The traffic data corrupted by noise and missing entries often lead to the poor performance of Intelligent Transportation Systems (ITS), such as the bad congestion prediction and route guidance. How to efficiently impute the traffic data is an urgent problem. As a classic deep learning method, Generative Adversarial Network (GAN) achieves remarkable success in image recovery fields, which opens up a new way for the traffic data imputation. In this paper, we propose a novel spatio-temporal GAN model for the traffic data imputation (STGAN). Firstly, we design the generative loss and center loss, which not only minimizes the reconstructed errors of the imputed entries, but also ensures each imputed entry and its neighbors conform to the local spatio-temporal distribution. Then, the discriminator uses the convolution neural network classifier to judge whether the imputed matrix conforms to the global spatio-temporal distribution. As for the network architecture of the generator, we introduce the skip-connection to keep all well preserved data unchanged, and employ the dilated convolution to capture the spatio-temporal correlation in the traffic data. The experimental results show that our proposed method obviously outperforms other competitive traffic data imputation methods.

Crowd counting provides an important foundation for public security and urban management. Due to the existence of small targets and large density variations in crowd images, crowd counting is a challenging task. Mainstream methods usually apply convolution neural networks (CNNs) to regress a density map, which requires annotations of individual persons and counts. Weakly-supervised methods can avoid detailed labeling and only require counts as annotations of images, but existing methods fail to achieve satisfactory performance because a global perspective field and multi-level information are usually ignored. We propose a weakly-supervised method, DTCC, which effectively combines multi-level dilated convolution and transformer methods to realize end-to-end crowd counting. Its main components include a recursive swin transformer and a multi-level dilated convolution regression head. The recursive swin transformer combines a pyramid visual transformer with a fine-tuned recursive pyramid structure to capture deep multi-level crowd features, including global features. The multi-level dilated convolution regression head includes multi-level dilated convolution and a linear regression head for the feature extraction module. This module can capture both low- and high-level features simultaneously to enhance the receptive field. In addition, two regression head fusion mechanisms realize dynamic and mean fusion counting. Experiments on four well-known benchmark crowd counting datasets (UCF_CC_50, ShanghaiTech, UCF_QNRF, and JHU-Crowd++) show that DTCC achieves results superior to other weakly-supervised methods and comparable to fully-supervised methods.

Symmetry in traffic patterns is a fundamental aspect of intelligent transportation systems, aiming to precisely predict traffic flow in real time despite the complex interplay of spatial and temporal factors. This paper presents a novel method of traffic forecasting that incorporates parameters related to road symmetry into a Graph Convolution Network model. Our model is crafted to dynamically adjust to real-time changes in road conditions, including the presence of symmetric and asymmetric road layouts, which substantially influence traffic flow. We have developed a GCN model that not only accounts for standard traffic flow metrics but also integrates a matrix representing road symmetry. The model undergoes training and validation on the METR-LA dataset, showcasing a significant enhancement in prediction accuracy. In the comparative analysis of state-of-the-art methods, our model demonstrated a significant enhancement in performance, achieving 30.68% improvement in Mean Squared Error (MSE) and a 24.28% improvement in Mean Absolute Error (MAE) over the best-performing method. The implications of our research are profound for urban planners, traffic management systems, and navigation service providers, as it offers a more dependable tool for forecasting traffic conditions, aiding in road design, and refining route planning strategies based on the symmetry of road networks.

To reduce the hardware cost of massive multiple-input multiple-output (MIMO) systems, low-bit analog-to-digital converters (ADCs) and sparse arrays are widely used. Compared with traditional 1-bit and 2-bit quantization techniques, 1.5-bit quantization uses two symmetric non-zero thresholds to quantize signal power into three levels, thereby balancing quantization complexity against system performance. However, the quantization loss introduced by 1.5-bit quantization is still significant and leads to degradation in DOA estimation performance. To improve the DOA estimation accuracy of 1.5-bit sparse massive MIMO systems, a covariance matrix estimation method is proposed. This method exploits the Toeplitz property of the covariance matrix of sparse arrays and the relationship between 1.5-bit quantized signals and their unquantized counterparts to transform the covariance matrix estimation problem for 1.5-bit sparse arrays into a non-convex optimization problem with equality constraints. We then further exploit the properties of 1.5-bit quantized signals to relax this problem into a convex problem and solve it via semidefinite programming. Once the covariance is estimated, the DOAs can be recovered by subspace-based methods. Numerical results show that the proposed method achieves higher estimation accuracy than 1.5B-MUSIC and 1-bit covariance-fitting baselines on 1.5-bit sparse arrays, and is competitive with structured covariance-fitting baselines applied to unquantized data, especially on coprime arrays in low-snapshot scenarios.

Multi-source radar data fusion has become increasingly vital for advancing weather monitoring and forecasting. However, effectively integrating Doppler radar with an X-band phased-array radar remains challenging. Doppler radar offers only low and inconsistent spatial resolution, whereas an X-band phased-array radar provides high resolution but is limited by short detection range, severe signal attenuation, and high deployment costs, constraining its use to localized monitoring. To address the aforementioned challenges, this paper proposes the Multi-source Radar Reflectivity Complementary Enhancement method (MSR-CE). By constructing a paired training dataset, real X-band phased-array radar reflectivity data serve as the starting samples for the forward diffusion process, while paired S-band Doppler radar reflectivity data act as conditional guidance. Leveraging a conditional diffusion model, the method generates high-resolution pseudo X-band phased-array reflectivity fields. Additionally, a Radar-Physics-Aware Loss (RPA Loss) is introduced to enhance spatial detail fidelity and physical consistency. Experiments on multi-source radar observations from Northeast China in 2025 demonstrate that MSR-CE achieves an SSIM of 0.892 and a PSNR of 41.6 dB, outperforming traditional interpolation methods and state-of-the-art generative approaches in radar reflectivity enhancement.

This paper proposes a tropical cyclone formation prediction network based on pyramid attention feature extraction and multi-scale feature fusion, aiming to enhance the prediction of whether tropical cloud cluster (TCC) precursors intensify to tropical storm (TS) strength by integrating multi-source reanalysis data. First, reanalysis data from ERA5 and NCEP/NCAR are represented as images at different scales and labeled according to tropical cyclone (TC) formation events derived from the TCC and IBTrACS datasets. The labels include information on whether TC formation occurred and its location. Then, a feature extraction module based on a Pyramid Attention Mechanism (PAM) is designed to extract features related to TC formation. Next, the features at different scales are input into a PAM-based feature fusion module that dynamically generates weights for different features to perform weighted fusion and unify the feature scales. Finally, a lightweight Convolutional Neural Network (CNN) is designed as the prediction module to predict TC formation occurrence and location. Experimental results over five independent data splits show that at a lead time of 24 hours, the proposed method achieves a Probability of Detection (POD) of $$0.865 \pm 0.012$$ , a False Alarm Ratio (FAR) of $$0.326 \pm 0.016$$ , and a location prediction Root Mean Square Error (RMSE) of $$0.795 \pm 0.033$$ grid units ( $$\approx$$ 437 km), demonstrating competitive performance in balancing POD and FAR. Ablation studies confirm the contribution of each proposed module to overall performance.