A terahertz(THz) all-metal metasurface with a high Q-factor has been proposed,consisting of four diamond-shaped holes. By altering the size of adjacent diamond-shaped holes, the symmetry of the structure can be broken, resulting in quasi-Bound states In the Continuum(BICs). By changing the polarization angle of the incident wave, the frequency of the excited quasi-BICs remains unchanged, exhibiting polarization-independent characteristics. Simulation results show that, by optimizing structural parameters, when the size deviation is 3 μm, the quasi-BICs achieved an ultranarrow bandwidth of less than 1 GHz, corresponding to a Q-factor greater than 1 300. Experimental test results show that the fabricated sample(with a deviation of 30 μm) had a measured Q-factor of 40.Compared to circular hole structures, the diamond-shaped hole structure can achieve stronger bound fields due to its sharp corners. The proposed metasurface with polarization-independent quasi-BICs has enormous practical application potential in real-time chemical and biomolecular sensing.

A wideband terahertz metamaterial absorber is proposed based on the phase transition properties of VO2, which consists of two VO2 pattern layers, two polyimide dielectric layers and a metal reflector layer. The absorption characteristics, electric field distribution and tunability of the absorber are simulated and analyzed, and the results show that the bandwidth of the designed absorber with an absorption rate of more than 90% is 2.56 THz. The absorber stacks two layers of periodic cells with the same structure but different sizes, effectively extending the bandwidth. Moreover, by controlling the phase transition of VO2 from the insulating state to the metallic state, the absorptivity can be continuously adjusted. In addition, by studying the absorption performance of the designed metamaterial absorber under different polarization angles and incident angles, it is found that the absorber is polarization independent and absorption insensitive at large incident angles. The designed absorber is expected to be widely used in areas of terahertz communication, imaging and detector.

In view of the high requirements of electronic countermeasures signal analysis, the large amount of intelligence processing data, and the difficulty of multi-dimensional information analysis and extraction in complex electromagnetic environment, the big data processing method of cross-industry data mining standard process, namely Cross-Industry Standard Process for Data Mining(CRISP-DM), is studied, and an processing and analysis platform for historical data of offline electronic countermeasures electromagnetic signal is designed based on the characteristics of the massive historical reconnaissance data of the electronic countermeasures intelligence system and the existing electromagnetic target knowledge base. By exploring the technical path of mining electromagnetic target parameters, time rules of the target, multi-target association rules and other intelligence analysis, the application of analysis methods such as clustering, the mining of time series and association rules in the processing of electronic countermeasure electromagnetic signal historical data is realized. The information features of unknown electronic target such as clustering, target quantity and scale prediction, multi-target association and cooccurrence rule are analyzed and obtained. The results show that the characteristics and correlation laws of electromagnetic target parameters are obvious, and the fitting correlation degree of target time characteristics reaches 0.825. This work lays a foundation for further research and practical application.

To address the recognition challenge of Multi-h Continuous Phase Modulation (Multi-h CPM) signals with varying modulation parameters, this paper proposes a modulation recognition algorithm grounded in fuzzy entropy theory. This theory transcends the binary approach of distance and count-based similarity in approximate entropy, opting for a membership function to assess similarity and more accurately reflect the complexity of time series. The algorithm separates and calculates the fuzzy entropy of the in-phase and quadrature components of the received signal, utilizing these values as classification features for a Support Vector Machine(SVM). Experiments demonstrate that the algorithm achieves 100% recognition accuracy for full-response rectangular shaped Multi-h CPM signals across various modulation index sets at signal-to-noise ratios above 6 dB, and enables modulation recognition with a minimal number of symbols.

Addressing the issues of low precision in short baseline time-difference direction finding,ambiguity in long baseline phase-difference direction finding, and the difficulty of traditional microwave technologies to meet the long-distance transmission requirements for distributed deployment of electronic warfare equipment, a distributed long-distance single-baseline high-precision direction finding technology based on microwave photonics combining time-difference and phase-difference is proposed. This technology employs direct modulation mode microwave photonic transmission equipment,overcoming the impact of time-difference and phase-difference variations caused by long-distance transmission of microwave signals, ensuring stable transmission of microwave photonics over long distances. Furthermore, an improved time-difference surface fitting interpolation method is adopted to resolve the phase-difference ambiguity issue. Simulations and experimental tests have verified the feasibility of this technology for implementing high-precision distributed long-distance single-baseline direction finding.

In the uplink grant-free access scenario of the satellite Internet of Things(IoT), users directly send data to the satellite without the need for dynamic resource allocation, which makes it difficult for the receiving end to distinguish active users in advance. To recover the transmitted data,Active User Detection(AUD) is required. Currently, code domain methods are commonly used for active user identification, but the density of IoT devices far exceeds that of mobile communication users.Relying solely on code domain identification can lead to a dramatic increase in the dimension of address codes, reducing frame efficiency. Moreover, many collection and monitoring IoT terminals have weak capabilities and cannot report their own positions, resulting in the loss of this dimension of access auxiliary information, which is not conducive to improving the performance of active user detection.Aiming at the future development trend of satellite-borne phased array, this paper introduces the concept of spatial domain identification. By cascading identification in the spatial and code domains, a multi-domain joint method is used for active user detection. Simulation results show that the proposed method can enhance the identification capability when there is an address code conflict.

In response to the issues with existing 3D point cloud object recognition algorithms based on deep learning methods, such as the lack of feature interaction between points in multi-layer perceptrons, reliance on Euclidean distance between point clouds, and failure to consider the correlation at the feature channel level, we propose an attention mechanism-based 3D point cloud(PAttenCls) object recognition algorithm. The spatial attention mechanism based on points is employed to explore the attention values between points, achieving adaptive neighborhood selection for point clouds; meanwhile,the channel attention mechanism based on points adaptively assigns weights to feature channels,enabling feature enhancement. Additionally, a geometric uniformization module is added to the network to address the different feature patterns of different local regions' geometric structures. The proposed algorithm achieves a recognition accuracy of 93.2% on the ModelNet40 dataset and an accuracy of 80.9%on the most difficult subset of the ScanObjectNN dataset, and its effectiveness is verified on real-world data. Experiments have proven that the proposed algorithm can better extract feature information from point clouds, making the point cloud recognition results more accurate.

In response to the development needs of miniaturization and integration of satellite payload technology, as well as the bottleneck issues of high power and low passive intermodulation in space applications, this paper proposes a design method for a TE101 mode waveguide filter that is easy to 3D print, based on the characteristics of integrated forming technology of 3D printing and filter synthesis theory. A waveguide filter in the Ka band was designed and processed using 3D printing technology, with the entire filter structure formed in one piece. The measured and simulated results are in good agreement; the weight is 70% of that of the traditional screw fastening scheme of the same size; the filter's return loss is better than 22 dB, the in-band insertion loss is better than 0.3 dB, and the equivalent Q value before plating is about 1 100. Simulation results show that this filter has characteristics such as single-mode operation bandwidth and large power capacity; at the same time, the filter structure is suitable for integrated forming by 3D printing, eliminating the need for assembly and adjustment, and avoiding the nonlinear contact caused by the connection between the traditional filter cavity and the cover plate. The measured passive intermodulation level of the filter is lower than -130 dBm,which verifies the accuracy and feasibility of 3D printing technology for filter processing. It is of great significance for the integrated design and engineering application of 3D printed broadband filters.

The interlayer vertical transition structure is usually utilized to realize the miniaturization and high-density design of module structure, and plays an indispensable role in millimeter wave circuit wiring. In this paper, a vertical transition structure from microstrip to stripline is studied, which can be used in Q-band. A "water droplet" matching structure is adopted at the end of microstrip line, and the vertical transmission of Radio Frequency(RF) signals is carried out in the form of metal through holes connecting transmission lines. The three-dimensional High-Frequency electromagnetic Simulation Software(HFSS) is employed to establish the structure and compare the performance with that of other similar structures. The optimization results show that the reflection coefficient of this structure is less than -20 dB and the in-band insertion loss is better than 0.3 dB in 42~47 GHz, which can meet the requirements of engineering application. Compared with other vertical transition structures of the same type, it reduces the occupied area and maintains good transmission performance.

In response to the increasingly prominent issues of electromagnetic spectrum usage safety for communication and radar equipment, as well as the difficulty in quantifying the risk assessment of electromagnetic spectrum parameter leakage, this paper employs numerical and analytical methods to model the electromagnetic characteristic parameters of emission features, reception sensitivity, and antenna radiation characteristics for both friendly and adversary frequency-dependent equipments. By integrating the complete communication link structure of electromagnetic spectrum parameter leakage,the analysis is conducted from the perspectives of equipment effectiveness and the impact on the radio wave environment. Utilizing the concept of grid division, a method for calculating the probability of electromagnetic spectrum parameter leakage is proposed, and the calculation method is verified through simulation. The leakage probabilities for a ground-based fixed radar L1 and a 1 000-hop transmission station M1 within the ranges of 300~450 km and 350~450 km, respectively, are calculated. This achieves a quantified assessment of the risk of electromagnetic spectrum parameter leakage, solving the problem of difficult risk quantification in electromagnetic spectrum parameter leakage assessment.

A four-channel X-band chip transceiver System in a Package(SiP) module was realized by integrating multiple Monolithic Microwave Integrated Circuits(MMICs) and passive power division network into a very compact volume using silicon-based 3D heterogeneous integration technology. The module is Package on Package(PoP) stacked by two silicon-based packages, and the different packages are interconnected through the Ball Grid Array(BGA) mode. Chips are installed on the upper and lower surfaces of the inner cavity of a single package. The internal package is vertically interconnected using Through Silicon Via(TSV) and the module dimensions are 14 mm×14 mm×3.2 mm. The test results show that within 8~12 GHz, the transmission saturation output power of the module is ≥30.5 dBm, the receiving gain is ≥24.5 dB, the noise figure is ≤3 dB, the receiving input P-1 is ≥-26 dBm, and it also features 6-bit digital phase shifting and 6-bit digital attenuation capabilities, weighs approximately 1 g.It can be widely used in microwave transceiver system.

Millimeter-wave and sub-millimeter-wave detectors, when operating in geostationary orbit, are influenced by environmental thermal radiation in the performance of their subsystems. Based on in-orbit temperature field data, finite element analysis is employed to simulate the thermal deformation of the quasi-optical feed network; the GRASP software is adopted to fit the positional and surface parameters after the device's thermal deformation, and to simulate the changes in electrical performance when the quasi-optical feed network is subjected to on-orbit environmental thermal radiation. The results show that environmental thermal radiation has a certain impact on the electrical performance of the quasi-optical feed network, and by adopting temperature control measures, it is possible to avoid performance changes caused by temperature variations in orbit.

Composite materials are widely used in general aircraft, with usage rates exceeding 90% in various types of aircrafts. Therefore, defects in composite materials can severely affect flight safety.This paper designs an accurate defect detection method for aircraft composite materials based on infrared imaging detection technology, as well as a prototype non-destructive testing system. The non-destructive testing system consists of a data acquisition system, signal source group, infrared thermal imager group,and infrared image processing reconstruction system, which can identify the type and planar size of defects in the aircraft composite material structure and accurately locate the depth. A light-excited infrared imaging detection system for aircraft composite materials was built, and detection experimentswere conducted on composite materials at different temperatures, with research results provided. Theseresearch results will expand the application scenarios of domestic infrared non-destructive testing technology, and the commercialization of research findings will be beneficial in solving maintenance and repair issues during the general aviation of aircraft.

Fire disaster can cause great harm to the safety of people and property, and how to detect flame intelligently and efficiently is of great significance. In order to achieve accurate flame recognition under high space conditions, an infrared camera with two degrees of freedom that can detect environmental conditions in all directions is designed, and the target detection algorithm YOLOv5 is improved combined with deep learning. The K-Means clustering algorithm is employed to obtain nine width and height dimensions of clustering center by matching and replace the original network anchor parameters. Considering the relative proportion of the target frame, the loss function is optimized and applied to the Raspberry Pi to achieve flame recognition. The test results show that it takes 2.9 s for the improved YOLOv5 algorithm to detect a single sheet on the Raspberry Pi, which is less than that for the original YOLOv5 algorithm by 78%. The accuracy of the system is 100%, and the confidence of identifying the target frame is above 0.9. The proposed system can identify the flame fast and accurately.

Human pose estimation primarily relies on capturing joint points from visual image information to obtain global posture information of limbs and torso. Currently, depth learning methods based on visible light have high detection accuracy, but the risk of privacy leakage limits their practical application. Infrared detectors of the same cost can highlight human targets more effectively, but due to their lower imaging resolution and poor image quality, the detection accuracy is reduced. Inspired by visual Transformers, this paper introduces MobileViT-FPN to extract key human body points, using MobileViT to capture the relationship between local and global joint features, and then using Fixed Pattern Noise (FPN) to aggregate these representational information at multiple scales. Combined with an improved OpenPose for key point clustering, the estimated results are outputted. In the key point cascading phase, the attention mechanism allows the model to adaptively focus on the area of interest,enhancing the recovery of occluded parts. Experiments show that this method can real-time detect infrared human targets with varying scales and partial occlusions, accurately depicting human posture.

With the development of the 5th Generation Mobile Communication Technology(5G),various application scenarios continue to emerge. Network slicing can construct multiple logically independent virtual networks on a common physical network to meet the diverse service requirements of mobile communication networks. In order to enhance the ability of mobile communication networks to allocate resources on demand according to the traffic of each slice, this paper proposes a network slicing resource management algorithm based on deep reinforcement learning. The algorithm uses two Long Short-Term Memory(LSTM) networks to predict statistical data that cannot be reached in real time, and extracts dynamic characteristics of business data volume caused by user mobility, and then makes bandwidth allocation decisions that match the needs of slice services in combination with the Advantage Actor-Critic(A2C) algorithm. Experimental results show that compared with existing methods, this algorithm can improve the spectral efficiency by about 7.7% while ensuring the user's delay and rate requirements.

After acquiring the power grid dispatching signals, traditional deep confidence identification systems are mostly used for anomaly data extraction, which can only obtain the anomaly information parameters contained in low-dimensional data, resulting in a lower Area Under the Curve (AUC) value of the final data extraction result. Therefore, in order to improve the AUC value of the anomaly data extraction results of the power grid dispatching signals, an anomaly data extraction method for power grid dispatching signals based on data mining algorithms is proposed. The power griddispatching signals are processed using the Independent Component Analysis(ICA) algorithm to remove noise from the signals. The denoised signals are then subjected to wavelet decomposition to obtain multiple sub-signal datasets. Clustering algorithms in data mining algorithms are employed to analyze the sub-signal datasets to obtain the characteristics of the data samples, and data feature classification is completed considering the attribute feature density index to obtain the anomaly data characteristics.Finally, with the assistance of the Support Vector Data Description(SVDD), the abnormal data in the power grid dispatching signals are detected, and summarizing this part of the data can complete the anomaly data extraction. The experimental results show that the AUC value of the anomaly data extraction results obtained after applying the proposed method is always greater than 0.85, proving its superior application effect.