Abstract：Radar Cross Section (RCS) is a key physical quantity that measures the ability of a target to scatter electromagnetic waves when illuminated by radar waves. It directly determines the visibility of the target under radar detection and is of great significance for the development of stealth equipment and anti-stealth technology. This paper reviews the principles, method classifications, technical characteristics, and development trends of RCS measurement technology. It elaborates on the basic definition and influencing factors of RCS, including target shape, material properties, and radar frequency bands. It then focuses on analyzing four mainstream testing methods: outdoor far-field testing, indoor compact range testing, indoor near-field testing, and dynamic testing, discussing their principles, advantages, and limitations. Furthermore, it points out the current challenges in RCS testing, including testing efficiency, multi-physical fields, extremely low RCS measurement, complex electromagnetic environment simulation56, and expansion into new frequency bands. Finally, it prospects the future directions of technological development. This paper can serve as a reference for the research and application of RCS testing technology.

Abstract：As a core task in the field of 3D vision, point cloud generation plays a vital role in scenarios such as point cloud shape completion, upsampling, and synthesis. It is widely used in key application areas including autonomous driving, robot navigation, and medical imaging. Due to the inherent unordered nature, sparsity, and structural complexity of point cloud data, traditional geometric modeling methods struggle to efficiently generate high-quality and diverse point cloud samples. In recent years, point cloud generation techniques based on deep generative models have developed rapidly and become a research hotspot, significantly improving the quality and efficiency of point cloud generation. This paper reviews the latest progress and current challenges in point cloud generation algorithms based on deep generative models and discusses potential future research directions.

Abstract：Based on Fan-Out Wafer-Level Packaging (FOWLP) technology, a compact X-band quad-channel transceiver module was designed and fabricated. This module heterogeneously integrates chips manufactured using Complementary Metal-Oxide Semiconductor (CMOS) and Gallium Arsenide (GaAs) processes, enabling functionalities such as low-noise amplification, amplitude-phase control, and power amplification output for received RF signals. Within the module, interconnections and fan-out among chips are achieved through redistribution layers (RDL), while vertical connections to the printed circuit board (PCB) are realized via ball grid arrays (BGA). The final module dimensions are a mere 12.8 mm × 10.4 mm × 0.5 mm. Test results indicate that the module achieves a reception gain ≥27 dB, noise figure≤4 dB, transmission gain≥26.7 dB, and saturated output power ≥24 dBm. Furthermore, the RMS accuracy of the 4-bit digital attenuation≤1 dB, and the RMS accuracy of the 6-bit digital phase shifter≤6°, all meeting design expectations. By consolidating the characteristics of two distinct material chips on a high-density resin-based packaging platform, this module demonstrates superior performance.

Abstract：Industry 4.0 has an unprecedented demand for the use of industrial Ethernet, and the security problem of Ethernet data transmission is also increasingly prominent. Based on TMS320C6657 Gigabit Ethernet and AES encryption algorithm, this paper designs a data encryption transmission system for Gigabit Ethernet. Use TMS320C6657 Gigabit Ethernet interface and off-chip PHY chip to complete the design of Gigabit Ethernet hardware interface, and complete the design of communication interface software in combination with NDK network development package. The encryption efficiency of AES algorithm in TMS320C6657 is studied, and the efficiency optimization of gigabit Ethernet data encryption transmission system is completed through the on-chip IPC module. Experiments show that the design can effectively improve the security of gigabit Ethernet data transmission.

Abstract：This paper demonstrates a design of high integrated and miniaturized multi-core general-purpose control computer module, which can be implemented through the system-in-packaging (SiP) technology. This module integrates nine large-scale integrated circuits, including multi-core DSP, FPGA, SDRAM, and Flash. It employs a high-temperature co-fired ceramic (HTCC) tube shell and utilizes a double-sided dual-cavity layout packaging design technology to achieve miniaturization, generalization, and high-density integration of complex system circuits. Additionally, to address issues such as the integrity of high-speed SerDes circuit signals, the difficulty of heat dissipation from high-power active devices, and the integrity of large-scale circuit power supplies within the module, a high-reliability design was achieved through comprehensive simulation and optimization methods covering mechanical, electrical, and thermal aspects. Small-scale trial production and functional performance testing have also been completed. The product measures 39 mm×39 mm×6.49 mm and weighs 29 g. Compared to a system composed of discrete components, the volume and weight are reduced by over 80%. At ambient temperature(25°C) and high temperature(105°C), the DSP and FPGA functions within the microsystem module operate stably, memory read/write operations function normally, and the high-speed communication rate can reach 5 Gb/s. All functions and performance indicators meet the design expectations. The work presented in this manuscript can provide ideas and methods for the miniaturization, high integration, and high-reliability design of similar system-level packaging products.

Abstract：To address the technical challenge of excessive network timebase tracking error caused by 32 kHz clock temperature drift in the idle state of multi-mode baseband chips, this paper proposes a clock control scheme integrating high-frequency clock extended calibration, two-stage PLL cascade architecture, and hardware automatic switching mechanism. The high-frequency calibration unit collaborates with the Time Processing Unit (TPU) and Low Power Mode (LPM) module to maintain the operation of high-frequency clocks in the idle state for dynamic calibration of 32 kHz clock temperature drift. Meanwhile, the two-stage PLL cascade is utilized to generate high-precision reference clocks, and hardware circuits are employed to achieve glitch-free switching between high and low-frequency clocks.

Abstract：Event detection systems based on video analysis have been successful in many fields, but there are still issues such as missed detections, false positives, and insufficient accuracy in the field of highway anomaly detection. To address these issues, this paper improves YOLOv8 and uses artificial intelligence video analysis technology to propose a new Highway-YOLOv8 model and construct an intelligent video analysis system that can be used for highway anomaly detection. First, this paper designs a novel Deep Channel-by-Space Attention Mechanism (DCSM), which leverages channel and spatial interaction information to effectively enhance the model's field-of-view perception capabilities. Second, to address the loss of small object information in deep convolutional layers, this paper introduces a Multi-Stage Fusion Mechanism (MSFM), which significantly improves the model's small object detection capabilities. Finally, this paper adopts the advanced Wise-IoU loss function to replace the original YOLOv8 loss function, which effectively accelerates model convergence and improves detection accuracy. Experimental results show that Highway-YOLOv8 achieves a 2% improvement in mAP across all categories on the highway anomaly dataset compared to the original YOLOv8, with the highest improvement reaching 5% for single categories such as vehicles. This not only significantly enhances detection accuracy but also effectively reduces false negatives and false positives.

Abstract：This study aims to comprehensively identify the network security risks of multimodal large language models, construct a security evaluation system, and propose response strategies to enhance their security and reliability. The research method first analyzes the risk categories, constructs a risk identification model through multimodal data fusion, and uses the Analytic Hierarchy Process and Fuzzy Comprehensive Evaluation Method to construct a security evaluation system. Then, the effectiveness and stability of the model and system are verified through experiments. The research results indicate that the risk identification model can accurately identify risks, and the security evaluation system can reasonably assess the security situation. This study provides innovative ideas for the security risk identification of multimodal large language models. The constructed model and system can significantly enhance the risk identification ability, providing theoretical support for the security research of multimodal large language models and promoting the safe development of large language models.

Abstract：In this paper, a study is conducted on the shortened Reed-Muller (RM) code used in TETRA trunking system, and an improved algorithm based on majority-logic decoding algorithm is proposed. Firstly, the received 30 soft information bits are sorted in descending order, and the sign of the last few bits of the sorted sequence with low reliability are reversed in turn to obtain many new sequences. Then, the majority-logic decoding algorithm is used to decode the new sequences. Finally, the cost function of the decoding results are calculated, and the one with the smallest cost function is selected as the final decoding result. The algorithm improves the impact of low reliability symbols in received data on decoding results and enhances decoding performance. The algorithm proposed in this paper can approach the maximum likelihood decoding effect, and has good performance even at low SNR ratios.

Abstract：With the widespread application of drone technology in power grid operation and inspection, image compression transmission technology is becoming increasingly important. To solve the problems of insufficient computing power in drone image processing, long compression time and poor performance of inspection images, a drone inspection image compression method based on asymmetric encoding is proposed. Firstly, by removing redundant image space and quantifying image colors, the compression rate of drone inspection images has been effectively improved. Secondly, by compressing the model and adapting it with lightweight design, the computational requirements for compressing inspection images of power lines can be reduced. On this basis, the peak signal-to-noise ratio of the image was improved through high-quality image restoration enhancement. Finally, the method proposed in this paper was simulated and applied in a certain provincial electric power science research institute, with a peak signal-to-noise ratio of 34.2 dB and an average compression and recovery time of 4 632 ms. The proposed method can effectively improve the compression speed of unmanned aerial vehicle inspection images and enhance the quality of restored images.

Abstract：Aiming at the contradiction between the convergence speed and the steady state error of the fixed step size least mean square(LMS) algorithm, a novel variable step-size algorithm is proposed. The algorithm is based on the Sigmoid function and establishes a new type of functional relationship between the step size factor and the error signal. In the initial stage of the algorithm iteration, the step factor is larger and the convergence speed is fast, and after the convergence, a small step factor is used and the steady state error is small. The effect of the introduced control parameters on the change of the algorithm step-size is analyzed and compared with the convergence curve of the classical variable step-size algorithm. The experimental results show that the mean square error of the improved algorithm at a signal-to-noise ratio of 30 dB is reduced by 1.9 dB and 0.4 dB in the mean noise reduction (MNR) compared with the classical variable step-size algorithm.The number of iterations required for the convergence of the algorithm to the steady state is 238, 276, and 329 at signal-to-noise ratios of 10 dB, 20 dB, and 30 dB respectively.

Abstract：Existing Fuzz tools for browser JavaScript engines are insufficient in detecting non-crashing vulnerabilities. To address this issue, an enhanced vulnerability detection method combining memory detection and large language models (LLMs) is proposed. This method improves the Fuzz tool's ability to identify potential vulnerabilities and reduces the false-negative rate. Additionally, the integration of LLMs with conventional mutators generates more complex and targeted test cases under the stimulus of Proof-of-Concept (PoC) sets, effectively exploring deeper code paths and increasing code coverage. Experimental results show that the improved tool significantly enhances branch coverage and vulnerability detection capabilities compared to existing tools.

Abstract：Due to the low efficiency, easy omission or misjudgment of manual inspection, automatic drawing review based on BIM models has received increasing attention, and how to automatically calculate building size is one of the key issues. In this paper, a new method for calculating key building size is proposed. Firstly, through intelligent clustering analysis of the axis grid in the Revit model, the geometric relationships and spatial layouts of various building units in complex building structures are automatically identified and organized. Then, the relevant elevations of the specified building are filtered out through regular expressions and the type is determined. Finally, by comparing the area of the roof layer and the roof of the computer room, the accurate building height can be obtained. The testing in actual engineering scenarios on the Revit platform verifies that the proposed method can enhance the consistency and accuracy of BIM model data.

Abstract：This study proposes a voting-mechanism-based anomaly detection technology to address cybersecurity challenges in power monitoring systems. By integrating machine learning models, deep learning models, and rule-based detection systems with majority/weighted voting mechanisms, the system enables multi-perspective collaborative analysis of industrial control network traffic. Dynamic weight adjustment optimizes model performance, while real-time detection (millisecond-level response) and feedback mechanisms enhance robustness. Experimental results demonstrate significant improvements: attack detection accuracy increased from 85% to over 93%, false positive rate reduced to below 5%, and false negative rate lowered to 9%, outperforming traditional single-model approaches. The technology effectively mitigates power outages caused by cyberattacks and reduces operational costs, accelerating the digital transformation of the power industry. Future work will integrate federated learning and edge computing to extend applications to distributed energy systems, offering comprehensive solutions for smart grid security.

Abstract：Aiming at the problem that the 230 MHz RF signal cannot be fully covered due to attenuation, an RF power amplifier is designed to further improve the output power. Its core is RD70HVFIV power amplifier tube. It mainly includes DC power supply bias network, input and output impedance matching network and gain stabilization network, among which impedance matching network adopts multi-capacitor parallel mode, which is convenient for engineering debugging and correction of design deviation. The detailed design process is given and measured. The measurement results show that the linear gain of 10 dB can be achieved in the commonly used 1~5 W input power range, and the gain is not linearly increased in the 5~10 W input power range, and the limit output power is about 65 W. After engineering testing, the power amplifier designed in this paper can effectively solve the problem that the signal can not be covered due to signal attenuation.

Abstract：With the rapid development of electronic technology, the demand for ultrashort wave and wideband amplifier module in electronic surveillance equipment is increasing day by day. This article designs a high amplitude phase consistency ultrashort wave and wideband amplifier module based on practical needs. Impedance transformation scheme is adopted to achieve the impedance of two input terminals from 200 Ω to 50 Ω. Small signal amplification and one channel RF output are achieved through amplification circuits and balanced to unbalanced conversion. In addition, gas discharge tubes and TVS tubes are used to achieve lightning protection design. The test results show that the gain of this amplifier module is more than 15 dB and the noise figure is less than 2.9 dB in the band of 570 MHz under the conditions of 15 V working voltage. Through the process control, this amplifier module's phase consistency is less than ±3°, amplitude consistency is less than ±0.5 dB, and has the advantages of small size, light weight, low failure repair rate, etc. It is in the leading position of technology in the domestic similar products, and widely used in the detection equipment already.

Abstract：In order to solve the technical problem of poor antenna isolation between adjacent frequency bands in BeiDou third-generation satellite navigation scene, this paper proposes a dual-band dual-polarized BeiDou transceiver antenna with high isolation, which can effectively solve the problem that BDS B1 band is susceptible to interference when BDS L band and BDS B1 band work at the same time. By introducing a broadband high isolation coupler, the high isolation of the antenna radiation patch feed port is realized. Secondly, the compact coupling feeding method is used to realize the miniaturization design of the antenna. Finally, the dual-band dual-polarized BeiDou antenna is verified. The measured isolation between the antennas is 37.8 dB. The test results are in good agreement with the design results, which verifies the effectiveness of the design and can meet the multi-frequency application requirements of the BeiDou third-generation satellite navigation antenna.

Abstract：A scheme for dynamically adjusting the output voltage of a switching power supply is proposed to address the shortcomings of the current power supply system's inability to dynamically adjust the voltage online. This solution is compatible with multiple wired communication protocols and is controlled by the system. During operation, the output voltage of the switching power supply is dynamically adjusted to meet the power supply requirements of the load. Simultaneously, this solution achieves synchronous monitoring of output voltage and transmits it back to the system's upper computer, thereby realizing online adjustment and monitoring of output voltage.

Abstract：It is important for point-to-point communication technology to control flow. Regarding the lower transmission rate of communication channels in shortwave facsimile communication, the stop-and-wait flow control and protocol flow control methods are respectively adopted in the shortwave facsimile communication module to achieve the communication flow control between the main controller and the Modem, as well as between the main controller and the shortwave waveform module. This effectively solves the flow control methods between different channels and realizes reliable and stable facsimile communication in the shortwave radio facsimile communication module.

Abstract：As aviation navigation systems grow in complexity, fault diagnosis becomes crucial for ensuring stable system operation. This paper presents a fault location and visualization system for the Thales DVOR 4000 navigation system, based on Paddle OCR. To enhance data quality, the system employs IQR and Z-Score methods for data cleaning and to build a baseline data model. It also uses Python and the PyQt5 UI library to create a visualization interface, enabling real-time fault information display and alerts. To address the high computational resource consumption of Paddle OCR, the paper introduces the multiprocessing library, achieving system multitasking via multithreading and multiprocessing parallelism. This significantly boosts data acquisition efficiency and system performance. Experimental results show that the system has a data recognition accuracy of nearly 100%, and it excels in fault diagnosis accuracy and visualization effects. It can effectively improve diagnostic efficiency, and reduce operational and maintenance costs.

Abstract：In the realm of data link burst communication systems, the Round Trip Time (RTT) synchronization method is employed for time synchronization and ranging, with measurement accuracy being limited by the code rate and thus unable to be significantly enhanced. This paper introduces an enhancement technique that incorporates carrier phase measurements into the RTT ranging process to improve the accuracy of distance measurements. At the hardware level, this solution employs a zero-intermediate frequency (IF) transceiver architecture, shared local oscillators, and a coupled closed-loop system.At thesoftware algorithmic level, the measurement values are processed in a comprehensive manner, and the cycle slip is resolved to enhance the measurement accuracy. Through the application of these technologies, while maintaining the fundamental process of RTT-based ranging, they overcame the need for continuous tracking of carrier phase measurements and utilized the high resolution capability of carrier phase measurements to enhance measurement accuracy. This method has been validated through laboratory prototype testing, demonstrating a significant improvement in measurement accuracy, approximately two orders of magnitude, which offers a viable solution for high-precision ranging applications in burst communication systems.

Abstract：During the construction of airborne Synthetic Aperture Radar (SAR) hardware systems, amplitude-phase errors are inevitably introduced in the transmit-receive channel links, which deteriorate the actual imaging performance of the SAR system. To address this issue, an amplitude-phase error compensation scheme for airborne SAR echo data based on internal calibration is designed, and imaging processing is completed in combination with an improved autofocus imaging algorithm. This scheme accurately estimates the amplitude-phase error values based on the reference, transmit, and receive calibration data in the internal calibration design, and compensates the echo data according to the amplitude-frequency and phase-frequency characteristics. Meanwhile, the traditional autofocus algorithm is improved by introducing quadratic phase terms, which enhances the estimation accuracy of high-order azimuth phase errors and optimizes the two-dimensional focusing effect of images. The processing results of measured data demonstrate that, after applying the proposed scheme, the range resolution improves by 6.6% and the azimuth resolution increases by 25.9%. Moreover, the ISLR meets the specified requirements of -10 dB and PSLR meets the specified requirements of -13 dB. Through the analysis of amplitude-phase error compensation and the comparative study of different imaging schemes, this research effectively enhances the image focusing accuracy, offering crucial guidance for the engineering implementation and practical applications of airborne SAR systems.