Search Results (72)

To address the limitations of insufficient feature representation, large model size, and high deployment cost in network traffic classification, a lightweight classification framework based on multi-teacher knowledge distillation is proposed. The framework consists of two heterogeneous teacher networks and a compact student network to enable end-to-end traffic classification under constrained computational resources. The teacher networks incorporate complementary spatio-temporal modeling strategies, including a bidirectional temporal convolutional network (BiTCN) enhanced with attention mechanisms and convolutional neural network (CNN), and a parallel spatio-temporal fusion architecture integrating bidirectional long short-term memory (BiLSTM) and CNN. Knowledge from the teacher ensemble is distilled into a lightweight CNN-based student network through soft-target supervision, leading to improved generalization capability with significantly reduced model complexity. Experimental results demonstrate that effective knowledge transfer is achieved while reducing model parameters by more than 80%, and performance gains of about 1–3% are obtained compared with baseline methods. These results indicate strong potential for practical deployment in resource-constrained network environments. Full article

To accurately predict the rate of penetration (ROP) for steeply inclined coal seam blocks, this paper proposes a data-driven ROP prediction method incorporating feature processing. First, Savitzky–Golay (SG) filtering is applied to key continuous monitoring parameters to mitigate the impact of noise on model training. Subsequently, features are comprehensively screened across linear, monotonic, and nonlinear dependency dimensions using the Pearson correlation coefficient, Spearman correlation coefficient, and mutual information evaluation, identifying structural parameters significantly contributing to ROP. Based on this, a Time Convolution Network (TCN)-Bidirectional Long Short-Term Memory (BiLSTM)-Attention prediction model is constructed: TCN extracts local temporal patterns, BiLSTM captures forward and backward dependencies, and the attention mechanism adapts weight distribution for information across different time steps. This architecture significantly enhances the model’s ability to capture complex operational variations and improves prediction accuracy. Experimental results demonstrate that compared to benchmark models such as BiLSTM, TCN-BiLSTM, and BiLSTM-Attention, our method achieves superior performance across all evaluation metrics and exhibits strong generalization capabilities on diverse operational datasets. Full article

Against the backdrop of the global energy transition toward deep electrification, the natural gas industry faces challenges, including increased load forecasting uncertainty and frequent extreme weather impacts. To enhance natural gas load forecasting accuracy and support system resilience planning, this study constructs a forecasting model based on quadratic decomposition and hybrid deep learning, incorporating an electricity substitution coefficient to characterize the coupling substitution effect between electricity and natural gas. Under the basic scenario, the VMD-WPD-TCN-BiGRU model is proposed. It employs variational mode decomposition and wavelet packet denoising for secondary signal denoising, combined with a time-series convolutional network and bidirectional gated recurrent unit to extract temporal features. Experiments demonstrate that, compared to mainstream methods such as CNN, BiLSTM, SVM, and XGBoost, this model achieves statistically significant reductions in MSE (11.11–96.21%), MAE (0.89–76.50%), and MAPE (4.10–67.94%), significantly improving forecasting accuracy. In the deep electrification scenario, the introduction of the electricity substitution coefficient further optimizes peak load forecasting for system pressure days under extreme low temperatures, elevating the overall R² to 0.9905 in the deep electrification scenario. Research indicates that the proposed model not only effectively improves the accuracy of short-term natural gas load forecasting but also provides quantitative support for enterprises to plan peak-shaving facilities, optimize pipeline networks, and respond to extreme weather emergencies in data silo environments. This contributes to strengthening the adaptability and long-term resilience of natural gas systems during the energy transition, thereby supporting the sustainable development of energy infrastructure. Full article

With the increasing penetration of offshore wind power, extreme marine conditions pose significant challenges to forecasting accuracy and grid stability. To address this issue, this study proposes a robust offshore wind power forecasting framework based on multi-source feature fusion and a hybrid TCN–BiLSTM–KAN architecture. Specifically, a Temporal Convolutional Network (TCN) is employed to extract local multi-scale temporal features and suppress high-frequency disturbances, followed by a Bidirectional Long Short-Term Memory (BiLSTM) network to capture long-term temporal dependencies. A Kolmogorov–Arnold Network (KAN) is further integrated as a nonlinear mapping module to approximate complex dynamics under extreme marine conditions. The model is validated using a real-world offshore wind power dataset with a 15 min forecasting horizon, where balanced samples are constructed across different operating conditions. Experimental results demonstrate that, under extreme conditions, the proposed model achieves an RMSE of 3.58 MW and an R² of 97.84%, with RMSE reductions of 56.8% and 42.3% compared to CNN-BiLSTM and Transformer-KAN, respectively. Furthermore, cross-site validation confirms that the model maintains stable predictive performance, indicating its preliminary spatial generalization capability. Overall, the proposed framework provides an effective solution for enhancing forecasting reliability and supporting secure grid integration of offshore wind power under extreme marine environments. Full article

Atrial fibrillation is the most prevalent sustained cardiac arrhythmia and a major risk factor for stroke, heart failure, and premature mortality. Automatic detection remains challenging due to the variability of electrocardiogram (ECG) morphology, noise, and the paroxysmal nature of atrial fibrillation events. This study proposes a comprehensive framework that integrates optimised segmentation, feature extraction, and advanced deep learning architectures to improve detection accuracy. A coalescence window is introduced to dynamically cluster arrhythmic episodes, aligning computational analysis with clinical event distributions. Multiple classifiers are investigated, ranging from traditional machine learning models to state-of-the-art deep neural networks, including Temporal Convolutional Networks (TCNs), Convolutional Neural Networks (CNNs), and Bidirectional Long Short-Term Memory (BiLSTM). Experimental evaluation on a balanced dataset of ECG signals demonstrates the superior performance of deep learning models, with the best architecture achieving high accuracy and F1-score, significantly outperforming traditional approaches. Furthermore, the proposed pipeline is designed to be modular and resource-aware, supporting potential deployment in real-time and edge computing environments. These results highlight the feasibility of scalable atrial fibrillation monitoring systems that bridge algorithmic innovation with clinical applicability, ultimately contributing to earlier diagnosis and improved patient management. Full article

Accurate fault diagnosis in rotating machinery is critical for predictive maintenance and operational reliability in industrial applications. Despite the effectiveness of deep learning, many models underperform due to manually selected hyperparameters, which can lead to premature convergence, overfitting, weak generalization, and inconsistent performance across binary and multi-class classification. To address these limitations, the study proposes a novel hybrid hyperparameter optimization framework that combines Robotic Brain Storm Optimization (RBSO) with Manta Ray Foraging Optimization (MRFO) to optimally fine-tune deep learning architectures, including MLP, LSTM, GRU-TCN, CNN-BiLSTM, and Transformer-LSTM models. The framework leverages RBSO for global search to promote diversity and prevent premature convergence, and MRFO for local search to enhance convergence toward optimal solutions, with their combined effect improving predictive model performance and methodological generalization. The approach was validated on three benchmark datasets, including Case Western Reserve University (CWRU), industrial machine fault detection (TMFD), and the Machinery Fault Dataset (MaFaulDa). Before optimization, Transformer-LSTM model achieved 98.35% and 97.21% accuracy on CWRU binary and multi-class classification, 99.52% and 98.57% on TMFD, and 98.18% and 92.82% on MaFaulDa. Following hybrid optimization, Transformer-LSTM exhibited superior performance, with accuracies increasing to 99.72% for both CWRU tasks, 99.97% for TMFD, and 99.98% and 98.60% for MaFaulDa, substantially reducing misclassification. These results demonstrate that the proposed RBSO–MRFO framework provides a scalable, robust, and high-accuracy solution for intelligent fault diagnosis in rotating machinery. Full article

Lithium-ion batteries are extensively employed in new energy vehicles, where accurate State of Charge (SOC) estimation is fundamental for optimal battery management. However, existing methods often rely on single-model approaches and fail to leverage the complementary advantages of multiple models. This study proposes an innovative hybrid estimation model integrating a Temporal Convolutional Network (TCN) that efficiently captures long-range temporal dependencies via dilated convolution and residual blocks, with a Bidirectional Long Short-Term Memory Network (BiLSTM) that extracts bidirectional context information to enhance the accuracy of SOC estimation. First, the Panasonic datasets are utilized, with current, voltage, and cell temperature selected as input features. Subsequently, the proposed model is evaluated under various temperature conditions and driving cycles, demonstrating high accuracy and robustness. Finally, comparative experiments are conducted against traditional methods, such as standalone TCN and Long Short-Term Memory (LSTM) networks, under both 10 °C and −10 °C operating conditions. The results show that the hybrid model achieves superior performance in error metrics. Specifically, based on a second-order resistor-capacitor network, at −10 °C, the Root Mean Squared Error is reduced by 0.948%, and at 10 °C, it decreases by 0.398%. Additionally, the Maximum Absolute Error is lowered by 2.751% at −10 °C and by 2.192% at 10 °C. These improvements highlight the model’s significant potential as an effective solution for SOC estimation in lithium-ion batteries. Full article

This study explores the feasibility of estimating tibiofemoral joint contact forces using deep learning models trained on in vivo biomechanical data. Leveraging the comprehensive CAMS-Knee datasets, we developed and evaluated two machine learning network architectures, a bidirectional Long Short-Term-Memory Network with a Multilayer Perceptron (biLSTM-MLP) and a Temporal Convolutional Network (TCN) model, to predict medial and lateral knee contact forces (KCFs) across various activities of daily living. Using a leave-one-subject-out validation approach, the biLSTM-MLP model achieved root mean square errors (RMSEs) as low as 0.16 body weight (BW) and Pearson correlation coefficients up to 0.98 for the total KCF (${\mathrm{F}}_{\mathrm{tot}}$) during walking. Although the prediction of individual force components showed slightly lower accuracy, the model consistently demonstrated high predictive accuracy and strong temporal coherence. In contrast to the biLSTM-MLP model, the TCN model showed more variable performance across force components and activities. Leave-one-feature-out analyses underscored the dominant role of lower-limb kinematics and ground reaction forces in driving model accuracy, while EMG features contributed only marginally to the overall predictive performance. Collectively, these findings highlight deep learning as a scalable and reliable alternative to traditional musculoskeletal simulations for personalized knee load estimation, establishing a foundation for future research on larger and more heterogeneous populations. Full article

To enhance intelligent decision-making for tunneling operations in complex geological conditions, this study proposes a high-precision prediction method for TBM cutterhead torque using engineering data from the west return-air roadway of the Shoushan No. 1 Mine in Pingdingshan, Henan (China). A multisource dataset integrating geological exploration data, TBM electro-hydraulic parameters, and surrounding rock–TBM interaction indicators was constructed and preprocessed through outlier removal, interpolation restoration, and Savitzky–Golay filtering to extract high-quality steady-state features. To capture the mechanical properties of composite strata, the equivalent strength parameter of composite strata and an integrity-classification index were introduced as key predictors. Based on these inputs, a hybrid TCN–BiLSTM–Logarithmic Attention model was developed to jointly extract local temporal patterns, model global dependencies, and emphasize critical operating responses. Testing results show that the proposed model consistently outperforms TCN, BiLSTM, and TCN-BiLSTM baselines under intact, transitional, and fractured rock conditions. It achieves an RMSE (19.85) and MAPE (3.72%) in intact strata, while in fractured strata RMSE (29.55) and MAPE (10.82%) are reduced by 23.5% and 22.7% relative to TCN. Performance in transitional strata is likewise superior. Overall, the TCN–BiLSTM–Logarithmic Attention model demonstrates the highest prediction accuracy across intact, transitional, and fractured strata; effectively captures the mechanical characteristics of composite formations; and achieves robust and high-precision prediction of TBM cutterhead torque in complex geological environments. Full article

Constructing a temperature and humidity prediction model for greenhouse-grown tomatoes is of great significance for achieving resource-efficient and sustainable greenhouse environmental control and promoting healthy tomato growth. However, traditional models often struggle to simultaneously capture long-term temporal trends, short-term local dynamic variations, and the coupling relationships among multiple variables. To address these issues, this study develops an iT-LSTM-CA multi-step prediction model, in which the inverted Transformer (iTransformer, iT) is employed to capture global dependencies across variables and long temporal scales, the Long Short-Term Memory (LSTM) network is utilized to extract short-term local variation patterns, and a cross-attention (CA) mechanism is introduced to dynamically fuse the two types of features. Experimental results show that, compared with models such as Gated Recurrent Unit (GRU), Temporal Convolutional Network (TCN), Recurrent Neural Network (RNN), LSTM, and Bidirectional Long Short-Term Memory (Bi-LSTM), the iT-LSTM-CA achieves the best performance in multi-step forecasting tasks at 3 h, 6 h, 12 h, and 24 h horizons. For temperature prediction, the R² ranges from 0.96 to 0.98, with MAE between 0.42 °C and 0.79 °C and RMSE between 0.58 °C and 1.06 °C; for humidity prediction, the R² ranges from 0.95 to 0.97, with MAE between 1.21% and 2.49% and RMSE between 1.78% and 3.42%. These results indicate that the iT-LSTM-CA model can effectively capture greenhouse environmental variations and provide a scientific basis for environmental control and management in tomato greenhouses. Full article

Continuous heart rate monitoring is crucial for early cardiovascular disease detection. To overcome the discomfort and limitations of ECG in home settings, we propose a multimodal temporal fusion network (MM-TFNet) that integrates ballistocardiography (BCG) and photoplethysmography (PPG) signals. The network extracts temporal features from BCG and PPG signals through temporal convolutional networks (TCNs) and bidirectional long short-term memory networks (BiLSTMs), respectively, achieving cross-modal dynamic fusion at the feature level. First, bimodal features are projected into a unified dimensional space through fully connected layers. Subsequently, a cross-modal attention weight matrix is constructed for adaptive learning of the complementary correlation between BCG mechanical vibration and PPG volumetric flow features. Combined with dynamic focusing on key heartbeat waveforms through multi-head self-attention (MHSA), the model’s robustness under dynamic activity states is significantly enhanced. Experimental validation using a publicly available BCG-PPG-ECG simultaneous acquisition dataset comprising 40 subjects demonstrates that the model achieves excellent performance with a mean absolute error (MAE) of 0.88 BPM in heart rate prediction tasks, outperforming current mainstream deep learning methods. This study provides theoretical foundations and engineering guidance for developing contactless, low-power, edge-deployable home health monitoring systems, demonstrating the broad application potential of multimodal fusion methods in complex physiological signal analysis. Full article

Addressing the challenges associated with wind turbine blade root loads—including nonlinearity, strong coupling effects, high computational complexity, and the limitations of conventional mathematical-physical modeling approaches. This paper proposes a wind turbine blade root load prediction model that integrates Variational Mode Decomposition (VMD) optimized by the Red-billed Blue Magpie Algorithm (RBMO) and a combined Temporal Convolutional Network (TCN)—Bidirectional Long Short-Term Memory (BiLSTM)—Attention mechanism. First, the RBMO algorithm optimizes VMD parameters. VMD decomposes data into multiple sub-sequences, which are combined with environmental and operational parameters to form input components for the TCN-BiLSTM-Attention ensemble prediction model. Finally, the RBMO algorithm determines the optimal hyperparameter configuration for the combined model. Prediction outputs from each component are then aggregated and reconstructed to yield the final blade root load prediction. Predictions are compared against actual data and results from other forecasting models. Results demonstrate superior predictive performance for the proposed model, effectively enhancing the accuracy of blade root load prediction for wind turbines. Full article

The model of a liquid desiccant dehumidification air conditioning (LDAC) system is one of the key foundations for achieving efficient cooling, dehumidification and regeneration, and saving energy consumption. The data-driven modeling method does not need to understand the complex heat and mass transfer mechanism and equipment physical information, thus the modeling complexity is greatly reduced. This paper proposes a temperature and humidity prediction model integrating the Black Kite Algorithm (BKA), Bidirectional Temporal Convolutional Network (BiTCN), Bidirectional Long Short-Term Memory (BiLSTM), and Self-Attention mechanism (SA). The model extracts local spatiotemporal features from sequence data through BiTCN, enhances the understanding of contextual dependencies in temporal data using BiLSTM, and employs the SA to assign dynamic weights to different time steps. Furthermore, BKA is adopted to optimize the hyperparameter combinations of the neural network, thereby improving prediction accuracy. To validate the model performance, an experimental platform for an LDAC system was established to collect operational data under multiple working conditions, constructing a comprehensive dataset for simulation analysis. Experimental results demonstrate that compared to conventional time-series prediction models, the proposed model achieves higher accuracy in predicting outlet temperature and humidity across various operating conditions, providing reliable technical support for system real-time control and performance optimization. Full article

The performance of integrated Global Navigation Satellite System and Inertial Navigation System (GNSS/INS) navigation often declines in complex urban environments due to frequent GNSS signal blockages. This poses a significant challenge for autonomous driving applications that require continuous and reliable positioning. To address this limitation, this paper presents a novel hybrid framework that combines a deep learning architecture with an adaptive Kalman Filter. At the core of this framework is a Temporal Convolutional Network and Bidirectional Long Short-Term Memory (TCN-BiLSTM) model, which generates accurate pseudo-GNSS measurements from raw INS data during GNSS outages. These measurements are then fused with the INS data stream using an Adaptive Noise-Regulated Iterated Extended Kalman Filter (ANR-IEKF), which enhances robustness by dynamically estimating and adjusting the process and observation noise statistics in real time. The proposed ANR-IEKF + TCN-BiLSTM framework was validated using a real-world vehicle dataset that encompasses both straight-line and turning scenarios. The results demonstrate its superior performance in positioning accuracy and robustness compared to several baseline models, thereby confirming its effectiveness as a reliable solution for maintaining high-precision navigation in GNSS-denied environments. Validated in 70 s GNSS outage environments, our approach enhances positioning accuracy by over 50% against strong deep learning baselines with errors reduced to roughly 3.4 m. Full article

As the global cement industry moves toward energy efficiency and intelligent manufacturing, refined control of key processes like precalciner outlet temperature is critical for improving energy use and product quality. The precalciner’s outlet temperature directly affects clinker calcination quality and heat consumption, so developing a high-accuracy prediction model is essential to shift from empirical to intelligent control. This study proposes a TCN-BiLSTM hybrid neural network model for the accurate prediction and regulation of the outlet temperature of the decomposition furnace. Based on actual operational data from a cement plant in Guangxi, the Spearman correlation coefficient method is employed to select feature variables significantly correlated with the outlet temperature, including kiln rotation speed, high-temperature fan speed, temperature A at the middle-lower part of the decomposition furnace, temperature B of the discharge from the five-stage cyclone, exhaust fan speed, and tertiary air temperature of the decomposition furnace. This method effectively reduces feature dimensionality while enhancing the prediction accuracy of the model. All selected feature variables are normalized and used as input data for the model. Finally, comparative experiments with RNN, LSTM, BiLSTM, TCN, and TCN-LSTM models are performed. The experimental results indicate that the TCN-BiLSTM model achieves the best performance across major evaluation metrics, with a Mean Relative Error (MRE) as low as 0.91%, representing an average reduction of over 1.1% compared to other benchmark models, thereby demonstrating the highest prediction accuracy and robustness. This approach provides high-quality predictive inputs for constructing intelligent control systems, thereby facilitating the advancement of cement production toward intelligent, green, and high-efficiency development. Full article