Aug 2026· PLoS ONE· Vol 21, pp. e0356376· 0 citations· 30 references
Medicine
TL;DR
Combining deep learning methods with physical orbital models can effectively improve the accuracy of space object orbit prediction and provides an effective approach for orbital error prediction, space situational awareness, and collision warning.
Abstract
To address the strong dependence of space object orbit prediction on physical models and initial conditions, as well as the difficulty of completely eliminating prediction errors, this study proposes a satellite orbit prediction correction method that integrates an attention mechanism with a long short-term memory (LSTM) network. Taking the LAGEOS satellite as the research object, the proposed method uses position error, velocity, and acceleration features extracted from historical orbital data to train a deep learning model for predicting one-day-ahead orbital errors and correcting the SGP4 orbit prediction results. The experimental results show that the ATLSTM model outperforms the LSTM, support vector machine (SVM), back propagation neural network (BP), and bidirectional long short-term memory (BiLSTM) models in both orbital error prediction and correction. The residual ratios of ATLSTM in the X, Y, and Z axes are reduced to 3.68%, 4.77%, and 2.37%, respectively, effectively improving the accuracy of satellite orbital error prediction. Further analysis indicates that a reasonable setting of the number of neurons helps improve model performance, while the prediction difficulty increases with the extension of the prediction duration, suggesting that the ATLSTM model is more suitable for short-term orbital error prediction and correction. In addition, validation results for satellites at different orbital altitudes demonstrate that the proposed model has certain generalization capability. In summary, combining deep learning methods with physical orbital models can effectively improve the accuracy of space object orbit prediction and provides an effective approach for orbital error prediction, space situational awareness, and collision warning.
The proposed residual prediction method effectively solves the problem of insufficient utilization of environmental field features in short-term typhoon prediction and provides efficient and accurate deep learning technical support for typhoon disaster early warning.
Zewen Ming, Jinyuan Liu· International Conference on...· 0 citations
Comprehensive experiments on the NGSIM dataset validate the proposed model, demonstrating robust performance across structured highway driving scenarios and both the accuracy and computational efficiency of the proposed architecture.
Yang Li, Chengqian Jin, Zhikang Li· IEEE Access· 0 citations
Aiming at the deficiencies of traditional satellite attitude prediction methods in prediction accuracy, feature extraction capability and on-board applicability, this paper proposes a lightweight fusion model of LSTM and attention mechanism. The model adopts a single-layer LSTM to extract temporal features from attitude data and combines with the attention mechanism to enhance the ability to focus on key information. Multiple data preprocessing strategies are introduced to improve the quality of input data, and a relatively complete comprehensive evaluation system is constructed. Experimental results show that the model can realize the joint prediction of three-axis attitude angles and three types of disturbance torques including gravitational gradient torque, solar radiation pressure torque and magnetic torque. The prediction results are in good agreement with the measured data, which can provide effective support for satellite attitude control systems.
Minghui Wu, Zhijian You· Digital Signal and Computer...· 0 citations
To address the challenges of difficult fault feature extraction and low diagnostic accuracy caused by noise interference in quadrotor unmanned aerial vehicles (UAVs) under complex flight environments, this paper proposes a UAV fault diagnosis model (ResCNN-LSTM-ATT) that integrates residual convolutional neural network (ResCNN), bidirectional long short-term memory (BiLSTM), and an attention mechanism. The proposed model seeks to concurrently acquire deep spatial features and long-term temporal dependencies from noise-corrupted airborne multi-source sensor time-series data. Specifically, ResCNN is first employed to extract local spatial features from high-dimensional observational data and suppress noise via residual connections and one-dimensional convolutional neural networks. Secondly, a BiLSTM structure is constructed to capture the bidirectional temporal dependencies of the flight data. Subsequently, an attention mechanism is introduced to weight the outputs of the BiLSTM, focusing on key fault time steps. Finally, the fused spatiotemporal features are fed into the classifier. Experimental results on typical UAV fault diagnosis tasks, including motor, gyroscope, and magnetometer faults, demonstrate that the proposed method achieves superior diagnostic accuracy over existing state-of-the-art approaches while maintaining a compact architecture suitable for real-time onboard deployment.
Yucheng Wu, Tian Xie, Sen Yang· Engineering Research Express· 0 citations
High-precision state-of-charge (SOC) prediction is critical for electric vehicle (EV) safety and performance. To address the high computational complexity of existing data-driven methods, which rely on long historical sequences, this paper proposes a purely data-driven end-to-end SOC prediction framework based on sliding-window technology. The framework adopts an SCSSA-optimized convolutional neural networks (CNN)-long short-term memory (LSTM)-attention hybrid model that integrates a CNN for local feature extraction, an LSTM for modeling temporal dependencies, and an attention mechanism for adaptive feature weighting, with an improved sparrow search algorithm for global hyper-parameter optimization. Experiments are conducted using 29 months of operational data from 20 EVs. Results show that the proposed method achieves 14.8%, 8.9%, and 17.6% improvements in mean absolute error, root mean square error, and mean absolute percentage error, respectively, compared with the best benchmark model, with R2 consistently above 0.96. The method demonstrates excellent robustness across seasonal variations and diverse charging patterns, laying a solid technical foundation for SOC prediction in battery management systems.
Xing Zhang, Ju-Qiang Feng, Shunli Wang et al.· Engineering Research Express· 0 citations
Accurate weather prediction is crucial for sectors such as agriculture, disaster management, transportation, and energy. Traditional numerical weather prediction (NWP) models rely on complex physical equations and high computational resources, but often struggle with localized patterns and nonlinear spatio-temporal relationships. Recently, machine learning approaches, particularly hybrid CNN-LSTM models, have emerged as effective alternatives by combining spatial and temporal learning capabilities. This paper explores the theory, architecture, and performance of CNN-LSTM models, where CNNs extract spatial features from meteorological data and LSTMs capture temporal dependencies for sequential forecasting. The study reviews existing methods, including statistical, NWP, and deep learning models such as CNNs, RNNs, GRUs, and attention mechanisms. The proposed methodology includes data preprocessing, feature normalization, spatial encoding, temporal modeling, and supervised multi-step prediction. Performance is evaluated using MAE, RMSE, and R² metrics. Results show that hybrid CNN-LSTM models outperform baseline methods in accuracy, robustness, and forecasting across multiple time horizons. The paper concludes with future directions, including explainable AI, physics-informed models, and integration with NWP systems.
Peter Okello· International Journal of App...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
