Total electron content (TEC) serves as a key parameter characterizing ionospheric conditions. Accurate prediction of TEC plays a crucial role in improving the precision of Global Navigation Satellite Systems (GNSS). However, existing research have predominantly emphasized spatial variations in the ionosphere, neglecting the periodic changes of the ionosphere with the diurnal cycle. In this paper, we propose a TEC prediction model, which simultaneously considers both spatial and temporal characteristics to extract spatiotemporal features of ionospheric distribution. Additionally, we integrate several space weather element datasets into the prediction model framework, allowing the generation of multiple space weather feature values that represent the influence of space weather on the ionosphere at different latitudes and longitudes. Moreover, we apply Gaussian process regression (GPR) interpolation to geomagnetic data to characterize impact on the ionosphere, thereby enhancing the prediction accuracy. We compared our model with traditional image-based models such as convolutional neural networks (CNNs), convolutional long short-term memory networks (ConvLSTMs), a self-attention mechanism-integrated ConvLSTM (SAM-ConvLSTM) model, and one-day predicted ionospheric products (C1PG) provided by the Center for Orbit Determination in Europe (CODE). We also examined the effect of using different numbers of space weather feature values in these models. Our model outperforms the comparison models in terms of prediction error metrics, including mean absolute error (MAE), root mean square error (RMSE), correlation coefficient (CC), and the structural similarity index (SSIM). Furthermore, we analyzed the influence of different batch sizes on model training accuracy to find the best performance of each model. In addition, we investigated the model performance during geomagnetic quiet periods, where our model provided the most accurate predictions and demonstrates higher prediction accuracy in the equatorial anomaly region. We also analyzed the prediction performance of all models during space weather events. The results indicate that the proposed model is the least affected during geomagnetic storms and demonstrates superior prediction performance compared to other models. This study presents a more stable and high-performance spatiotemporal prediction model for TEC.

A Grid Ionospheric Vertical Error (GIVE) prediction model based on GAT‐TCN is proposed in this paper. The model integrates the Graph Attention Network (GAT) with the Temporal Convolutional Network to enhance prediction accuracy while ensuring the integrity of the Satellite‐Based Augmentation System, thereby improving its availability. This model predicts the GIVE values for the next 24 hr based on the historical data from the previous 24 hr. Moreover, multiple space weather parameters and the Rate of TEC Index are employed to characterize the temporal and spatial features of space weather and ionospheric disturbances. To validate the effectiveness of the proposed prediction model, this study not only conducted comparative experiments with multiple models and ablation experiments on key features, but also performed an in‐depth comparison between the proposed prediction model and the real‐time Kriging model. The results show that, while meeting the integrity requirements, the proposed prediction model improves the error bounding performance by 55.19% and increases the availability contour coverage by 9.65% during quiet space weather periods compared to the real‐time Kriging model. During active periods, the error bounding performance improves by 53.13%, and availability contour coverage increases by 12.39%.

With the development of the rail transportation business, the perception of abnormal events in rail transportation has become particularly important. It is necessary to explore and innovate the technology by acquiring point cloud information through a new perception method - LiDAR(Light Detection and Ranging). In this paper, we propose a method of rail transit abnormal event sensing based on multi-source point cloud information. It can sense rail transit abnormal events in different areas of rail transit station lobby scenes. We use Matlab-based statistical filtering, plane fitting, and point cloud rotation to complete the pre-processing, and extract the valid point cloud data contained in the original point cloud. Then we use the voxel-centered difference method to perceive the dynamic targets in the rail transit environment. And we use the model of human characteristics and the deep learning method based on PointNet to complete the perception of human point cloud data in the dynamic targets. Finally, the human behavior patterns are discriminated based on the enclosing frame and skeleton extraction methods. The experimental results show that the proposed method can effectively sense the dynamic human targets in the station lobby scenes and provide data support for subsequent abnormal event detection.