The radar radiation source signals hold extremely high reconnaissance value. Accurately positioning these signals constitutes one of the key technologies in safeguarding the security of the electromagnetic space. The positioning error in multi-station scenarios is influenced not only by the accuracy of positioning parameter estimation but also by the geometric configuration of the positioning platform. This paper focuses on the direction of arrival (DOA), frequency difference of arrival (FDOA), and time difference of arrival (TDOA) methods, analyzing the optimal configuration, optimal detection area, and optimal position dilution of precision in both elevation-known and elevation-unknown scenarios. Specifically, the paper constructs a signal receiving model, establishes the corresponding positioning equations, and performs dimensional normalization on these equations to derive measurement values in meters. Through differential processing, the position dilution of precision is obtained, which is then used as the optimization function to determine the optimal configuration, optimal detection area, and optimal position dilution of precision. Simulation results validate the accuracy of the proposed formulas.

This paper presents a new method for specific emitter identification (SEI) using the re-parameterization visual geometry group (RepVGG) neural network model and Gramian angular summation field (GASF). It converts in-phase and quadrature (IQ) signals into 2D feature maps, retaining both time and frequency domain features. Compared to residual network 18-layer (ResNet18) and Hilbert transform methods, this approach offers higher accuracy, faster training, and a smaller model size, making it ideal for hardware deployment.

In this paper, an adaptive neural backstepping control method based on barrier Lyapunov function is proposed for hypersonic vehicle considering full state constraints. The longitudinal dynamic of hypersonic vehicle can be divided into two subsystems, i.e., altitude subsystem and velocity subsystem and the controllers are designed with backstepping method, respectively. In the designing process, the radial basis function neural networks are used to approximate the unknown nonlinear functions of longitudinal dynamic, therefore, the accuracy requirement of hypersonic vehicle model is largely reduced. In order to handle the explosion of complexity issues occurring in the backstepping method, a tracking differentiator is introduced to calculate the differential of virtual control law. The barrier Lyapunov function is constructed to overcome the full system dynamic state constraints and an auxiliary system is designed for overcome the input state saturation issue. The stability is carried out based on Lyapunov theory, and the signals of closed-loop system established are uniformly ultimately bounded. The simulation results show that the controller designed for hypersonic vehicle can guarantee the good tracking performance.

Micro-Doppler parameter estimation is crucial for moving targets. However, conventional methods face limitations like inadequate time-frequency (TF) resolution and poor generalization, while existing deep learning approaches often treat TF analysis as a fixed preprocessing step. To overcome these challenges, this paper introduces a radar micro-Doppler parameter estimation method based on a gated dual-path dynamic-wavelet convolutional network (GDWCN). The GDWCN is an end-to-end deep learning framework that maps raw radar signals to micro-motion parameters by integrating clutter suppression, gated dual-path module, feature extraction, and parameter regression. Its core innovation is a gated dual-path module that combines dynamic convolution and learnable wavelet convolution, selecting the optimal processing path based on input signal characteristics. For the Inspire 2 drone, GDWCN reduced the mean absolute error (MAE) of frequency estimation by approximately 38% compared to the enhanced time-frequency micro-Doppler network, and its relative error by approximately 69% compared to the short-time Fourier transform (STFT), and 58% over the local maximum synchroextracting transform. Ablation studies further confirm the efficacy of the clutter suppression module and the attention mechanism.

Ship detection in synthetic aperture radar (SAR) image is crucial for marine surveillance and navigation. The application of detection network based on deep learning has achieved a promising result in SAR ship detection. However, the existing networks encounters challenges due to the complex backgrounds, diverse scales and irregular distribution of ship targets. To address these issues, this article proposes a detection algorithm that integrates global context of the images (GCF-Net). First, we construct a global feature extraction module in the backbone network of GCF-Net, which encodes features along different spatial directions. Then, we incorporate bi-directional feature pyramid network (BiFPN) in the neck network to fuse the multi-scale features selectively. Finally, we design a convolution and transformer mixed (CTM) detection head to obtain contextual information of targets and concentrate network attention on the most informative regions of the images. Experimental results demonstrate that the proposed method achieves more accurate detection of ship targets in SAR images.

To address the challenges of multi-scale differences, complex background interference, and unstable small target positioning in visual inspection of power towers, the existing methods still face issues such as insufficient feature interaction and unstable confidence estimation, which lead to performance degradation in complex backgrounds and occlusion conditions. This paper proposes a precise inspection method for key power tower components using autonomous drone positioning. To this end, this paper makes three key improvements to the you only look once version 11 (YOLOv11) framework. First, it constructs C3k2-adaptive multi-receptive field block (C3k2-AMRB), combining multiple dilated convolutions with a reparameterization mechanism to significantly expand the receptive field and enhance multi-scale feature extraction. Second, it designs a hierarchical wavelet interaction unit (HWIU), which leverages high- and low-frequency decomposition and reconstruction of wavelet transform (WT) to achieve cross-scale semantic alignment, enhancing feature discriminability in complex backgrounds. Third, it proposes a distribution-aware confidence refinement head (DACR-Head), which adaptively calibrates classification confidence based on the statistical characteristics of the predicted bounding-box corner distribution, improving the localization stability and accuracy of small targets. Experiments on the inspection of power line assets dataset (InsPLAD) dataset show that the integrated approach achieves a component detection accuracy at intersection over union (IoU)=0.5 (CDA₅₀) of 88.3% and a component detection robustness (CDR_50:95) of 69.8%, respectively, improvements of 4.4% and 7.0% over the baseline.

In wireless communication scenarios, especially in low-bandwidth or noisy transmission conditions, image data is often degraded by interference during acquisition or transmission. To address this, we proposed Wasserstein frequency generative adversarial networks (WF-GAN), a frequency-aware denoising model based on wavelet transformation. By decomposing images into four frequency sub-bands, the model separates low-frequency contour information from high-frequency texture details. Contour guidance is applied to preserve structural integrity, while adversarial training enhances texture fidelity in the high-frequency bands. A joint loss function, incorporating frequency-domain loss and perceptual loss, is designed to reduce detail degradation during denoising. Experiments on public image datasets, with Gaussian noise applied to simulate wireless communication interference, demonstrate that WF-GAN consistently outperforms both traditional and deep learning-based denoising methods in terms of visual quality and quantitative metrics. These results highlight its potential for robust image processing in wireless communication systems.

This paper presents a vision-based navigation framework for micro air vehicles (MAVs) operating in confined warehouse environments. To address the trade-off between low localization accuracy in mapless methods and high computational demands in map-based approaches, the proposed system leverages topology-aware path guidance using monocular vision. Navigation is driven by a digital instruction format (DIF) that encodes both the path index and target junction, enabling autonomous navigation without environmental modifications. The framework comprises a cascaded perception–encoding–control pipeline. For structured paths, foreground pixel density trend analysis with sliding window smoothing for robust junction recognition, while lateral proportional-integral-derivative (PID) control ensures accurate path tracking. For geometric trajectories, the control logic incorporates L-junction triggers, fixed-angle turns, and spatial yaw correction to accommodate sharp corners and curved segments. ROS-Gazebo simulations validate the method’s effectiveness, achieving up to 94.40% junction recognition accuracy (92.01% on average), trajectory tracking errors below 0.1 m, and terminal localization deviations under 0.2 m. These results validate the method’s accuracy, stability, and suitability for computationally constrained MAV platforms in warehouse automation.