Reconfigurable Intelligent Surfaces (RIS), as the core technology for 6G wireless communications and Integrated Sensing and Communication (ISAC), play a vital role in enhancing signal quality and improving communication efficiency. To address the critical issue of damaged or contaminated RIS elements during design and deployment, this paper proposes a failure diagnosis method based on Electromagnetic Physical Constraints (EMPC). First, a low-dimensional measurement matrix is generated via spatial encoding of sparse signals from impaired RIS elements, subject to physical constraints imposed by the RF device structure. Subsequently, to accurately model the multi-modal characteristics of signal distributions in multipath channels, a Scaled Gaussian Mixture (SMG) model is adopted to fit the distribution of sparse coefficients. Leveraging the spatial structural correlation within the RIS array, a center-oriented parameter propagation acceleration strategy is designed, significantly enhancing computational efficiency. Building upon this foundation, the model parameters of the RIS are mapped onto a Riemannian manifold using Information Geometry (IG) theory. This transforms the fault diagnosis problem into one of measuring the information distance between coordinates on the manifold. A fast Kullback-Leibler Divergence (KLD)-based difference detector is designed to rapidly diagnose the locations and quantity of failed RIS elements. The experimental evaluation comprehensively validates the method's performance through extensive numerical simulations (encompassing noise impacts, device parameters, diverse fault modes, and channel models) and practical measurements in a microwave anechoic chamber. The results are analyzed from multiple dimensions to discusss the method's effectiveness, robustness, and practical applicability.

Low-altitude Unmanned Aerial Vehicle (UAV) detection using LiDAR range images faces persistent challenges. These include sparse features for long-range targets, large scale variations caused by viewpoint changes, and severe interference from complex backgrounds. To address these issues, we propose an improved detection framework based on YOLOv10. First, we design a Swin-Conv hybrid module that combines sparse attention with deformable convolution. This module enables the network to focus on informative regions and adapt to target geometry. These capabilities jointly strengthen feature extraction for sparse, long-range targets. Second, we introduce Attentional Feature Fusion (AFF) in the neck to replace na & iuml;ve feature concatenation. AFF employs multi-scale channel attention to softly select and adaptively weight features from different levels, improving robustness to multi-scale targets. In addition, we systematically study how the viewpoint distribution in the training set affects performance. The results show that moderately increasing the proportion of low-elevation-view samples significantly improves detection accuracy. Experiments on a self-built simulated LiDAR range-image dataset demonstrate that our method achieves 88.96% mAP at 54.2 FPS, which is 4.78 percentage points higher than the baseline. Deployment on the Jetson Orin Nano edge device further validates the model's potential for real-time applications. The proposed method remains robust under noise and complex backgrounds. The proposed approach achieves an effective balance between detection accuracy and computational efficiency, providing a reliable solution for real-time target detection in complex low-altitude environments.

Shape from polarization (SfP) provides physically informative cues for surface normal recovery by exploiting the polarization state of reflected light. Existing SfP methods remain limited by the imperfect nature of the physical model and the scarcity of labeled datasets. Recent diffusion models are attractive for SfP because of their strong generative capability in geometry estimation. Their direct application remains challenging due to the need for polarization-guided diffusion and reduced stochastic ambiguity during inference. To address this issue, we propose PcdSfP, a polarization-conditioned diffusion model for shape from polarization. The unpolarized intensity, degree of polarization (DoP), and angle of polarization (AoP) are encoded as semantic anchors and polarimetric conditioning cues to guide a pre-trained latent diffusion backbone toward geometrically consistent normal recovery. A spatially adaptive noise injection strategy is further introduced to reduce ambiguity during inference, and joint supervision in latent and pixel spaces is adopted to preserve high-frequency boundaries and local geometric details. Experiments demonstrate that the proposed method effectively integrates the physical observables of polarimetric imaging with the structural priors of diffusion models, achieving improved accuracy and detail recovery in surface normal estimation across both object-level and scene-level SfP tasks.

Objective To address the challenges of low detection efficiency and high false negative rates in X-ray industrial inspection of printed circuit boards—resulting from insufficient three-dimensional defect samples—as well as the limitations of traditional data augmentation methods such as flipping and cropping in terms of the quantity and diversity of generated defect samples,and further considering the inadequate reconstruction performance of existing defect generation algorithms on images with complex textures and fine defect details,this paper proposes a three-dimensional defect sample generation method based on the latent diffusion model(LDM).The method aims to encode large-scale,hard-to-train 3D defect samples into a latent space for diffusion,thereby overcoming the high computational cost and resource demands associated with training 3D data in conventional diffusion models like the denoising diffusion probabilistic model(DDPM).Moreover,by leveraging the iterative diffusion properties of noise-based diffusion models,the proposed model is redesigned to improve the completeness and authenticity of defect detail reconstruction.This approach offers a fast and effective 3D data augmentation solution for subsequent tasks such as 3D defect classification and detection. Methods Based on the LDM,an improved design is proposed to achieve three-dimensional defect sample augmentation.First,a vector-quantized generative adversarial network(VQGAN)is trained to compress three-dimensional data into a latent space,which improves the efficiency of 3D sample generation while preserving image feature information.Additionally,a reconstruction alignment loss is introduced during the autoencoder reconstruction process.By aligning the latent features obtained from VQGAN training with the output features of the pre-trained visual model DINOv2,the latent space features are standardized to retain richer semantic information.This enhances reconstruction quality,accelerates model training convergence,and improves reconstruction efficiency.Next,with VQGAN frozen,the LDM is trained.A conditional encoder incorporates a conditional input mask into the diffusion process to control the location and size of defects,thereby generating 3D defect images that satisfy the specified conditions.Finally,an attention module is integrated into the decoder of the diffusion model,enabling the model to focus on defect details and further improving the overall quality of defect generation. Results and Discussions A comparison of the proposed method with several existing models for defect image detail generation(Fig.5)shows that the proposed approach preserves defect details more completely and achieves higher overall image quality,with PSNR and SSIM values of 31.2476 dB and 0.9688,respectively,ranking the best and second among all compared methods,respectively(Table 1).Furthermore,through slice-wise comparison of the 3D images generated by each model(Fig.6),the proposed method demonstrates reduced loss of defect details and exhibits closer resemblance to real samples.An ablation study is conducted to evaluate the contributions of the reconstruction alignment loss and the multi-head deep convolutional transposed attention(DMTA)module.The results compare the generation quality(Table 2)and running time(Table 3)of the original LDM model against models augmented with each component separately.Both the reconstruction alignment loss and the DMTA module improve generation quality to varying degrees.Moreover,the inclusion of the reconstruction alignment loss substantially reduces the training time of the autoencoder VQGAN and the latent diffusion process,accelerating model convergence without increasing the number of trainable parameters,thus enhancing the efficiency of 3D defect sample generation.Finally,classification performance tests are performed on the generated 3D defect samples(Table 4).The proposed method improves classification accuracy by 18.0 percentage points compared with the unaugmented dataset.This outcome confirms that data augmentation can effectively mitigate performance degradation in classification models caused by insufficient defect samples,further validating the effectiveness of the proposed approach. Conclusions To address the issue of insufficient 3D defect samples in printed circuit board inspection,we propose a 3D defect sample generation method based on an improved LDM.By employing a defect mask map to guide the generation process of the LDM and introducing a reconstruction alignment loss during the training of the autoencoder VQGAN,the training difficulty of 3D defect samples is significantly reduced without increasing the number of model parameters,while the reconstruction performance is effectively improved.Moreover,the incorporation of the DMTA attention module enhances the model's ability to capture detailed defect information,further improving the quality of generated samples.Validation through comparative classification tests confirms that the proposed model can effectively alleviate problems caused by limited defect sample data in downstream tasks.Although the improved model enhances the generation quality of 3D defect samples to some extent,it still faces challenges such as high parameter counts and long training times.Future work will focus on further optimization to improve both the quality and efficiency of 3D defect sample generation.Additionally,the proposed model holds theoretical potential for generating various types of 3D defect images.Therefore,subsequent research will explore its application to other defect categories commonly found in printed circuit board solder joints,such as solder bridges and pillow-type defects,with the aim of expanding the model's capability to generate diverse defect types and mitigate data scarcity across different solder joint defect classes.

In this paper, we propose a new time-frequency analysis (TFA) method, namely scaling-basis chirplet extracting transform (SBCET). Based on the time-frequency representation (TFR) results obtained by scaling-basis chirplet transform (SBCT), the method introduces a new "extraction operator" to extract the time-frequency (TF) energy associated with the signal to portray the TF energy distribution information of the signal with high accuracy. SBCET can also obtain a TFR with concentrated energy and high resolution for non-stationary signals with close frequency intervals and intense background noise. The effectiveness and superiority are proved by numerical signal processing and experimental verification.

To overcome the shortcomings of existing time-frequency (TF) analysis (TFA) methods in analyzing signals containing cross-instantaneous frequencies (IFs), this paper proposes an adaptive TFA technique combined with image processing methods based on local maximum synchrosqueezing transform (SST). The core idea of the proposed algorithm is to localize the filtering of signals containing several different IF components using kernel functions containing several different directions, respectively, to achieve energy separation at the crossing frequencies. In turn, the local maximum SST is used to rearrange the TF energy to the true IF ridges of the signal to improve the TF energy concentration. Simulation data demonstrates that the proposed algorithm has higher energy aggregation and better noise immunity, especially for signals with cross-IFs. Applying the proposed method to animal acoustic and radar wave signals of pedestrians can accurately describe the differences in the frequency change patterns and the temporal distribution of energy in the signals, thereby providing a judgment basis for effectively identifying and classifying the signals.

To achieve ultrahigh resolution, accuracy, repeatability, and sensitivity in stress birefringence measurements, a polarimetry measurement method based on photoelastic modulation was proposed. Two photoelastic modulators (PEMs) were cascaded and operated at different frequencies. The multi-harmonic terms of the modulation signals were demodulated by using digital phase-locked technology. Real-time, in situ calibration of phase modulation amplitudes of the PEMs were performed. The simultaneous measurement of stress birefringence retardance and fast axis azimuth was achieved. A Soleil-Babinet compensator was employed to validate the accuracy and repeatability of the method, and the stress birefringence distribution in a CaF 2 crystal and BK7 glass were then measured. The experimental results revealed a superior standard deviation of retardance, measuring below 0.043 nm, and a maximum relative deviation of less than 0.71%. The single-point data measurement time was only 20 ms. This method offers an effective approach and istrument for assessing stress in optical materials or components, such as glass and crystals.

To propose a method for testing the dynamic change rate of the throat diameter based on the characteristics of the jet noise field, this study measured the effect of the size variation of the throat diameter on the jet noise during engine nozzle operation. Steady-state flow field simulation and transient noise simulation based on the Ffowcs Williams–Hawkings equation were carried out using large-eddy simulation for nozzles with different sizes of the throat diameter. The jet noise’s sound pressure level (SPL) is discussed and analyzed. By performing a Fourier transform on the noise signal obtained from the simulation, the influence rule of throat diameter size on the magnitude and frequency characteristics of jet noise is analyzed. The results show that the jet noise signal mainly manifests in the low-frequency region. With a gradual increase in frequency, the corresponding amplitude will first increase and then decrease. The effect of the change in the nozzle throat size on the structure of the jet wave system is more prominent; with the increase in the throat diameter, the SPL of the jet noise tends to increase, and the peak frequency of the SPL tends to decrease. The size of the nozzle throat diameter affects the change in the main frequency of the noise, which decreases as the nozzle throat diameter becomes larger.

The reconstruction algorithms play an important role in acoustic thermometry and are a research hotspot. This paper proposes a novel reconstruction method to improve the accuracy of temperature field reconstruction methods based on acoustic tomography. First, three mesh models are constructed by using the finite element method, which overcomes the defects of oversimplified classical discretization models. Secondly, the coefficient matrix A of the acoustic thermometry forward problem is calculated offline, and then the temperature values of several points in the measurement are obtained by solving the inverse problem using the improved regularization method based on singular value decomposition. Finally, the temperature profile of the entire region is obtained through interpolation. The correctness and feasibility of the proposed method are validated through simulations and experimental testing with three different temperature fields. Results show that the proposed method has satisfactory reconstruction accuracy and provides a different solution for reconstructing the ultrasound temperature field.

Automatic modulation classification (AMC) is vital in cognitive communication systems. Existing AMC methods are mainly designed for Gaussian noise channels, but research shows that non-Gaussian impulsive noise in wireless communication systems cannot be ignored. This paper proposes a novel AMC method based on Log-domain 3D constellation diagrams and multi-scale dual-convolutional 3DCNN (3D Convolutional Neural Networks). Firstly, we adopt non-linear logarithmic function transformation to effectively suppress non-Gaussian impulsive noise and transform multi-type modulated signals from the signal time domain to the graph domain, constructing Log-domain constellations. Secondly, the 2D constellations are projected onto three-dimensional space to form logarithmic domain 3D constellation diagrams, aiming to provide more discriminative feature dimensions for deep learning networks. Then, a new multi-scale dual-convolutional 3DCNN (MDC-3DCNN) is proposed, where the number and scale of dual-convolutional structures are the hyperparameters of MDC-3DCNN, to achieve feature extractions on both spatial and planar dimensions simultaneously for modulation classification. Numerical simulation results and actual measurement experiments demonstrate that the proposed method can effectively perform high-precision and robust classification of modulated signals in non-Gaussian impulsive noise channels.

In the context of a complex marine environment, extracting and recognizing underwater acoustic target features using ship-radiated noise present significant challenges. This paper proposes a novel deep neural network model for underwater target recognition, which integrates 3D Mel frequency cepstral coefficients (3D-MFCC) and 3D Mel features derived from ship audio signals as inputs. The model employs a serial architecture that combines a convolutional neural network (CNN) with a long short-term memory (LSTM) network. It replaces the traditional CNN with a multi-scale depthwise separable convolutional network (MSDC) and incorporates a multi-scale channel attention mechanism (MSCA). The experimental results demonstrate that the average recognition rate of this method reaches 87.52% on the DeepShip dataset and 97.32% on the ShipsEar dataset, indicating a strong classification performance.

In the fields of film coating, microelectronics, biosensing, and nanomaterial characterization, there is a growing demand for rapid and highly sensitive ellipsometry. Photoelastic modulation ellipsometry has emerged as a method capable of achieving unparalleled measurement speeds and sensitivities. However, the performance of this ellipsometry technique is contingent upon the precise calibration and long-term stability control of photoelastic modulation modulator (PEM). In this study, a differential frequency photoelastic modulation ellipsometry was proposed. Two PEMs operating at distinct frequencies were cascaded, and multi-harmonic terms of the modulation signals were extracted using digital phase-locking technology. Real-time/in-situ calibration and control of PEM was accomplished within a single Field-Programmable Gate Array. The developed system facilitated high-speed and high-sensitivity ellipsometric parameter measurements. To validate the proposed methodology, an experimental system was meticulously constructed, and performance tests were conducted on air and SiO2 film samples. Additionally, a tensile force experiment was performed on a Cu/polyimide film. Results demonstrated that the proposed approach achieved highly sensitive ellipsometry measurements within milliseconds. The system exhibited repeatability and sensitivity up to 10−3 ° with an integration time of 200 ms. This research provided valuable insights into high-precision, high-sensitivity, and rapid in-situ ellipsometry measurements, offering promising applications in online real-time film technology detection.

利用炸点声音进行定位是目前热门的定位方法,为此设计了 一个阵列式炸点声音定位系统,完善了所采用的阵列式炸点声音采集电路,包括大动态声音处理模块、多量程信号比较模块、数据处理模块等.通过Multisim仿真验证了阵列式炸点声音采集电路功能,其中大动态声音放大电路可以高倍数放大声音信号,将获取的微小信号放大成可处理的声音信号,多量程信号比较模块采用多阈值结构进行处理,比较后输出具有声音信息的逻辑信号.再通过单片机STM32连接上位机进行数据传输,采用MATLAB进行数据分析定位,实现对炸点位置的初步定位.

Modulated signal detection has been rapidly advancing in various wireless communication systems as it's a core technology of spectrum sensing. To address the non-Gaussian statistical of noise in radio channels, especially its pulse characteristics in the time/frequency domain, this paper proposes a method based on Information Geometric Difference Mapping (IGDM) to solve the signal detection problem under Alpha-stable distribution ( ff -stable) noise and improve performance under low Generalized Signal-to-Noise Ratio (GSNR). Scale Mixtures of Gaussians is used to approximate the probability density function (PDF) of signals and model the statistical moments of observed data. Drawing on the principles of information geometry, we map the PDF of different types of data into manifold space. Through the application of statistical moment models, the signal is projected as coordinate points within the manifold structure. We then design a dual-threshold mechanism based on the geometric mean and use Kullback-Leibler divergence (KLD) to measure the information distance between coordinates. Numerical simulations and experiments were conducted to prove the superiority of IGDM for detecting multiple modulated signals in non-Gaussian noise, the results show that IGDM has adaptability and effectiveness under extremely low GSNR.

针对目前ADC采样存在的集成化、成本和功耗相互制约等问题,实现了一种基于TDC的多通道高速采样方法,精度不受ADC采样位宽限制,只与所用FPGA芯片的硬件参数和软件设计有关,且采样通道数只与可配置的LVDS 口的数量和FPGA芯片资源大小相关.与传统ADC采样方法不同,首先,设计采集电路,用已知连续阈值信号作比较器阈值电压,TDC记录信号越过阈值时间获得采样点;其次,探究阈值信号和比较器对采样信号的影响,统计各影响因素的误差并寻找最优解;最后,总结采样规律并分析其适用性.

针对水下动态目标定位场景中,传感器节点能量有限且水下噪声影响节点获取目标测量信息导致定位精度低的问题,提出一种基于最大一致的最少Sigma点无迹Kalman滤波算法.该算法采用最少Sigma点集的无迹变换策略,与传统无迹Kalman滤波n维系统时2n+1个Sigma点的计算量相比改进为n+1个Sigma点的计算量,节省网络能耗;并采用最大一致策略来提高定位精度,同时引入虚拟向量避免节点同值问题.仿真结果表明,该算法定位精度高,计算量小,验证了所提改进算法的可行性和高效性.

在使用一发一收低频反射超声方法对大厚度介质进行测厚过程中,采集到的反射超声信号存在有效波和直达波混叠的现象,常规的一维超声信号处理方法难以对其进行有效处理.针对该类信号混叠问题,提出了基于线性Radon变换的自适应相减处理方法,从二维角度出发,根据在时空域中,直达干扰呈规律一致性,而有效底面回波呈推移差异性的特点,通过将信号阵列转换至线性Radon域的方法,对经过投影汇聚的信号数据进行自适应相减算法处理,实现对直达干扰波的有效压制及分离,从而提高低频超声回波数据的分辨率.通过仿真数据处理验证了该方法的有效性,相应厚度解算相对误差不超过0.7％.仿真和实验结果表明,提出的方法能够很好地解决大厚度介质测厚情境中混叠超声信号的分离问题.

In response to the problems of biased estimation of instantaneous frequency (If) and poor noise immunity in current time–frequency (Tf) analysis methods, the adaptive scale chirplet transform (ASCT) is proposed in this paper. The core idea of the proposed algorithm is to use a frequency-dependent quadratic polynomial kernel function to approximate the IF of the signal and to use the time-varying window length to overcome the frequency resolution problem due to the change in signal modulation. This method can dynamically select suitable parameters and overcome the disadvantage of unfocused energy of TF distribution. The experimental results show that the ASCT algorithm has high TF aggregation and can suppress noise interference well. In practical signal processing, the advantage of the ASCT algorithm is that it can accurately depict the characteristic frequency of the signal and detect the fault in the bearing signal. Both simulation and experimental results prove the strong realistic relevance of this algorithm.

Electromagnetic environment anomaly radiation source detection (EMEARD) can provide a basis for electromagnetic (EM) equipment behavior cognition and battlefield threat assessment. Existing methods mainly focus on generating EM space situations but ignore the research on anomaly signal detection. In this article, the connections between information geometry and performance of sensor networks for anomaly detection are explored. Nonlinear estimation methods based on information geometry can be directly used for anomaly detection. However, Riemann distance as a metric function faces the disadvantages of extensive computation and high complexity. To tackle this issue here, we propose a method based on image-like situation difference mapping, namely, ILSDM, which takes the data-driven strategy to obtain the analytic expression of situation difference mapping distance function (DMDF), and anomaly detection is equivalent to the mapping reconstruction of information geometry. First, the EM space situation is acquired through the distributed sensor network. In addition, to solve the perceptual blind area problem, the sensors’ data are estimated by a spatial autocovariance optimal interpolation algorithm (kriging). Second, we apply the Gaussian mixture model (GMM) to fit the statistical features of each sensing node into probability density function (pdf), and adjacent sensors with similar pdf are fused by the neighbor average fusion method (NAF). Then, based on the information geometry, the fused pdfs are mapped to the Riemannian manifold space as a series of coordinates, and we further propose to use an image-like situation (ILS) to represent the EM space situation. Finally, DMDF measures the information distance between different ILSs to achieve anomaly detection. Numerical simulations and real experiments were conducted to prove the superiority and feasibility of our method for anomaly detection in EM space.