Abstract：To address the problems of uneven heating and low energy utilization caused by standing wave distribution in traditional microwave ovens, this paper designs a new type of microwave oven based on the principle of artificial surface plasmon, which contains an H-plane waveguide horn antenna excitation device and a metal fold slow-wave structure. The core design parameters of the excitation device are determined, and an array-type metal fold structure is proposed. Simulation results show that this structure can excite uniform surface waves at a center frequency of 2.45 GHz, with a uniformity width of the transverse electric field of 210 mm and a longitudinal height of 150 mm in the cavity. The reflection coefficient is less than -10 dB in the working frequency band of 2.45~2.46 GHz. Uniform heating can be achieved under no-load and different load conditions. After the prototype was fabricated and tested, it was confirmed that this microwave oven effectively improved the standing wave defect of traditional microwave ovens. The energy efficiency of the prototype reached 52%, and the heating uniformity and energy utilization rate were good, providing a new solution and experimental basis for the structural design and performance optimization of new microwave ovens.

Abstract：To address the non-uniform heating issue in conventional microwave ovens caused by cavity standing waves, this paper proposes and validates a novel uniform heating scheme based on an array of dielectric surface waveguides fed by a horn antenna in free space. The scheme utilizes a horn antenna to excite a symmetrically arranged dielectric surface waveguide array, which in turn launches surface waves within the cavity. Full-wave simulations demonstrate that at 2.45 GHz, the proposed scheme significantly enhances heating uniformity, with its spatial coverage of uniform heating zones substantially surpassing that of conventional direct-feed configurations. Furthermore, simulations performed on a compact microwave oven model under both no-load and loaded conditions (with 1 000 mL of water) show a uniform electric field distribution inside the cavity, high energy absorption efficiency, and a voltage standing wave ratio (VSWR) below 3 across the operating frequency band. This study provides an innovative technical pathway characterized by structural simplicity, high power capacity, and ease of engineering integration for solving the problem of microwave heating non-uniformity.

Abstract：Portable telemetry devices, due to their exceptional mobility, ease of transportation, and the capability for temporary augmentation, provide robust supplementary support for telemetry tasks. Among the components of portable telemetry devices, the quality of antenna performance directly impacts the overall system′s performance and decisively determines the range of applications it can be used for.Most existing portable telemetry devices are equipped with directional antennas, which have a narrow beamwidth. This poses significant limitations when these systems are added to special scenarios such as mobile platforms at sea. To address this, an array of multiple antennas is proposed, utilizing multi-antenna signal synthesis techniques to design a phased-array antenna that offers hemispherical coverage.Simulated testing and practical application validation demonstrate that this antenna achieves coverage from 0° to 360° in azimuth and from 0° to 90° in elevation. The gain fluctuates less than 2 dB within the elevation range of 0° to 90°, and it exhibits a gain greater than 8 dBi at low elevations. This outstanding performance makes it highly suitable for receiving telemetry signals from high-speed mobile targets at sea, significantly broadening the application scope and practical effectiveness of portable telemetry devices.

Abstract：This paper proposes a miniaturized analog reflective phase shifter operating in the ISM band. Based on a 3 dB directional coupler as the core topology, the circuit integrates fan-shaped resonators at key nodes of the transmission line. By fully leveraging the capacitive compensation characteristic and space-compact advantage of the fan-shaped structure, a significant reduction in circuit size is achieved, with a 23 % miniaturization compared to the traditional structure. Measured results demonstrate that within the frequency band of 2.4 GHz~2.5 GHz, the phase shifter exhibits a stable phase shift of 173 °±2 °, an insertion loss better than 2.3 dB with a fluctuation of less than 1 dB, and a return loss greater than 15 dB across the entire tuning range. This device combines compact size with simplified structure, demonstrating excellent practical value in applications such as industrial sensing, smart home systems, and short-range communication.

Abstract：Traditional PCB design for RF filter circuits primarily focuses on impedance matching, with limited consideration given to the control of electromagnetic coupling paths, which restricts port isolation in filtering units. To address this, a method to achieve high isolation on PCBs based on a "field-circuit-shielding" synergistic design is proposed. This method integrates three core techniques: thinning the critical dielectric layer to enhance vertical field confinement; optimizing the coplanar ground spacing to suppress in-plane lateral coupling; and employing a high-density via array with the minimum process-achievable pitch to construct a 3D electromagnetic shielding wall. Verified through 3D electromagnetic simulation and physical measurements, the synergistic design approach improves the inter-port isolation by over 16 dB within the 1.6~2.1 GHz frequency band, without significant degradation in insertion loss. This method provides a quantitative engineering reference for the packaging and electromagnetic compatibility design of RF filtering modules and passive filters.

Abstract：In mobile communications, satellite communications, and radar systems, reducing signal interference has become a key objective. Flat-topped dielectric resonator antennas (DRAs) exhibit strong anti-interference capability, high design flexibility, rich resonant modes, and compact dimensions. Based on the principle of pattern superposition, both the low-order DRA mode and the slot mode are simultaneously excited to achieve a flat-topped radiation pattern in both the E- and H-planes. By attaching two substrates with relatively high permittivity to the DRA, the resonant frequency of the DRA mode is reduced. Additionally, metallic walls are incorporated around the substrate to confine the electromagnetic field propagation, thereby modifying the directivity and sidelobe levels. Ultimately, a compact dielectric resonator antenna operating at 3.0 GHz with a flat-topped radiation pattern has been designed. The radiation pattern of the proposed antenna demonstrates a flat-top characteristic in both the E- and H-planes. Compared with other flat-topped antennas, the proposed structure offers the advantages of simple configuration and compact size.

Abstract：To address the issues of large size, heavy weight, and complex structure of sum-and-difference power divider networks in active phased array antennas, a compact planar sum-and-difference power divider network based on microstrip-waveguide hybrid transmission line is proposed. The network employs four three-branch hybrids and four 1:8 Wilkinson power dividers to achieve 1:32 power distribution. The sum-and-difference comparator and power divider are designed in a coplanar configuration, operating in the Ku-band. Measured results demonstrate that the amplitude consistency (RMS) is better than 0.2 dB, phase consistency (RMS) is better than 4°, the null depth of difference beams in both elevation and azimuth planes is greater than 20 dB, and the voltage standing wave ratio (VSWR) of all ports is less than 1.8. The fabricated prototype features compact structure, small size , light weight , and ease of fabrication, making it suitable for airborne and missile-borne phased array antenna applications.

Abstract：Monte Carlo (MC) simulation is widely employed for the reliability analysis of circuits and systems. However, with advances in manufacturing technology, the number of process parameters affecting circuit performance has increased dramatically, and the runtime of a single circuit simulation has become increasingly expensive, resulting in excessively long overall computation times for standard MC methods. To address this issue, in this work, an improved fast MC algorithm is proposed to substantially accelerate circuit reliability evaluation. The method hierarchically extracts critical parameters and constructs a sparse regression model to replace costly circuit simulations, enabling rapid screening of samples that are likely to fail within the MC framework. Following this prescreening, the candidate samples predicted as potential failures are re-simulated to obtain an accurate count of failures, from which the circuit failure rate is estimated. To ensure that the predicted number of failing samples adequately covers the true failures, a ranking-based strategy is applied to the model predictions on a test set to determine a reasonable resimulation range. Experimental results on circuits implemented in 28 nm technology show that the proposed fast MC algorithm achieves a speedup of 1 146×~1 886× over the standard MC method without loss of accuracy. Compared with other fast MC approaches, the proposed method provides 2.6×~7.8× speed, demonstrating its effectiveness and practical value for advanced circuit reliability analysis.

Abstract：With the continuous improvement in the integration level and complexity of embedded chips, the bottlenecks of traditional testing methods in terms of security, flexibility, and testing efficiency have become increasingly prominent. Therefore, this paper takes a general-purpose Microcontroller Unit (MCU) as the research object, conducts an in-depth analysis of the core technologies of chip testing modes, and proposes a chip testing scheme that integrates high-security test mode access and hierarchical test data allocation. This scheme achieves secure access to the test mode through a 63-bit key authentication mechanism. Meanwhile, it constructs a hierarchical architecture consisting of IO direct control, register control, and Flash system configuration value control to ensure the flexible allocation of test data. The scheme can fully cover the testing requirements of multiple modules of the chip and provide reliable support for chip testing in high-security scenarios.

Abstract：Aiming at the problem that low output voltage of the existing laser power converter is difficult to meet the power supply requirements of electronic devices, this paper presents the design and fabrication of a six-junction GaAs laser power converter. Experimental results demonstrate that under 808 nm laser irradiation with an irradiance of 2.09 W/cm², the six-junction GaAs laser power converter achieves a photoelectric conversion efficiency of 57.87%, an open-circuit voltage of 6.72 V, and an output power of 1.21 W. These performance metrics meet the charging demands of miniature smart devices, including consumer electronics, electric-drive mobile platforms, and drones.

Abstract：With the wide application of RF signals in high-speed transmission fields, traditional discrete RF architectures are facing bottlenecks of resource and performance problems, while RFSoC with high integration and low power consumption has becoming a good solution. This paper focuses on the research of filter optimization of domestic RFSoC, and proposes a project for the filtering of high-speed data after the sampling of domestic RFSoC using SUPERDSP. Experiment has shown that this project has the same performance of traditional FIR filtering, but also a substantial reduction of hardware usage. It provides theoretical and practical references for the optimal design of domestic RF systems, and has great significance on engineering.

Abstract：In the irradiation scenario, data inversion may occur in the storage unit. To enhance data reliability, a correction mechanism based on dynamic self-refreshing and extended Hamming code is designed. This circuit consists of an encoding module, a decoding and error correction module, a data storage module, and a dynamic self-refreshing management module. It has the function of correcting one-bit errors and detecting two-bit errors, as well as low dynamic power consumption. The mechanism is implemented using Verilog hardware description language and a simulation verification system is built. Simulation results show that this error correction mechanism can effectively correct single-bit errors and detect and report double-bit errors, improving the reliability of the storage unit. This mechanism has been widely applied in the switching chips developed by our company and has passed the irradiation upset test, verifying its effectiveness.

Abstract：With the development of space detection satellite technology,it has become a reality for a single-satellite to integrate sound,light,radio and other detection means.Multiple method working together can yield better benefits. A typical application is to precisely match the target radiation signal with it orbit.This paper develops a space target track maching technology based on Kinematic principle. Using the photo data and the satellite almanac data,combined with radio signal data,the matching of the target radiation signal and target track is achieved.The test results show that the method solves the correlation problem between image target intelligence and signal target intelligence to a certain extent.

Abstract：Advanced Persistent Threat (APT) attacks pose severe risks to critical information infrastructure due to their high degree of stealth, persistence, and phased progression. The attack scenarios of APT constitute typical complex attack patterns. Traditional point-based detection methods struggle to effectively reconstruct the complete picture of such attacks from massive and heterogeneous event log sets. This paper proposes an APT attack scenario reconstruction method based on the fusion of event semantic correlations. First, a multi-level, semantic-enhanced event log correlation framework integrated with Large Language Model (LLM) capabilities is designed, by incorporating security entity and attack behavior semantics, it constructs a correlation structure from "attack short chains" to "attack long chains". Second, to address the uncertainty in mapping attack behaviors to attack phases, a Hidden Markov Model (HMM)-based attack phase inference model is constructed. By integrating event semantic information, this model maps attack long chains to standard kill chain phases, thereby identifying complete APT attack chains. Experimental results demonstrate that, on a simulated dataset containing noise and APT attacks, the proposed method can effectively identify complete and partially observed APT attack scenarios, significantly improving the accuracy and robustness of scenario reconstruction.

Abstract：RDMA network is gaining popularity for high performance communication of distributed applications. As virtualiza tion has become the de-facto layer of datacenters, RDMA network also needs to be virtualized. It is common for the datacenter to have virtual machines (VMs), containers and sometimes container-in-virtual-machine as its basic resource provisioning entity. However, RDMA is virtualized in different ways for VMs or containers. There is no coordination between the two resource users. To solve this problem, this paper proposes uniRDMA (unified virtual RDMA), which is a unified RDMA virtualization framework for both VMs and containers. Through a single point of management, RDMA resources could be flexibly allocated to VMs and containers. With uniRDMA, the server machine does not have to choose which virtualization form it is hosting, VMs or containers. In this evaluation, uniRDMA can be deployed in hybrid virtualization environments and achieve similar performance to those of SR-IOV and bare-metal RDMA for real applications.

Abstract：This study designed and implemented an integrated mask layout analysis software based on Python to enhance the efficiency of mask pattern analysis in semiconductor manufacturing. The software offers complete-layout and region-based analysis modes, supporting three light transmittance calculation schemes: direct computation, overlapping graphics merging, and grid partitioning-recalculation. Key functionalities include light transmittance calculation, manufacturing effect preview/saving, and dimensional measurement tools featuring both arbitrary-point measurement and endpoint snapping. Testing demonstrates that the software significantly reduces processing time while maintaining computational accuracy in light transmittance analysis, providing a reliable analytical solution for mask design and fabrication.

Abstract：The positioning technology inside aircraft cabins holds significant importance in fields such as navigation and positioning, material supervision, and emergency rescue. The positioning method based on Wi-Fi Received Signal Strength Indicator (RSSI) has become a key research direction in cabin positioning due to its advantages of easy accessibility and low cost. Among various approaches, the positioning method that combines RSSI with a distance attenuation model features simple operation and strong portability; however, there is still room for improvement in its positioning accuracy. This paper focuses on researching this specific method and proposes an Access Point (AP) layout optimization approach oriented to geometric figure factors. By analyzing the propagation relationship between distance measurement errors and Position Dilution of Precision (PDOP), an optimization model for positioning errors is established, and its analytical expression is derived. Furthermore, the improved Rosen gradient projection method is employed to obtain the optimal layouts under different numbers of APs within the rectangular cabin space. Experimental results demonstrate that the proposed optimization method can effectively reduce the impact of geometric figure factors, thereby enhancing positioning accuracy. This study provides a theoretical basis and optimization strategy for the deployment of in-cabin positioning systems.

Abstract：Addressing the high failure rate of intermediate frequency (IF) receivers based on analog devices, this paper focuses on analyzing the amplitude stability, anti-interference capability, and high-precision angle measurement performance brought by quadrature modulation，it also explains the functions of each module in the receiver and analyzes the fault diagnosis method of this system in equipment maintenance through theoretical derivation and system composition. The analysis shows that this design is one of the key technologies for modern missile seekers to achieve high-precision and high anti-interference target tracking, providing reference for the design of digital receivers and radar signal processing.

Abstract：This paper investigates a scenario where multiple low Earth orbit satellites equipped with integrated sensing and communication base stations collaboratively provide communication and sensing services to ground users. It is assumed that each low Earth orbit satellite base station is equipped with a uniform linear array antenna, enabling them to communicate with their respective ground users while cooperatively performing joint sensing of targets within the region of interest. In this context, the objective of this paper is to maximize the minimum detection probability in the target area by jointly designing the beamforming schemes of multiple low Earth orbit satellites, while ensuring the signal-to-interference-plus-noise ratio constraints at the ground users and the maximum transmit power constraints of the satellites. Due to the multi-antenna configuration of the satellites, this optimization problem is highly non-convex. To address this, the paper proposes a semidefinite relaxation-based optimization algorithm to obtain high-quality solutions. Finally, numerical simulations are conducted to validate the performance advantages of the proposed scheme compared to other baseline methods, demonstrating that the joint beamforming design effectively balances the trade-off between communication and sensing performance.

Abstract：To meet the demand for next-generation high-performance computing in the commercial space sector and improve the on-orbit reliability of large-scale CPUs, the paper takes the domestically produced high-performance server processor, Phytium S5000C-M, and conducts in-depth research on the boot security enhancement technology and dynamic state detection methods based on RAS under C2C mode, developing a dual-CPU-based processing system platform. The technical advantages of the S5000C-M processor in terms of autonomous control, high-performance computing, and security features are analyzed in detail. Additionally, designing boot security measures for the S5000C star-service computer system, covering hardware to software, including key technologies such as reliable boot zone construction, secure storage partitioning, and firmware integrity verification. Based on the RAS architecture, it completes the dynamic state monitoring design and verifies monitoring methods for key indicators such as processor core anomalies, memory errors, and bus transmission exceptions. The research holds significant engineering importance for the application of the new generation of high-performance processors in the field of commercial aerospace.

Abstract：Spacecraft star sensor image processing is a crucial component of Field Programmable Gate Array(FPGA) payload missions. This functionality significantly reduces the data transmission requirements of the payload, accelerates decision-making speed, and enhances the processing capability of the payload unit. Star sensor algorithms (e.g., star extraction, star pattern recognition, and attitude determination) play a vital role in improving the efficiency and real-time performance of on-board FPGA image processing. However, star sensor algorithms involve a large number of parameters and complex computational processes, consuming substantial storage and FPGA hardware resources. Traditional triple modular redundancy (TMR) reinforcement tools offer limited support for on-board algorithm modules, lack focus on critical aspects, and result in high resource consumption and significant performance degradation, making them inadequate for current mission requirements. This paper proposes an FPGA reinforcement technique for star sensor images based on high-level synthesis (HLS) constraints. HLS constraints enable system-level functional reinforcement of FPGAs, allowing for targeted reinforcement design based on the criticality of system and algorithm principles. This approach effectively reduces both the time and resource costs associated with reinforcing logically complex FPGAs.

Abstract：This paper proposes an improved silkworm health detection algorithm based on YOLOv11n (MSDA-YOLO) to address the issues of delayed disease recognition and insufficient detection accuracy in traditional silkworm farming. The core innovations of the algorithm include: 1) introducing the Multi-scale Grouped Dilated Convolution in the backbone network and combining it with C3k2 to propose the C3k2MS module, which enhances the model’s ability to extract features at different scales; 2) introducing the DySample dynamic upsampler for the first time in silkworm disease detection, which improves the model’s ability to perceive spatial details and maintain geometric structure for small targets; 3) using the Adaptive Spatial Feature Fusion strategy to modify YOLOv11’s detection head, optimizing multi-scale feature fusion and suppressing feature conflicts between scales. Experimental results show that MSDA-YOLO achieves superior performance, with a precision of 87.0%, recall of 77.8%, mAP@0.5 of 87.3%, and mAP@0.5:0.95 of 52.8%, surpassing the baseline YOLOv11n and other mainstream YOLO models, demonstrating its clear advantage in silkworm health detection.

Abstract：Image caption generation aims to generate semantically coherent text descriptions for images. Existing dynamic networks are mostly limited to adjusting convolutional kernels, lacking spatial multi-scale and channel-level modeling, and the cross-layer feature fusion methods are simple, making it difficult to mine semantic complementarity between layers. To address these issues, this paper proposes an image caption generation model based on dynamic path and local visual fusion. The dynamic path encoder builds a multi-level routing space through spatial and channel modeling units and adaptively allocates path weights based on a spatial-channel joint routing mechanism, enabling the network structure to dynamically adjust according to the input image. The local visual fusion module enhances semantic complementarity through cross-layer feature aggregation. Experiments on the MS-COCO dataset demonstrate that the proposed method has competitive performance, verifying the effectiveness of the proposed strategy.