Electromagnetic calculation and SAR echo simulation algorithm play an important role in radar system design, radar algorithm verification, and automatic target recognition algorithms research. To solve the problem of traditional electromagnetic calculation and SAR echo simulation algorithm being too time-consuming, an acceleration method for the entire process based on GPU is proposed. Firstly, the three core computing parts of the traditional algorithmray tracing, electromagnetic calculation,and SAR echo simulation are divided into multiple threads. Then, by using the GPU general computing platform, all thread units are processed in parallel to accelerate the simulation speed. Finally, the experimental results show that the acceleration ratio of this method is about 450 compared with singlecore CPU and about 50 compared to a 10-core CPU. This method achieves a good acceleration effect on the premise that the generated data is consistent with the traditional algorithm.

This article analyzes and verifies the reliability of high repetition rate narrow pulse width multi-wavelength lasers. A block diagram method model of high repetition rate narrow pulse width multi-wavelength lasers was established, and reliability analysis of the laser was conducted at the initial design stage. The reliability results of each unit of the laser were quantitatively calculated, including failure rate λ and Mean Time Between Failures (MTBF), with an expected MTBF of the entire laser system being 1 798.8 hours. To meet the design requirements of the laser, reliability optimization of the laser was achieved through selecting high-quality crystals, first-level derating design of semiconductor lasers, and redundant design of power control. After optimization, the MTBF of the entire laser system was increased to 2 260.9 hours. The reliability optimization design results of the laser system were validated, and the validation results showed that the laser's MTBF could reach 2 400 hours.

In response to the structural dynamics optimization design requirements of a certain satellite-borne precipitation radar, based on finite element analysis, the Kriging surrogate model is used to approximate the finite element model of the satellite-borne precipitation radar. The improved genetic algorithm is then employed to optimize the honeycomb panel structure parameters,resulting in a better design solution. Finally, the effectiveness of this method is demonstrated through a case study of the structural dynamics optimization design of a certain satellite-borne precipitation radar and mechanical experiments.

In this paper, an improved Surface Acoustic Wave (SAW) temperature and pressure sensor with two chambers is proposed to solve the nonlinear coupling problem in the integrated structure suggested in earlier studies, and the design and experimental study of an acoustic surface wave all quartz pressure sensor is carried out. Based on the finite element method and perturbation theory, the response mechanism of the quartz-based SAW pressure sensor is analyzed, the coupled mode theory is used to optimize the design of the three-resonator-type sensitive element, and the glass paste bonding is used to realize the quartz cross-lead hermetic encapsulation, and the preparation of the SAW allquartz pressure sensor is realized. The test results show that the developed SAW all quartz pressure sensor has a pressure range of 0~500 kPa, a linearity of 0.415% FS, a pressure sensitivity of 551 kHz/MPa, and a temperature coefficient of sensitivity of 0.134% over the operating temperature range of 0 ℃~120 ℃. The development of this surface acoustic wave all quartz pressure sensor lays the foundation for the subsequent realization of wireless passive measurement of the sensor.

Quantum cascade technology, based on resonant tunneling and intersubband transitions in a multiple-quantum well or superlattice structure can generate light source and detect optical signals, it is the theoretical cornerstone of quantum cascade laser(QCL) and quantum cascade detector (QCD), which has wide application prospect in detection, remote sensing, communication, ra-dar, and other fields. After three decades of research, quantum cascade technology has made significant progress in basic research,product performance, application system research and scene testing. In this paper, the principle and development history of quantum cascade technology are briefly introduced first. Subsequently, the calculation approaches of intersubband energy level structure and electron transport dynamics in quantum cascade devices are elaborated. Next, the research progress of quantum cascade technology is mainly reviewed, including mid- and far-infrared high power QCL, mid-and far-infrared widely tunable QCL, terahertz QCL, high performance QCD, and single-chip photonic integration of QCL and QCD. Finally, the commercially available QCL and QCD products, as well as their application status, are introduced.

In the radar imaging equipment test, the traditional real scene test method is difficult to construct, the scene is limited, and the test risk is high, so it is urgent to solve the problems of insufficient testing and incomplete evaluation of the target recog-nition algorithm. Aiming at the existing problems, this paper designs a test system for the target recognition algorithm, which can provide the processing and labeling of SAR image and inverse SAR image, as well as the automatic operation, environment configu-ration and performance evaluation of target recognition algorithm. Compared with the traditional test method, the system has the ad-vantages of low cost, short test time, strong controllability and extensibility.

The traditional deep learning-based Polarimetric Synthetic Aperture Radar (PolSAR) feature classification method extracts image local features by stacking convolutional layers, which makes it difficult to establish long-range dependencies. It is not-ed that Transformer, a deep learning model based on a self-attention mechanism that captures global pixel-to-pixel correlations, has achieved success in image classification tasks. Meanwhile, the PolSAR feature classification task has demonstrated better classification results in the complex domain compared to the real domain. Therefore, Transformer is introduced into the complex domain, and a hybrid model of Transformer and Unet based on the complex domain (CT-Unet) is proposed for PolSAR feature classification. This model combines Transformer with CNN for feature extraction on PolSAR data of complex type. The experimental results of PolSAR feature classification using the Xi'an dataset and German dataset show that the proposed model can effectively improve the accuracy of PolSAR feature classification. Transformer is expected to make up for the shortcomings of convolutional neural net-works in the PolSAR feature classification task.

With the widespread use of network encryption protocols, traditional network traffic classification technology has been challenged. The current method has the following limitations: first, the model is highly dependent on the depth feature, which requires the labeled training data set to be large enough in scale, otherwise the model will have difficulty generalizing to new data;second, the model only focuses on one modal feature of traffic, and the feature differentiation of the same mode of traffic from differ-ent categories may not be obvious. To solve these problems, a deep learning-based encryption traffic classification model called Par-allel Transformer Net (PTNet) is proposed in this paper. Based on the semi-supervised idea of pre-training and fine-tuning, the mod-el makes full use of a large amount of unlabeled traffic data on the network for pre-training, and then fine-tunes on the basis of a small amount of labeled data. Additionally, the model extracts the flow characteristics of load and packet length sequences in parallel to carry out multi-mode feature fusion. Three different traffic classification tasks and their corresponding datasets (Android, USTC-TFC, and CSTNET-TLS1.3) show good results, with classification accuracies reaching 95%, 98%, and 97%, respectively.

This paper introduces a novel design for a miniaturized wide-stopband hairpin filter that aims to overcome the chal-lenges of parasitic passbands at the harmonic frequencies encountered in traditional microstrip hairpin filters. By incorporating three 1/4-wavelength open-circuit microstrip lines at the input and output of the conventional hairpin filter, the design effectively suppresses parasitic passbands. Additionally, the application of cross-coupling techniques facilitates the miniaturization of the device. The design, fabrication, and testing phases culminated in the development of a bandpass filter with a 400 MHz bandwidth and a center frequency of 3 GHz. The filter demonstrates an insertion loss of 2.5 dB and achieves effective suppression of parasitic passbands be-low 13 GHz, with an out-of-band suppression of 24 dB. The new filter design is characterized by its simplicity and low complexity,with dimensions of just 23 mm by 27.7 mm, meeting the current demands in the communications field for high-performance, minia-turized filters, and demonstrating significant potential for future applications.

Commercial IP core is currently used in the development of high-speed memory device to achieve date error correction,the maximum achievable encoding and decoding speed is 800Mbps， which can only rely on multiple IP cores working simulta-neously to meet the requirements of gigabit high-speed data access rate. In view of the bit error characteristics of spaceborne storage data, this paper proposes an improved RS decoding algorithm, which reduces the number of iterations and the amount of computation in the decoding process by downgrading the residual polynomial in the encoding algorithm and the syndrome polynomial in the decoding algorithm, and adopts the sub-term multiplexing technology for the finite field multiplier of the basic operation unit in the decoding algorithm. The implementation results indicate that the maximum speed the encoder and decoder can achieve 10.5 Gbps，while the encoder resources is reduced by 15% and decoder resources reduced by 40% compared to a single commercial IP core,which can meet the application needs of high-speed of memory platform.