Abstract：In vessel detection from low-altitude UAV perspectives in inland rivers, traditional algorithms struggle to accurately detect small vessels due to issues such as small target size, vessel occlusion, complex backgrounds, light reflection, and wave disturbances. To address these problems, this study proposes an improved algorithm based on YOLOv11n—YOLO11-FFW (YOLO11—FEM FFM_Concat WIoUv2). To enhance the feature extraction ability for small vessel targets, the Feature Enhancement Module (FEM) is introduced, which expands the receptive field through multi-branch atrous convolution and integrates multi-scale contextual information. To improve multi-scale feature expression in complex backgrounds, the Feature Fusion Module Concat (FFM_Concat) is introduced, incorporating a learnable weight recalibration mechanism on top of the BiFPN structure, achieving adaptive fusion of high- and low-level features. To increase the model's robustness in scenarios with water surface reflection, occlusion, and dense targets, the loss function is improved to WIoUv2, dynamically balancing localization and classification losses. Experimental results show that compared to YOLOv11, YOLO11-FFW achieves a 1.4% increase in mAP@0.5, a 0.8% increase in precision, and a 2.4% increase in recall, which is verified to be effective in detecting small vessels in inland river scenarios from complex UAV perspectives.

Abstract：To address the conflict between the limited edge computing capacity of Unmanned Aerial Vehicles （UAVs） and the demand for high-precision multi-modal 3D detection， this paper proposes a lightweight detection network tailored for airborne platforms， termed Lite-VAFNet， building upon VAF-Net. A grid dimensionality-reduction detection head is constructed to compress the parameter volume by approximately 63%， thereby alleviating device storage constraints. A Linear-Bottleneck fusion module is designed to execute feature interaction with linear complexity， effectively eliminating the peak memory bottleneck. Furthermore， a collaborative acceleration framework integrating spatial resampling and logit distillation is introduced to overcome memory access limitations while compensating for the accuracy degradation induced by quantization. Experiments on the KITTI benchmark demonstrate that Lite-VAFNet achieves a 3D AP （Mod.） of 85.24% with merely 14.60 M parameters， significantly outperforming state-of-the-art models such as BEVFusion. This research strikes an optimal balance between accuracy and efficiency while substantially reducing resource consumption， exhibiting exceptional potential for edge deployment.

Abstract：In response to the scheduling complexities introduced by the integration of distributed generation and the limitations of existing single-dimensional evaluation methods, this study constructs a comprehensive evaluation index system for Artificial Intelligence (AI)-enabled distributed power scheduling. The system encompasses three dimensions: economy, environmental friendliness, and stability. The research employs a combination of the Analytic Hierarchy Process (AHP) and the entropy method, based on cluster analysis, for comprehensive weighting. Additionally, a multi-task deep learning algorithm incorporating an attention mechanism is designed to synchronously predict multi-dimensional indicators. Simulation results demonstrate that, compared to traditional scheduling schemes, the AI-enabled scheduling approach based on this evaluation system achieves a better overall balance across key indicators such as total cost, clean energy penetration rate, and power supply reliability. This provides effective theoretical and methodological support for the coordinated optimization of economy, environmental protection, and stability in power systems.

Abstract：Cellular-connected unmanned aerial vehicles (UAVs) show great potential in 5G networks. To address the challenge of maintaining stable connections with ground base stations during communication tasks, this paper investigates an UAV trajectory optimization problem. The objective is to jointly minimize the task completion time and communication outage duration while maximizing communication throughput as the UAV travels from the starting point to the destination within a given area. Considering the non-convex nature of the problem, a multi-step learning-based Dueling Double Deep Q Network (D3QN) algorithm is adopted to achieve adaptive trajectory optimization through interactive learning between the UAV and ground base stations. Simulation results show that, compared with the direct flight strategy, this method reduces the task completion time by 28%, cuts down the communication outage duration by 42%, and increases the average system throughput by 35%, achieving significant improvements in task efficiency, communication stability and system throughput.

Abstract：To address the problems of low target recognition accuracy and poor real-time performance of unmanned aerial vehicles (UAVs) in complex environments, this paper proposes a UAV target recognition algorithm based on adaptive neural networks and multi-scale feature fusion. The algorithm employs an improved convolutional neural network to extract multi-level features, combines attention mechanisms to adaptively adjust feature weights, and enhances target representation capability through multi-scale feature fusion. Experiments on the DroneCrowd dataset demonstrate that compared to algorithms such as ResNet-50, YOLOv11 and EfficientNet, the proposed method achieves an average recognition accuracy of 94.3%, representing an improvement of 8.7 percentage points over ResNet-50 and 2.0 percentage points over YOLOv11; an F1 score of 93.9%; and a processing speed of 61.0 FPS, representing a 35% improvement over ResNet-50. The method exhibits excellent robustness, providing an effective solution for UAV target recognition.

Abstract：In recent years, airborne infrared small target detection technology has become a research hotspot in the military and civilian fields. However, in practical applications, the influence of factors such as complex background and low signal-to-noise ratio still make infrared small target detection a challenge. Therefore, this paper proposes an improved airborne infrared small target detection algorithm AIR-YOLOv7, and uses the example transfer learning method to analyze the characteristics of small infrared targets, expand the data set, and further improve the performance of the algorithm. The experimental results show that the AIR-YOLOv7 algorithm has a better performance in infrared small target detection in airborne complex scenes, with a mAP value of 97.09% and an FPS of 102.09. With only a small amount of expansion of the data set in this paper, the instance migration method increases the mAP value of the algorithm by 0.96 percentage points, which provides a theoretical basis for the subsequent hardware platform edge computing transplantation.

Abstract：In the mass production testing of FPGA devices, the internal interconnection resources consume a significant amount of testing time. How to reduce testing time and save testing costs has been a major challenge for mass production. To address this issue, taking the XCKU5P FPGA as a verification example, we used Vivado to perform timing simulation on the tested interconnection paths, and based on the actual measurements from the ATE, we obtained the transmission delay of the interconnection lines, which was consistent with the simulation results. Furthermore, we studied the interconnection line delay under different test temperatures and identified the reasonable and stable delay wait time for ATE testing of FPGA interconnection functions. Through validation, adding 100 ns to the minimum delay wait time required for low-temperature functional testing can meet the stability requirements of three-temperature testing, effectively reducing the unnecessary redundant wait time set based on experience, thus improving testing efficiency and reducing testing costs.

Abstract：In recent years, Field-Programmable Gate Arrays (FPGAs) have been increasingly adopted in critical domains such as high-performance computing, aerospace, and communication systems, with their functionalities becoming ever more vital. As a result, the importance of FPGA software testing has grown significantly. However, insufficient testing of Intellectual Property (IP) cores remains a persistent challenge. This paper proposes a code review-based black-box testing approach that integrates static code analysis with black-box testing to enhance test adequacy. We begin by reviewing the current landscape of FPGA software testing, analyzing the limitations of both dynamic and static analysis techniques, and underscoring the essential role of black-box testing in IP core validation. A black-box testing framework based on code review is then constructed, featuring two core components: rule checking and configuration checking, which aim to identify and resolve potential design issues. Case studies on the missing reset signal in a PLL IP core and impedance mismatch in RS422 interface LVDS primitives demonstrate the practical effectiveness of this method in FPGA testing projects.

Abstract： A micro-current detection system is designed in this paper, it can be used to detect the weak current signal of electron beam equipment from picoampere level to nanoampere level. The design uses a transimpedance amplifier I/V conversion circuit structure, considering device selection, circuit cascade amplification, pure power supply, filter noise reduction, embedded computing, shielding and grounding, etc., the picoampere-level micro-current is amplified. At the same time, a dual-range selection circuit is designed to expand the measurement range to the hundred-nana level, and the amplified electrical signal is collected and optimized by the MCU microprocessor. Experiments show that the system can realize the measurement of weak current signal, and can be adjusted to a large range, and the measurement error is less than 0.5%, and the linearity is excellent and the response is fast.

Abstract：This paper addresses the issues of poor anti-interference capability and unstable output in magnetic-electric wheel speed sensor circuits. Based on the principle of Hall-effect wheel speed sensors, a Hall-type wheel speed sensor circuit was designed using a 0.18 μm BCD process. The circuit includes functional modules such as an amplifier, a bandpass filter, and a hysteresis voltage comparator. To mitigate the offset voltage inherent in Hall devices, a voltage bias orthogonalization method employing four Hall elements was adopted. Simulations conducted with Cadence Virtuoso demonstrate that the proposed Hall sensor circuit exhibits strong anti-interference performance. It consistently outputs square wave signals with stable amplitude across an input voltage range of 4.5 V to 30 V and an operating temperature range of -40 ℃ to 125 ℃, enabling accurate measurement of vehicle speed.

Abstract：The image quality of the Scanning Electron Microscope (SEM) is highly dependent on the accuracy and stability of the output signals from its scanning circuitry. However, the impact of scanning signals on image quality remains insufficiently explored in existing research. To address this, a transmission mechanism from scanning signals to image quality was established, and the influence of scanning signal performance metrics on the image was analyzed. Using high-precision digital multimeters and oscilloscopes, actual scanning signals were accurately measured. Data processed via the endpoint fitting method enabled quantification of error levels in both field and line scanning signals under a 512×512 pixel mode. Furthermore, MATLAB was employed to simulate the impact of various errors on a standard grid image, and measured signal data were imported to evaluate their effect in practical imaging. The results indicate that the scanning signals exhibit good overall linear performance, with an average error of 1.884 pixels in the central region, superior to the 2.749 pixels observed in the peripheral region.

Abstract：The switching of strong light or the interweaving of shadows within the power plant area can affect the stability of feature points, causing the temporal information to break during the processing of high frame rate video streams. It is difficult to obtain the change direction of human feature points in continuous video frames, resulting in a relatively high average EER for behavior recognition. For this purpose, research on the recognition of abnormal human behaviors in intelligent security monitoring of power plants based on temporal attention enhancement has been carried out. The temporal attention enhancement module is introduced to enhance the short-distance and long-distance temporal features of surveillance videos. After fusion, a joint feature spanning multiple video segments is output to correlate the information of the segmented video frames. The distance-rotation angle representation method is used to calculate the change direction of human feature points in consecutive video frames, and abnormal behaviors are identified based on the direction relationship. On the test dataset, the design method outputs human feature information spanning multiple video segments. The AUC for its abnormal behavior recognition reached 0.92, and the mean EER was stable within 0.15, which was at a relatively low level.

Abstract：Utilizing unmanned aerial vehicles (UAVs) to charge energy-constrained sensor nodes can effectively mitigate the problem of service interruption caused by node energy depletion. However, due to the limited energy capacity of UAVs, minimizing the mission time while ensuring complete charging of sensor nodes is crucial for reducing UAV energy consumption. To address this challenge, this paper formulates an optimization problem aimed at minimizing UAV mission time, based on the UAV energy consumption model, the sensor node energy harvesting model, and the remaining energy levels of the nodes. Furthermore, a UAV scheduling algorithm is proposed, which transforms the optimization problem into a classic Traveling Salesman Problem (TSP) through node clustering and anchor point selection methods, and employs a genetic algorithm for path planning. Simulation experiments using synthetic data are conducted to evaluate the proposed algorithm. The results demonstrate that the proposed UAV flight strategy effectively reduces mission time under the same energy consumption conditions.

Abstract：3D reconstruction plays a critical role in various fields, including computer vision and artificial intelligence, medical imaging, architecture, and urban planning. To address the inefficiency of traditional manual modeling methods, this paper proposes a method based on Neural Radiance Fields(NeRF) that incorporates multi-scale fusion and attention mechanisms. The approach introduces a multi-scale feature module combined with graph convolutional networks to enhance the network's understanding of spatial structures, allowing for more accurate capture of both local and global geometric relationships. The multi-scale feature module extracts information at different levels, improving the accuracy and comprehensiveness of detail reconstruction, which in turn enhances overall reconstruction quality.Additionally, to further improve the model's robustness and precision, a feature pyramid network is introduced to ensure the network can effectively capture important information across different scales, particularly in complex scenes where details might otherwise be lost. The integration of the Squeeze-and-Excitation attention mechanism allows the model to adaptively focus on key regions in the image, enhancing the representation of important features and improving reconstruction performance in challenging environments.Experimental results demonstrate that the proposed method outperforms the NeRF model on a self-built building dataset, achieving SSIM, PSNR and LPIPS of 0.784, 25.42 and 0.183, respectively. These metrics show improvements of 4.39%, 3.29% and 15.84% over the NeRF model, indicating better handling of complex reconstruction tasks. This method provides a new approach for 3D reconstruction in various application domains.

Abstract：In order to meet the needs of diverse user roles, heterogeneous data, and complex and changing page display requirements in the production, operation and management system platforms of different industries, a Web page generation method based on nested combinatorial model is proposed. By adopting a unified modeling language and object-oriented programming technology, combined with Vue.js and Nuxt.js frameworks, flexible configuration and efficient rendering of page elements are achieved. The proposed method has demonstrated satisfactory application results in different engineering projects, significantly reducing the configuration complexity and maintenance costs while enhancing the user experience, and it has high value for practical promotion.

Abstract：Under the background of the artificial intelligence era, regular expression matching technology plays a crucial role, particularly in data cleansing and data extraction domains, where it provides technical support for high-quality data processing required by large language model training. However, traditional regular expression matching algorithms suffer from performance bottlenecks that limit their application scope. To address this challenge, this paper proposes a data-driven high-performance regular expression matching algorithm, denoted as YJJFA algorithm. The algorithm reduces the number of input characters that need to be processed in the state transition table by partitioning it into optimal trusted regions and untrusted regions, while leveraging non-memory-access vector comparisons of the untrusted character set (UBCS) to achieve low-time-cost processing of trusted characters. Experimental results show that the YJJFA algorithm achieves a throughput rate of 17.88~53.81 Gb/s on L7filter rules, representing an order-of-magnitude performance improvement over conventional DFA implementations.

Abstract：Tunable Diode Laser Absorption Spectroscopy (TDLAS) is widely applied in gas concentration measurements due to its high sensitivity. However, its detection performance is often limited by laser stability and the signal processing capability of the lock-in amplifier. In order to address this issue, this paper proposes a methane detection system based on FPGA. The system integrates the laser driver, which realizes accurate adjustment of injection current and stable maintenance of temperature, and the lock-in amplifier module for signal demodulation. The experimental results indicate that when the laser current is 62.5 mA, current fluctuation is ±1.5 μA, and the stability of temperature regulation module is ±0.0063 ℃ at 30 ℃. Methane absorption tests are carried out at a concentration of 2 to 20 μmol/mol over an optical path length of 15 m. At 20 μmol/mol, signal-to-noise ratio of the second harmonic signal is 42.66 dB, calculating detection limit is 146.98 nmol/mol. Continuous measurements below20 μmol/mol show an optimal averaging time of about 249 s and the Allan deviation is 17.25 nmol/mol. The comprehensive results indicate that the circuit system can meet the high sensitivity requirements in gas detection.

Abstract：Lossless compression algorithms work as the core algorithms for information processing and storage in the process-level networks of smart substations. The Huffman algorithm, a universal lossless compression method, is applied in communication networks of power system. Based on the general Huffman coding compression method and the scenario characteristics of smart substation process-level networks, this paper proposes an improved Huffman coding compression method incorporating dictionary computation which can be bypassed. This method demonstrates advantages in computational latency. Furthermore, this paper presents the implementation based on Field-Programmable Gate Array (FPGA) devices, providing a detailed analysis of the design methodology for key modules, along with simulation, verification, and result analysis.

Abstract：A high-power interleaved parallel buck converter suitable for a low-temperature refrigeration unit in a certain space was designed to meet the requirements of high power density, high efficiency, and high reliability for DC-DC converters. This article elaborates on the design process of the interleaved parallel buck converter and successfully develops a product with a rated power of 800 W. It achieves a conversion efficiency of 95% at peak power, achieving the design goal of high efficiency and verifying the correctness of theoretical analysis and methods. The circuit has been tested and verified to meet the performance requirements, and has been successfully applied to model flight missions, providing reference for the design of subsequent high-power converters.

Abstract：With the increasing demand for information security and data privacy, the domestic SM4 block cipher algorithm has demonstrated significant application value in information transmission fields such as government affairs and commerce. To address the performance bottlenecks of the SM4 algorithm in FPGA implementations, this paper proposes an efficient fully pipelined hardware architecture. By optimizing the hardware implementation of the S-box, the number of S-boxes per round iteration is reduced from four to one, and fast substitution is achieved using combinational logic, significantly lowering resource consumption. Additionally, a 32-stage fully pipelined encryption or decryption module is designed to enable parallel processing of multiple data blocks, compressing the encryption or decryption throughput to a single clock cycle. Experimental results on the Xilinx Zynq7045 platform show that the proposed design achieves a working frequency of 412 MHz and a throughput of 52.7 Gb/s without relying on additional memory resources, with a 20% improvement in throughput per unit area compared to existing solutions.

Abstract：In order to solve the problem of digital RF signal duplication interference and broadband noise interference threatening radar target detection, this paper proposes an interference suppression algorithm that combines sparse reconstruction and deep learning. The algorithm first uses deep learning to extract interference features and detect the interference type and location, and then adopts a space-time adaptive processing strategy based on sparse representation to suppress interference in the space and time domains. Simulation experiments show that compared with traditional methods, this algorithm significantly improves the target detection probability and output signal-to-interference-noise ratio under strong interference background, and can effectively suppress main lobe simulation interference and broadband noise interference, providing a new idea for radar interference suppression and having engineering application value.

Abstract：In this paper, an RFID tag array layout in 3D space is adopted to perceive the quantity, location and type of foreign objects in cotton. Taking the received signal strength (RSS) and phase of the tag as characteristic quantities, the Euclidean distance ratio algorithm and the least square method are proposed to improve the positioning success rate. Based on the differentiated influence of different foreign objects on signal characteristics, a decision tree model is constructed to achieve the perception of foreign object types. Experimental verification shows that the positioning error is controlled within 5% after optimization, and it can accurately identify four types of foreign substances in cotton, demonstrating excellent perception and recognition performance.