Search Results (1,026)

Intelligent following control for electric-drive unmanned ground vehicles (UGVs) is important for human–robot collaboration and autonomous mobility. However, target occlusion, detection noise, and short-term detection failures can reduce following stability and reliability. To address these challenges, this paper proposes a vision-based following method that integrates adaptive Kalman filtering with target-state-aware control. The proposed method updates the observation noise covariance online using an innovation-sequence sliding window and evaluates the validity of visual measurements based on the intersection over union between adjacent frames, thereby enabling target position prediction and compensation under unstable observation conditions. Meanwhile, the longitudinal velocity and lateral angular velocity are dynamically adjusted according to the target detection state, image-center offset, and distance variation to achieve continuous and stable following control. Experimental results show that the proposed method reduces the target position prediction error by approximately 10% compared with conventional Kalman filtering and maintains the longitudinal following distance at about 3 m, improving the robustness, continuity, and motion smoothness of electric-drive UGV target following in complex environments. Full article

Aircraft mass is a fundamental input for aircraft performance calculations, and estimation errors may degrade trajectory-prediction accuracy and affect operational safety. To improve the reliability of aircraft mass estimation, this study proposes a takeoff mass estimation method based on flight parameters from stable climb segments. First, the phase characteristics of Quick Access Recorder (QAR) data are analyzed, and a sliding-window-based screening method is developed to identify stable climb segments. A takeoff mass estimation model is then established using the Total Energy Model (TEM), and an improved differential evolution algorithm is designed to solve the resulting nonlinear optimization problem through iterative search. Validation using historical A320 flight data shows that the proposed method produces takeoff mass estimates with relatively small discrepancies from the QAR-recorded reference values. Batch experiments involving A320 and B737 flights further show that the DE algorithm achieves absolute percentage errors of 1.64% and 2.60%, respectively. These results demonstrate the applicability of the proposed method to the two aircraft types considered in this study and indicate its potential to support aircraft performance monitoring and trajectory prediction. Full article

Terrain types significantly affect the dynamics and locomotion performance of hexapod robots during walking. Terrain classification is a key solution to modify gait patterns in different terrains and recognize hazardous conditions. Perceiving the terrain with proprioceptive sensing is a robust and reliable approach in extreme conditions. In this paper, an efficient terrain classification approach for a hexapod robot is proposed. The proposed method combines a deep classification framework including the long short-term memory (LSTM) network and an effective statistical feature extraction. Proprioceptive inertial measurement unit (IMU) data is only used as the sensing system for the robot–terrain interaction. In the feature extraction process, four meaningful characteristic features, namely, the mean, median, Lomb–Scargle periodogram power spectral density (LPSD), and Welch’s power spectral density (WPSD), are extracted from the body orientation data using a sliding-window method. These features are combined and fed into the network to perform the training and testing processes. In the experiments, the proposed method is evaluated with commonly used soft computing and deep learning models. The classification performance for the concrete, pebble, and waxed tile terrains reaches 100% with the proposed method. The overall accuracy, precision, sensitivity, specificity, F1-score, and Matthew correlation coefficient are recorded as 95.45%, 96.36%, 95.28%, 98.86%, 95.49%, and 94.61%, respectively. These results demonstrate that the proposed approach delivers reliable classification performance with a low-cost and easy-to-implement solution. Full article

Background: The genus Anchusa L. includes plants used in Uyghur medicine for their anti-inflammatory and analgesic effects. However, in China, the botanical origin of HERBA ANCHUSAE (Niushecao, a Uyghur medicinal herb) is severely confused. Traditional identification methods and standard DNA barcodes do not work well for these close relatives. Chloroplast genomes are known to contain variable regions that can help distinguish species, yet no such study has been done for Anchusa. Therefore, We compared the complete chloroplast genomes of six Anchusa species and the main adulterants of Niushecao. Methods: We analyzed genome structure, repeat sequences, codon usage bias, and nucleotide diversity (Pi), as well as conducted comparative and phylogenetic analyses. Results: All genomes shared a typical ring-shaped quadripartite structure, ranged from 150,178 to 150,844 bp in size, and contained the same set of genes. Despite this overall conservation, we identified several highly variable spots, mostly located in non-coding intergenic spacer regions. Using two complementary approaches—sliding window analysis and mVISTA-based sequence visualization—we identified three overlapping regions (rbcL-psaI, petA-psbJ, and trnC-GCA-petN) as candidate DNA barcodes for species identification. Our phylogenetic tree showed that Anchusa strigosa Banks & Sol. is most closely related to the true medicinal species Anchusa azurea Mill. (Bootstrap support (BS) = 100%), while other look-alikes formed separate branches. Conclusions: These findings provide the first chloroplast genomic resources for this genus and offer potential molecular markers for authenticating Anchusa medicinal materials, laying a foundation for future development of molecular authentication methods. Full article

The rapid development of large model-driven agent applications, such as digital assistants and robots, requires 6G radio access networks (RAN) to deliver enhanced flexibility, adaptability, and low-latency capabilities. However, the existing RAN user plane (UP) architecture suffers from coarse decoupling granularity and significant cross-layer functional redundancy. These limitations severely hinder the on-demand orchestration and dynamic reconfiguration required by heterogeneous agent services. To address these challenges, this paper proposes a ComBERT-driven service-based RAN UP decoupling method, specifically targeting the functional coupling and redundancy between the PDCP and RLC sublayers. First, we develop a domain-specific language model, ComBERT, by pre-training a BERT model on a 3GPP protocol corpus and fine-tuning it on text-matching tasks to deeply comprehend protocol semantics. Subsequently, ComBERT is utilized to extract semantic features from UP functional components, employing a sliding window mechanism to overcome truncation in lengthy protocol texts and using cosine similarity to measure functional relevance. Finally, a threshold-based fusion algorithm is designed to identify and merge cross-layer redundant functions, thereby forming independent service units with distinct responsibilities. These fused units serve as the basic building blocks for scenario-specific orchestration. Simulation results demonstrate that the proposed method reduces the number of UP components by 12.5%, 18.7%, and 18.2% in eMBB, URLLC, and mMTC scenarios, respectively. Simultaneously, it decreases average processing delays by 7.9%, 10.2%, and 11.0% across these respective scenarios. Ultimately, this approach effectively improves the lightweight deployment, processing efficiency, and reconfiguration capabilities of the service-based UP, providing a crucial foundation for on-demand service orchestration in 6G networks tailored to agent services. Full article

In Deep Learning Remote Sensing, data quantity is rarely the limiting factor. A single high-resolution satellite image can yield thousands of training patches. What determines model performance, yet remains largely overlooked, is the quality of those patches. To date, the choice of sampling method has rarely been treated as a methodological decision. Conventional approaches, namely sliding-window and random sampling, introduce two compounding data-quality problems: severe class imbalance caused by the overproduction of background-only patches and negative learning arising from incomplete annotations, where unlabeled objects are implicitly treated as negative examples during training. To address these limitations at the data construction stage, we propose object-centric patch sampling, a model-independent strategy that anchors each training patch to the geometric centroid of an annotated object. This design ensures that every object-anchored patch contains at least one target instance and substantially reduces exposure to unlabeled regions that generate false-negative supervision signals; only a small, deliberately controlled proportion of background-only patches is retained to preserve contextual variety without reinstating background dominance. The method is evaluated on three heterogeneous remote sensing datasets spanning satellite (Sentinel-2, 10 m), aerial (NAIP, 1 m), and UAV (0.25 m) imagery, covering cotton field segmentation, rural building extraction, and water body delineation, respectively. Using a U-Net architecture under identical training conditions, the proposed approach achieves IoU scores of 0.929, 0.896, and 0.912 on the three datasets, respectively, outperforming sliding-window sampling by up to 19.6 percentage points in IoU and consistently delivering higher F1-scores across all experimental configurations. Evaluation under DeepLabV3+ gives a mean IoU of 0.9504 and a mean F1-score of 0.9772 in a multi-class segmentation task, indicating that the gains are not specific to a single architecture. Unlike model-level solutions such as focal loss or class reweighting, the proposed method improves training data quality at its source and integrates seamlessly into any deep learning pipeline without architectural modifications. Full article

Since heavy clutter seriously restricts the ability of radar to detect targets, it is significant to build the target detector under heavy clutter. For practical situations without the secondary data or prior knowledge of target and clutter, this paper proposes a polarization-space-time detector. First, a general radar model is constructed for multiple pulses, multiple arrays, and multiple polarizations. Based on the theory of ternary hypothesis, the secondary data free (SDF) GLRT detector is proposed, which can maintain the constant false alarm probability (CFAR) in inhomogeneous clutter. Then, this paper proposes a matrix transform operator and an adaptive detection method using sliding window. These two approaches do not need to know the steering vector of radar and the noncentral parameter of clutter in advance, so the SDF-GLRT detector can adapt to different application scenarios. In addition, this paper optimizes the polarization waveform of the radar system by constructing a projection matrix. This method yields closed-form solutions of the optimal polarization and worst polarization, rather than relying on numerical solution. Finally, the performances of the SDF-GLRT detector and three other detectors are compared by simulated and real data. The proposed SDF-GLRT maintains $P_{\mathrm{FA}}$ of $5.4\times {10}^{-3}$ and $2.6\times {10}^{-3}$ on two IPIX datasets (#54 and #310) at a design $P_{\mathrm{FA}}={10}^{-3}$, whereas the other detectors deviate to $0.0249$–$0.7405$. The optimal polarization yields a detection-probability gain of more than $0.22$ over the worst polarization at $\mathrm{SCR}=0$ dB. Full article

Continuous-wave mud pulse telemetry offers high spectral efficiency and transmission rates, making it an important technology for high-speed information transmission under complex well conditions. However, surface-received signals are highly susceptible to periodic pressure pulsations generated by mud pumps, which degrade phase extraction and symbol decision performance. Dual-pressure-sensor delayed differential processing can exploit the correlated propagation characteristics of pump noise between two measurement locations to suppress its correlated components; however, its performance depends on accurately matching the propagation delay and amplitude compensation coefficient. To specifically address the dynamic variation in the pump noise propagation relationship between two measurement locations under actual operating conditions, a joint delay–amplitude compensation method is developed, in which pump noise suppression is formulated as the joint estimation of the signal propagation delay and amplitude compensation coefficient. Built upon LMS-based time delay estimation, the proposed method employs an enhanced time-varying step-size LMS time delay estimation algorithm (HTVSS-LMSTDE) to improve dynamic retracking capability following changes in propagation delay. A sliding-window weighted least-squares method (SWLS) is further introduced to estimate the amplitude compensation coefficient and correct differential mismatch caused by variations in the amplitude transfer ratio. With non-pump interference modeled as additive white Gaussian noise independent of the telemetry signal and pump noise, simulation results demonstrate that, when the propagation delay and amplitude transfer ratio vary simultaneously, the proposed method yields delay estimates and amplitude compensation coefficients close to their theoretically optimal values. Field wellbore tests further verify that the proposed method effectively attenuates low-frequency pump noise interference in continuous-wave mud pulse telemetry signals while preserving the BPSK-modulated information. Full article

Gravity-matching navigation is one of the important means for the covert navigation of underwater vehicles. The production and application of gravity reference maps as core key technologies have a very significant impact on the accuracy of underwater navigation. Firstly, this paper adopts the factor analysis method to obtain the comprehensive results of nine characteristic parameters such as the standard deviation and roughness of the gravity reference map by setting a range sliding window. Secondly, the TERCOM algorithm is introduced to conduct simulation verification calculations within the sliding window. After comparing and verifying with the comprehensive results of the factor analysis characteristic parameters, the limitations of statistical methods in the evaluation of the adaptability of gravity reference maps are analyzed. Thirdly, intelligent processing methods such as the learning vector quantization neural network algorithm and the extreme learning machine are proposed. The characteristic parameters of some sliding window gravity reference maps and the simulation verification results of the TERCOM algorithm are used as training samples to predict the adaptability evaluation effect of underwater gravity navigation for other sliding windows. The results show that the prediction results are generally in good agreement with the simulation verification results of the TERCOM algorithm. Compared with the learning vector quantization neural network algorithm, the extreme learning machine algorithm exhibits superior performance in terms of classification accuracy and computational efficiency. Full article

In continuous sign language recognition (CSLR), existing methods predominantly adopt frame-wise feature extraction, neglecting temporal continuity and motion trajectory modeling, thereby struggling to capture complete spatio-temporal dynamics. Meanwhile, global cross-modal attention approaches typically directly model interactions between text and the entire video sequence but lack structural constraints, making them susceptible to interference from redundant frames, which leads to attention distribution dilution and undermines fine-grained motion alignment capability. To address these issues, this paper proposes CMST-Net, a Cross-modal Interaction and Spatio-temporal Feature Enhancement Method for Continuous Sign Language Recognition. CMST-Net comprises two principal modules: the Local–Global Cross-modal Fusion Module (LGCFM) and the Spatio-Temporal Feature Enhancement Module (STFEM). LGCFM introduces a synergistic modeling mechanism that combines local sliding-window attention with global attention, achieving a unified fusion of structured local alignment and global semantic modeling. STFEM incorporates multi-scale spatial dilated convolution and coordinate attention to extract fine-grained spatial features while leveraging channel partitioning and a hierarchical residual structure to enhance long-range temporal modeling capability; the two modules collaboratively yield high-quality spatio-temporal feature representations. Experiments on three public benchmark datasets (PHOENIX2014, PHOENIX2014-T, and CSL-Daily) demonstrate that CMST-Net can effectively improve continuous sign language recognition performance, achieving state-of-the-art performance on the PHOENIX2014 and CSL-Daily datasets and competitive results on the PHOENIX2014-T dataset. Full article

Background/Objectives: Consciousness fluctuations involve brain network reorganization, yet the underlying neural synchronization mechanisms remain unclear. This study examined the static and dynamic characteristics of alpha-band functional connectivity during propofol sedation from two dimensions: phase synchrony and amplitude coupling. Methods: Electroencephalography data from 20 healthy volunteers across baseline, mild sedation, moderate sedation, and recovery were analyzed. Source-level signals for 68 cortical regions of interest were reconstructed using sLORETA. Dynamic functional connectivity matrices for both weighted Phase Lag Index (wPLI) and amplitude envelope correlation (AEC) were computed using 5 s sliding windows. Dynamic connectivity states were identified through clustering analysis, and state occurrence rates were compared between drowsy and responsive participants across sedation levels. Results: Static analysis revealed a dissociation between the two metrics: during moderate sedation, wPLI showed significant suppression in posterior parieto-occipital regions, whereas AEC exhibited widespread whole-brain coupling enhancement. Dynamic clustering identified three wPLI states and five AEC states. Critically, although the two metrics exhibited spatially distinct dynamic reconfiguration patterns, with deepening sedation, the occurrence rate of the ventral connectivity pattern in wPLI and that of the medial prefrontal pattern in AEC both increased significantly, and these two patterns showed synergistic co-occurrence. This effect was more pronounced in the drowsy subgroup, with greater increases in both patterns. Conclusions: Propofol-induced alterations in consciousness are not characterized by linear attenuation along a single neural synchrony dimension, but rather by differential reorganization of phase- and amplitude-based functional connectivity across spatial configurations and temporal dynamics. Full article

The synthesis of efficient reversible logic circuits is critical for fault-tolerant quantum computing (FTQC). The primary motivation of this work is to overcome the inherent disadvantages of existing synthesis techniques: approximate heuristic methods often miss optimal solutions, while pure exact computational methods suffer from combinatorial explosion on deep circuits. While the strict NCT library (NOT, CNOT, Toffoli) is often preferred due to the high cost of distilling non-Clifford states required for arbitrary gates, standard physical implementations frequently utilize the broader NCV library (NOT, CNOT, V, V-dagger), requiring the decomposition of Toffoli gates into five elementary operations. To bridge this gap, this paper presents a unified, highly scalable methodology for the optimal design of reversible circuits across both libraries. First, a Mixed-Integer Non-Linear Programming (MINLP) formulation, linearized for the high-performance IBM ILOG CPLEX solver, is introduced to automate the exact generation of globally optimal strict NCT topologies. Second, a systematic four-phase optimization framework is proposed to reduce NCV costs. By replacing Toffoli gates with specific NCV decompositions, permuting control lines to match subsequent linear gates, and applying exact local searches via an extended MINLP solver on bounded sliding windows, significant gate cancellations are achieved. Applying this methodology to prominent primitives (MIG, SAYEM, URG, TSG, and MKG), we match global NCT optimality constraints and achieve highly optimized NCV Quantum Costs of 7, 14, and 12 for the MIG, TSG, and MKG gates, respectively, establishing best-known upper bounds that significantly outperform heuristic literature benchmarks. Full article

Industrial sensor-based fault diagnosis often requires continuous data acquisition, local data processing, and timely model inference on edge devices. Although deep learning-based diagnostic methods have achieved promising performance, many existing approaches rely on cloud-centered processing pipelines that introduce communication overhead and potential data privacy concerns. This paper presents a CPU–NPU heterogeneous edge fault diagnosis framework for industrial sensor data. The framework runs on an RK3588 local edge device and includes SQLite- and RingBuffer-based data management, sliding window generation, micro-batch construction, and model inference. The CPU is responsible for data access and buffering, preprocessing, and micro-batch preparation, while the NPU executes fault diagnosis models using the RKNN runtime environment. By performing inference locally, the framework reduces the continuous transmission of raw sensor data and supports real-time fault diagnosis under resource-constrained edge devices. Experimental results demonstrate high consistency between ONNX-based CPU inference and RKNN-based NPU inference after model conversion. Furthermore, the effects of different data input paths and micro-batch configurations are systematically evaluated. A cross-platform comparison between server-class CPU/GPU execution and embedded NPU deployment is also conducted in terms of latency, throughput, and energy efficiency. The results show that RingBuffer-based streaming input significantly reduces data access overhead, while the effectiveness of NPU acceleration depends on both model structure and micro-batch size. The cross-platform results further demonstrate the energy efficiency advantages of the RK3588 platform, making it more suitable for practical deployment in resource-constrained edge scenarios. These findings provide practical insights for deploying fault diagnosis models on heterogeneous edge devices. Full article

Satellite battery subsystems are fundamental for reliable operation during periods when solar power is unavailable. Their performance is strongly influenced by the harsh and variable thermal conditions of the space environment. Pre-launch thermal simulations are routinely employed to evaluate subsystem behaviour; however, their ability to fully represent in-orbit dynamics remains limited. Meanwhile, machine learning (ML) approaches have emerged for satellite health monitoring, yet their practical role alongside conventional thermal analysis is not clearly established. This study investigates satellite battery temperature prediction by integrating pre-launch thermal simulation, real in-orbit telemetry, and data-driven ML forecasting within a unified framework. A dataset of 27,552 temperature measurements from an operational low-Earth-orbit satellite was analysed. A sliding-window regression approach was used to predict one-hour-ahead minimum and maximum battery temperatures, consistent with operational thermal margins. Three models—linear regression, Random Forest, and Extreme Gradient Boosting—were trained using historical temperature data and evaluated against telemetry and simulation outputs using MAE, RMSE, and R² metrics. Results indicate that linear regression achieved the highest accuracy (R² up to 0.98, MAE ≈ 0.20 °C), outperforming more complex models. ML-based predictions captured thermal behaviour more effectively than static simulation outputs under nominal conditions, while all predicted values remained within the acceptable operational range (10–30 °C). Rather than replacing physics-based methods, this work demonstrates that interpretable ML models can serve as an effective real-time complement to thermal simulations, offering practical insights into model selection and enhancing satellite battery thermal monitoring. Full article

In critical infrastructure environments, cybersecurity situation-awareness platforms generate large volumes of alerts, including substantial numbers of false positives, placing a considerable burden on security analysts. At present, alert correlation methods mainly rely on rule-based matching or statistical clustering, and large language models often lack the domain-specific threat intelligence required for reliable security analysis. This paper proposes GRAL, which is an AI-driven framework that combines graph neural networks (GNN) for cross-asset temporal alert correlation, retrieval-augmented generation (RAG) for dynamic threat intelligence enrichment, and large language models (LLM) for semantic reasoning and verdict generation. A temporal heterogeneous graph attention network constructs alert-relation graphs within a 72 h sliding window, and temporal decay and multi-relational dependencies are captured. Powered by bge-m3 embeddings and a dense vector index, the RAG module retrieves the most relevant threat intelligence entries above a cosine similarity threshold of 0.75. A domain-specific dataset of 1000 annotated security alerts from a nuclear power operational environment was built, and Cohen’s Kappa reached 0.87. The experiments show that GRAL achieves a macro-averaged precision of 87.0%, a macro-averaged recall of 97.0%, and a binary false-positive rate of 9.1%, together with 92.5% alert compression. Generalisation capability is confirmed by cross-dataset evaluation on CICIDS2017 (93.0% accuracy and 92.5% F1-score) and UNSW-NB15 (89.4% accuracy and 89.8% F1-score). Full article