction. Repeat statistics were 20/30 palindromic repeats, 57/64 tandem repeats, and 156/159 simple sequence repeats (SSRs) in E. japonica/E. tubulosa, respectively. Leucine was the most frequently encoded amino acid, cysteine the least, and codon usage favored A/U at third positions. Five hypervariable loci—rps19, psbA, trnK, matK, and rps16 (Pi > 0.03)—were identified as candidate DNA barcodes. All trees consistently placed both species within Papilionoideae (Fabaceae) and recovered the closest relationship to Sophora macrocarpa. [Conclusions]: This study provides, for the first time, complete plastomes and candidate barcoding regions for two protected Euchresta species, supplying foundational resources for species identification, resource assessment, and conservation planning. Full article

In this study, we employed functional near-infrared spectroscopy (fNIRS) to explore dynamic functional connectivity (dFC) responses to olfactory stimulation in thirteen healthy control participants and seven patients with vascular dementia (VD). Participants underwent five rest and odor exposure cycles, and dFC was estimated using a sliding window correlation approach. The healthy control group exhibited limited changes, while the VD group exhibited more extensive fluctuations in both oxy- and deoxyhemoglobin dFC across multiple regions during several stimulation periods. Between-group analyses revealed differences, particularly during olfactory stimulation, with moderate to large effect sizes. These preliminary findings suggest that olfactory-evoked dFC may reflect altered brain network dynamics in VD and could potentially serve as a non-invasive, accessible tool to help understand vascular dementia. Full article

This paper proposes a deep reinforcement learning-based predictive control scheme to address cushion pressure prediction and stabilization in hovercraft systems subject to modeling complexity, dynamic instability, and system delay. Notably, this work introduces a long short-term memory (LSTM) network with a temporal sliding window specifically designed for hovercraft cushion pressure forecasting. The model accurately captures the dynamic coupling between fan speed and chamber pressure while explicitly incorporating inherent control lag during airflow transmission. Furthermore, a novel adaptive behavior cloning mechanism is embedded into the twin delayed deep deterministic policy gradient with behavior cloning (TD3-BC) framework, which dynamically balances reinforcement learning (RL) objectives and historical policy constraints through an auto-adjusted weighting coefficient. This design effectively mitigates distribution shift and policy degradation in offline reinforcement learning, ensuring both training stability and performance beyond the behavior policy. By integrating the LSTM prediction model with the adaptive TD3-BC algorithm, a fully data-driven control architecture is established. Finally, simulation results demonstrate that the proposed method achieves high accuracy in cushion pressure tracking, significantly improves motion stability, and extends the operational lifespan of lift fans by reducing rotational speed fluctuations. Full article

In response to the insufficient adaptability of power transformer remaining useful life (RUL) prediction under complex working conditions and the difficulty of multi-scale feature fusion, this study proposes an industrial time series prediction model based on the parallel Transformer–BiGRU–GlobalAttention model. The parallel Transformer encoder captures long-range temporal dependencies, the BiGRU network enhances local sequence associations through bidirectional modeling, the global attention mechanism dynamically weights key temporal features, and cross-attention achieves spatiotemporal feature interaction and fusion. Experiments were conducted based on the public ETT transformer temperature dataset, employing sliding window and piecewise linear label processing techniques, with MAE, MSE, and RMSE as evaluation metrics. The results show that the model achieved excellent predictive performance on the test set, with an MSE of 0.078, MAE of 0.233, and RMSE of 11.13. Compared with traditional LSTM, CNN-BiGRU-Attention, and other methods, the model achieved improvements of 17.2%, 6.0%, and 8.9%, respectively. Ablation experiments verified that the global attention mechanism rationalizes the feature contribution distribution, with the core temporal feature OT having a contribution rate of 0.41. Multiple experiments demonstrated that this method has higher precision compared with other methods. Full article

A prescribed-performance-based sliding mode control method with feed-forward inverse compensation is proposed in this study to improve the micropositioning accuracy and convergence speed of a piezoelectric actuator (PEA). Firstly, the piezo-actuated micropositioning system is described by a Hammerstein structure model, and an inverse Prandtl–Ishlinskii (PI) model was employed to compensate for its hysteresis characteristics. Then, considering modelling errors, inverse compensation errors, and external disturbances, a new prescribed performance function (PPF) with an exponential dynamic decay rate was developed to describe the constrained region of the errors. We then transformed the error into an unconstrained form by constructing a monotonic function, and the sliding variables were obtained by using the transformation error. Based on this, a sliding mode controller with a prescribed performance function (SMC-PPF) was designed to improve the control accuracy of PEAs. Furthermore, we demonstrated that the error can converge to the constrained region and the sliding variables are stable within the switching band. Finally, experiments were conducted to verify the speed and accuracy of the controller. The step-response experiment results indicated that the time taken for SMC-PPC to enter the error window was 8.1 and 2.2 ms faster than that of sliding mode control (SMC) and PID, respectively. The ability of SMC-PPF to improve accuracy was verified using four different reference inputs. These results showed that, for these different inputs, the root mean square error of the SMC-PPF was reduced by over $39.6\%$ and $52.5\%$, compared with the SMC and PID, respectively. Full article

Traditional time-frequency diagnostics for high-speed bearings face an entrenched trade-off between resolution and real-time feasibility. We present a fast Dual-Phase Short-Time Root-MUSIC pipeline that exploits Hankel structure via FFT-accelerated Lanczos bidiagonalization and Sliding-window Singular Value Decomposition to deliver sub-Hz super-resolution under millisecond budgets. Validated on the Politecnico di Torino aerospace dataset (seven fault classes, three severities), fDSTrM detects 150 μm inner-race and rolling-element defects with 98% and 95% probability, respectively, at signal-to-noise ratio down to −3 dB (78% detection), while Short-Time Fourier Transform and Wavelet Packet Decomposition fail under identical settings. Against classical Root-MUSIC, the approach sustains approximately 200 times speedup with less than ${10}^{-11}$ relative frequency error in offline scaling, and achieves 1.85 milliseconds per 4096-sample frame on embedded-class hardware in streaming tests. Subspace order pre-estimation with adaptive correction preserves closely spaced components; Kalman tracking formalizes uncertainty and yields 95% confidence bands. The resulting early warning margin extends maintenance lead-time by 24–72 h under industrial interferences (Gaussian, impulsive, and Variable Frequency Drive harmonics), enabling field-deployable super-resolution previously constrained to offline analysis. Full article

Modern crop breeding requires efficient tools for genetic identity and parentage verification to manage large-scale programs. To address this, we present GIPA (Genomic Identity and Parentage Analysis), a high-performance toolkit designed for these tasks. GIPA integrates key innovations: a sliding-window algorithm enhances accuracy by correcting genotyping errors, an intelligent system classifies samples by heterozygosity to streamline parentage analysis, and an integrated engine generates intuitive chromosome-level heatmaps. We demonstrate its utility in a soybean backcrossing scenario, where it identified a donor line with 98.02% genomic identity to the recipient, providing a strategy to significantly shorten the breeding program. In maize, its parentage module accurately identified the known parents of commercial hybrids with match scores exceeding 97%, validating its use for variety authentication and quality control. By transforming complex SNP data into clear, quantitative, and visual insights, GIPA provides a robust solution that accelerates data-driven decision-making in plant breeding. Full article

Location-based services (LBSs), which are used for navigation, tracking, and mapping across digital devices and social platforms, establish a user’s position and deliver tailored experiences. Collecting and sharing such trajectory datasets with analysts for business purposes raises critical privacy concerns, as both symmetry in recurring behavior mobility patterns and asymmetry in irregular movement mobility patterns in sensitive locations collectively expose highly identifiable information, resulting in re-identification risks, trajectory disclosure, and location inference. In response, several privacy preservation models have been proposed, including k-anonymity, l-diversity, t-closeness, LKC-privacy, differential privacy, and location-based approaches. However, these models still exhibit privacy issues, including sensitive location inference (e.g., hospitals, pawnshops, prisons, safe houses), disclosure from duplicate trajectories revealing sensitive places, and the re-identification of unique locations such as homes, condominiums, and offices. Efforts to address these issues often lead to utility loss and computational complexity. To overcome these limitations, we propose a new ($\xi$, $ϵ$)-privacy model that combines data generalization and suppression with sliding windows and R-Tree structures, where sliding windows partition large trajectory graphs into simplified subgraphs, R-Trees provide hierarchical indexing for spatial generalization, and suppression removes highly identifiable locations. The model addresses both symmetry and asymmetry in mobility patterns by balancing generalization and suppression to protect privacy while maintaining data utility. Symmetry-driven mechanisms that enhance resistance to inference attacks and support data confidentiality, integrity, and availability are core requirements of cryptography and information security. An experimental evaluation on the City80k and Metro100k datasets confirms that the ($\xi$, $ϵ$)-privacy model addresses privacy issues with reduced utility loss and efficient scalability, while validating robustness through relative error across query types in diverse analytical scenarios. The findings provide evidence of the model’s practicality for large-scale location data, confirming its relevance to secure computation, data protection, and information security applications. Full article

To improve the accuracy and stability of anomaly detection in lithium-ion batteries for electric bicycles, in this study, we propose a hybrid deep learning model that integrates a convolutional neural network (CNN), long short-term memory (LSTM) network, and attention mechanism to extract local temporal features, capture long-term dependencies, and adaptively focus on key time segments around anomaly occurrences, respectively, thereby achieving a balance between local and global feature modeling. In terms of data preprocessing, separate feature sets are constructed for charging and discharging conditions, and sliding windows combined with min–max normalization are applied to generate model inputs. The model was trained and validated on large-scale real-world battery operation data. The experimental results demonstrate that the proposed method achieves high detection accuracy and robustness in terms of reconstruction error distribution, alarm rate stability, and Top-K anomaly consistency. The method can effectively identify various types of abnormal operating conditions in unlabeled datasets based on unsupervised learning. This study provides a transferable deep learning solution for enhancing the safety monitoring of electric bicycle batteries. Full article

Current drone navigation systems primarily rely on Global Navigation Satellite Systems (GNSSs), but their signals are susceptible to interference, spoofing, or suppression in complex environments, leading to degraded positioning performance or even failure. To enhance the positioning accuracy and robustness of low-altitude drones in satellite-denied environments, this paper investigates an absolute visual localization solution. This method achieves precise localization by matching real-time images with reference images that have absolute position information. To address the issue of insufficient feature generalization capability due to the complex and variable nature of ground scenes, a visual-based image retrieval algorithm is proposed, which utilizes a fusion of shallow spatial features and deep semantic features, combined with generalized average pooling to enhance feature representation capabilities. To tackle the registration errors caused by differences in perspective and scale between images, an image registration algorithm based on cyclic consistency matching is designed, incorporating a reprojection error loss function, a multi-scale feature fusion mechanism, and a structural reparameterization strategy to improve matching accuracy and inference efficiency. Based on the above methods, a hierarchical absolute visual localization system is constructed, achieving coarse localization through image retrieval and fine localization through image registration, while also integrating IMU prior correction and a sliding window update strategy to mitigate the effects of scale and rotation differences. The system is implemented on the ROS platform and experimentally validated in a real-world environment. The results show that the localization success rates for the h, s, v, and w trajectories are 95.02%, 64.50%, 64.84%, and 91.09%, respectively. Compared to similar algorithms, it demonstrates higher accuracy and better adaptability to complex scenarios. These results indicate that the proposed technology can achieve high-precision and robust absolute visual localization without the need for initial conditions, highlighting its potential for application in GNSS-denied environments. Full article

Dynamic state estimation of floating offshore wind turbines (FOWTs) in complex marine environments is a core challenge for digital twin systems. This study proposes a joint estimation framework that integrates windowed dynamic mode decomposition (W-DMD) and an adaptive strong tracking Kalman filter (ASTKF). W-DMD extracts dominant modes under stochastic excitations through a sliding-window strategy and constructs an interpretable reduced-order state-space model. ASTKF is then employed to enhance estimation robustness against environmental uncertainties and noise. The framework is validated through numerical simulations under turbulent wind and wave conditions, demonstrating high estimation accuracy and strong robustness against sudden environmental disturbances. The results indicate that the proposed method provides a computationally efficient and interpretable tool for FOWT digital twins, laying the foundation for predictive maintenance and optimal control. Full article

In response to the challenge of dynamic adaptability in operational safety assessment for UAVs operating in complex operational environments, this study proposes a novel operational safety assessment method based on an Improved Support Vector Machine. An operational safety assessment index system encompassing four dimensions—operator, UAV platform, flight environment, flight mission—is constructed to provide a comprehensive foundation for evaluation. The method introduces a dynamic weighted information entropy mechanism based on a sliding window, overcoming the static features and delayed response of traditional SVM methods. Additionally, it integrates Gaussian and polynomial kernel functions to significantly enhance the generalization capability and classification accuracy of the SVM model in complex operational environments. Experimental results show that the proposed model demonstrates superior performance on test samples, effectively improving the accuracy of operational safety assessment for the Reconnaissance–Strike UAV in complex operational environments, and offering a novel methodology for UAV safety assessment. Full article

There have been significant breakthroughs in developing models for segmenting 3D medical images, with many promising results attributed to the incorporation of Vision Transformers (ViT). However, the fundamental mechanism of transformers, known as self-attention, has quadratic complexity, which significantly increases computational requirements, especially in the case of 3D medical images. In this paper, we investigate the UNETR++ model and propose a voxel-focused attention mechanism inspired by TransNeXt pixel-focused attention. The core component of UNETR++ is the Efficient Paired Attention (EPA) block, which learns from two interdependent branches: spatial and channel attention. For spatial attention, we incorporated the voxel-focused attention mechanism, which has linear complexity with respect to input sequence length, rather than projecting the keys and values into lower dimensions. The deficiency of UNETR++ lies in its reliance on dimensionality reduction for spatial attention, which reduces efficiency but risks information loss. Our contribution is to replace this with a voxel-focused attention design that achieves linear complexity without low-dimensional projection, thereby reducing parameters while preserving representational power. This effectively reduces the model’s parameter count while maintaining competitive performance and inference speed. On the Synapse dataset, the enhanced UNETR++ model contains 21.42 M parameters, a 50% reduction from the original 42.96 M, while achieving a competitive Dice score of 86.72%. Full article

Low Earth orbit (LEO) satellites, characterized by low altitudes, high velocities, and strong ground signal reception, have become an essential and dynamic component of modern global navigation satellite systems (GNSS). However, orbit decay induced by atmospheric drag poses persistent challenges to maintaining stable trajectories. Frequent orbit maneuvers, though necessary to sustain nominal orbits, introduce significant difficulties for precise orbit determination (POD) and navigation augmentation, especially under complex operational conditions. Unlike most existing methods that rely on Two-Line Element (TLE) data—often affected by noise and limited accuracy—this study directly utilizes onboard GNSS observations in combination with real-time precise ephemerides. A novel time-series indicator is proposed, defined as the geometric root-mean-square (RMS) distance between reduced-dynamic and kinematic orbit solutions, which is highly responsive to orbit disturbances. To further enhance robustness, a sliding window-based adaptive thresholding mechanism is developed to dynamically adjust detection thresholds, maintaining sensitivity to maneuvers while suppressing false alarms. The proposed method was validated using eight representative maneuver events from the GRACE-FO satellites (May 2018–June 2022), successfully detecting seven of them. One extremely short-duration maneuver was missed due to the limited number of usable GNSS observations after quality-control filtering. To examine altitude-related applicability, two Sentinel-3A maneuvers were also analyzed, both successfully detected, confirming the method’s effectiveness at higher LEO altitudes. Since the thrust magnitudes and durations of the Sentinel-3A maneuvers are not publicly available, these cases primarily serve to verify applicability rather than to quantify sensitivity. Experimental results show that for GRACE-FO maneuvers, the proposed method achieves near-real-time responsiveness under long-duration, high-thrust conditions, with an average detection delay below 90 s. For Sentinel-3A, detections occurred approximately 7 s earlier than the reported maneuver epochs, a discrepancy attributed to the 30 s observation sampling interval rather than methodological bias. Comparative analysis with representative existing methods, presented in the discussion section, further demonstrates the advantages of the proposed approach in terms of sensitivity, timeliness, and adaptability. Overall, this study presents a practical, efficient, and scalable solution for real-time maneuver detection in LEO satellite missions, contributing to improved GNSS augmentation, space situational awareness, and autonomous orbit control. Full article