This PDF file contains the front matter associated with SPIE Proceedings Volume DSCC300, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

Semi-supervised-Learning(SSL) providing a solution to leverage vast amounts of unlabeled data. In cognitive psychology, the Primacy-effect refers to the phenomenon where the initial information encountered tends to leave a deeper impression in human cognitive processes, serving as the basis for subsequent judgments. Inspired by the Primacy-effect, this paper proposes a novel semi-supervised image classification algorithm. The core idea of this algorithm is to mimic the beneficial effects of the Primacy effect on human cognitive processes, simulate similar phenomenon on artificial neural networks. In this paper, the initially labeled data is referred to as exemplar. The algorithm includes an exemplar prediction module, whose main function is to accurately identify examples, ensuring that the model forms a “deep impression” of them. We found that due to the scarcity of examples, it is easy to cause model overfitting. Therefore, we proposes the Weighted- Gradient-Chain technique. Additionally, Pseudo-labeling technique was employed, but during model training, we found that generated erroneous Pseudo-labels could introduce errors. To enhance the quality of Pseudo-label generation, this paper proposes a Pre-Pseudo-labeling method. A series of experiments were conducted on multiple datasets. The results indicate that the proposed model performed well.

The task of fact verification is to evaluate the truth of a claim-focused on plausible evidence, verifying the truth of a claim in public health is challenging because the need for reasoning across various retrievable pieces of evidence. Most fact-checking studies typically focus only on political claims, while few studies address fact verification on other topics. Therefore, this paper chooses to study fact verification in public health. To support this case study, this paper employs PUBHEALTH, a dataset containing claims related to public health. In this study, this paper presents an approach based on a semantic hierarchy. Different from most methods, which represent evidence sentences by concatenating or merging the characteristics of isolated evidence sentences, our method obtains the complex semantic hierarchy of evidence sentences through semantic analysis. Specifically, based on the XLNet model, the graph structure is initially utilized to redefine the relative distance of words. Then, a graph attention network and multi-layer graph convolutional network are used to propagate and aggregate information from neighboring nodes on the graph. Experiments on the PUBHEALTH dataset show that the proposed method achieves 76.2% in F1 score, which is substantially higher than the current highest score, indicating the effectiveness of the fact verification task based on semantic hierarchy.

When using convolutional neural networks (CNNs) for facial emotion recognition, down-sampling in the convolutional layers often leads to the loss of many features. These lost features also play a role in emotion recognition and can affect the effectiveness of emotion recognition. Therefore, we propose a facial emotion recognition method based on auxiliary classifiers and SENet modules to effectively alleviate the problem of feature loss during the convolution process. By using SENet modules, the model enhances its focus on different channel features during the feature extraction process, enabling the model to pay better attention to important features. Subsequently, auxiliary classifiers are introduced into the model to effectively utilize features extracted from different layers for auxiliary classification. Finally, the auxiliary classification results are adaptively weighted and fused with the main classifier's classification results. Experimental results demonstrate that introducing auxiliary classifiers and SENet modules further improves the model's accuracy, reaching 68.87%.

This article proposes a method for obtaining farmland information in high-resolution satellite images. Firstly, divide the pixels of satellite images into a set of soft object regions and correspond them to a class. Secondly, the representation of each object region is estimated by aggregating the classes of pixels in the object region. Then, design the Hamacher operator decision method to make decisions on the neutral fuzzy part. Finally, calculate the relationship weights and weighted aggregation between the target regions, with the highest aggregation being the category of the pixel region. The experimental results show that the proposed method can effectively obtain cultivated land information from high-resolution satellite images. Meanwhile, compared with existing mature methods, the scoring values of our method have significantly improved.

Aspect-based Sentiment Analysis (ABSA) is a specialized form of sentiment analysis that zeros in on pinpointing and harvesting sentiment details pertinent to distinct facets within textual content. In the field of sentiment analysis, traditional neural network models often face several challenges, including insufficient accuracy, low efficiency and poor generalization ability. Therefore, we propose a model called TLN-NET, this approach utilizes an adapted Sentic Net graph convolutional network and is designed to carry out sentiment analysis at the aspect level. First, we design a triple GLN (Graph Layer Normalization) architecture, which adds layer normalization technique to each GCN layer to maintain the stability of the gradient, thus improving the training efficiency of the model. We use the residual structure to combine the encoding output with the pooling output to enhance the expression ability of the model. Finally, we used Multi-layer Perceptron (MLP) and Softmax function to classify the final output, which improved the generalization ability of the model. We use the improved TLN-NET model to test on multiple datasets. The experimental results show that our proposed model has higher accuracy in judging aspect words, which is better than some state-of-the-art methods.

In vibration measurement, increasingly greater accumulative integration errors will be caused by the interference of DC components and low-frequency trend terms when an acceleration signal is integrated into velocity and displacement signals. To repress such errors, the principles and characteristics of time-domain and frequency-domain integration were analyzed, and the acceleration signal was preprocessed before integration. Afterward, the signal was subjected to low-frequency rejection processing so as to improve the integration accuracy. Moreover, the scope of application of the time-domain and frequency-domain integration method was analyzed and discussed. This method is of reference significance to the integration method of different acceleration signals.

The numerical solution of large sparse linear systems is a fundamental aspect of many large-scale scientific and engineering computations. The multifrontal method, as a prevalent direct solving approach, currently lacks a corresponding implementation on domestic Digital Signal Processor platforms. With the development of the domestic FT series high-performance digital signal processors, there is an urgent need to meet the demand for efficient numerical solutions of large sparse linear systems in practical engineering applications. In response to the above-mentioned issue, We implement and optimize the multifrontal method for the FT-M6678 platform. By analyzing the hardware architecture of the FT-M6678 and the characteristics of the multifrontal method algorithm, we employ compilation optimization, loop unrolling, and single-instruction multiple data vectorization techniques. This fully leverages the independent functional units and register resources of the platform, achieving instruction-level and data-level parallelism. Considering the storage hierarchy of the FT-M6678, we configure the first and second-level caches, set attributes for the cacheability of external memory, and design memory layout based on memory bandwidth. We allocate different data and code segments to distinct storage areas, optimizing the storage and access of data. Experimental test data is sourced from the University of Florida Sparse Matrix Collection. The acceleration ratio of the algorithm after optimization on the FT-M6678 platform ranges from 16.0~38.9. Compared to the TMS320C6678 platform, the performance improvement can reach up to 2.3 times.

Modern wireless communication has increasingly high requirements for communication systems. In order to meet the complex and everchanging application scenarios and performance requirements of 6G, a new waveform technology orthogonal time frequency space (OTFS) modulation has been developed in recent years. Most of the current research mainly focuses on the basic principle of OTFS system and the key algorithms of channel estimation and signal detection, and there are fewer researches on the coded OTFS system. this paper focuses on the low density parity check (LDPC) coded OTFS system signal detection BER performance is evaluated. By adding LDPC encoding and decoding modules at the beginning and end of the OTFS system, correlating the OTFS detector with the LDPC decoder. The performance of signal detection of the LDPC-encoded OTFS system is derived, and compared to the uncoded OTFS system, the performance of the LDPC-decoded receivers are all improved to a certain extent.

With the aging population in our society, there is a growing demand for the detection of vital signs. Millimeter-wave radar has gained significant attention due to its high accuracy and non-contact nature. To enhance the accuracy of heart rate prediction, this paper proposes a method based on a hybrid CNN-Transformer model. In this method, the CNN layer is utilized to extract local features of the heartbeat phase, compensating for the limited information mining capability of the Transformer encoder when dealing with multivariate datasets. Concurrently, the Transformer module is employed to capture dependencies between local and global features in the sequence data, aiding the model in better understanding the overall characteristics of the heartbeat signal. The experiments were conducted using an FM continuous wave radar device with a center frequency of 77 GHz. The experimental results demonstrate that the proposed model achieves an average absolute error of only 0.94 bpm, outperforming existing models when compared to medical-grade ECG acquisition equipment. This validates its superiority in heart rate prediction.

Aiming at the problems of detail loss, edge blurring and noise amplification in traditional infrared image enhancement methods, an algorithm based on improved weighted guided image filtering (IWGIF) and wavelet thresholding denoising (WTD) is proposed. The algorithm first uses IWGIF to filter and layer the input image to obtain the base layer and detail layer. For the base layer, the high-frequency component obtained through wavelet decomposition (WD) is processed through WTD and subsequently reconstructed using wavelet reconstruction (WR) along with the low-frequency component. Contrast is then enhanced by contrast limited adaptive histogram equalization (CLAHE). For the detail layer, its edge details are enhanced by gain calculation. Finally, the processed two layers are weighted and fused to obtain the enhanced image. In this paper, the algorithm is applied to images captured by infrared thermal imaging devices, and the results of the combination of subjective and objective evaluation indexes show that the method in this paper shows good performance in infrared image enhancement.

In the actual transmission process, the signal will be disturbed by a lot of noise, which directly affects the quality of the signal. The purpose of speech enhancement technology is to denoise the signal containing noise, which is a method to improve the effectiveness of signal and remove noise. This paper is based on the wavelet transform threshold denoising method, in the wavelet transform wavelet denoising processing, research a new threshold selection algorithm, to achieve a better noise reduction effect, so as to achieve speech enhancement. Write a speech signal processing program in Matlab software, add other recorded speech signals to the program, process the signal with Gaussian noise, then replace the proposed algorithm into the speech signal processing program to denoise, and observe the signal waveform diagram reconstructed at last. It is compared with the time domain waveform of speech signal, soft threshold and hard threshold, and determines whether the new algorithm can achieve speech enhancement according to whether the signal denoising effect is improved.

This paper is devoted to studying a new simulation technology of UHF partial discharge signal, aiming at improving the performance verification accuracy of UHF sensor. In this paper, a multi-channel UHF signal simulation source is researched and established with the help of PD simulation reproduction technology and programmable PD pulse sequence device, which provides a reliable tool for the verification and evaluation of intelligent sensing devices, contributes to the technical promotion and innovation in the field of intelligent sensing, and lays the foundation for the further development of new power systems.

In the research of synchronous location and map construction of indoor mobile robots, lidar is often used to build maps. However, three-dimensional lidar is expensive, two-dimensional lidar can only detect information above the plane, and the map constructed by depth camera is poor, so this paper proposes a low-cost map construction method based on the fusion of depth camera and lidar. First, the joint calibration method is used to make the two sensors in the same coordinate system, then the depth image is converted into laser data, and then the data is fused by selecting the minimum value method with error processing under the same Angle. Finally, the fusion data is released to the Gmapping mapping algorithm for the construction of two-dimensional maps. The effectiveness of the proposed method is verified by setting up gazebo simulation environment, and the experiment shows that the fusion method can build a complete map.

This paper introduces a method based on synchronous clock, which realizes blind separation of two kinds of superposition signals by Fourier series analysis. The sampled signal is accelerated to FFT by FPU floating-point. Through Fourier series analysis, four different combinations of signals are determined according to the Fourier series expansion of sine wave signal and triangle wave signal. A 'and B' signals are separated by inverting the signal frequency according to the position of the fundamental frequency signal. The A 'signal obtained by separation is extracted through the software zero-crossing comparator to extract the square wave of the synchronization signal. According to the frequency and phase control information incoming from the serial port screen, the expression of the two signals is back-derived to change the output frequency of SI5351, and the two DAC outputs are synchronously triggered by DMA according to the synchronous square wave signal extracted by the software zero-crossing comparator output two separate signals to realize the stable same frequency display on the oscilloscope.

With the increasing number of antennas in massive MIMO systems, traditional signal detection algorithms are faced with the challenge of increasing computing load and decreasing detection performance. To solve this problem, a new largescale multi-input multi-output signal detection algorithm, IPICNet, is proposed. Based on the iterative process of IPIC detection algorithm, the algorithm is transformed into a deep neural network structure, and the network structure is improved by introducing trainable parameters. The final simulation results show that the new algorithm has faster convergence speed while maintaining low complexity, and significantly improves the performance of IPIC detection algorithm.

File similarity is widely used in information security, vulnerability detection, image retrieval, IP protection and other fields. Hash algorithms can be used to compute the similarity between two identical or basically identical files. However, if sizes of files are different extraordinary, hash cannot accurately reflect their similarity. Analysis shows that the main reason for this problem is that the traditional similarity has nothing to do with file sizes, since files of different sizes are mapped to hash values of fixed length. For this reason, we propose a file-size-dependent similarity definition to solve this problem in this paper, and propose a file similarity algorithm which combines the traditional CTPH algorithm and the edit distance algorithm to solve the similarity problem of size-different files. An experimental environment was built to verify the accuracy and efficiency of the algorithm.

Reinforcement learning for optimizing traffic signal control is becoming increasingly popular, but the actual deployment of reinforcement learning algorithms poses certain challenges. The difficulties primarily lie in the following aspects:(1) The detection equipment installed at most intersections are mainly geomagnetic coils and video detectors and the detection range is limited. (2) Common academic traffic control algorithms usually allow random phase switching of signal phases. However, in practice, considering the habits of drivers and the existence of left-turn waiting lane, the application of random switching signal control mode may introduce potential risks. Therefore, it is necessary to use traffic signal control methods with fixed-phase sequence. In this paper, we propose the DACLight to address the deployment problems. Under the assumption of fixed phase sequences, DACLight utilizes the attention mechanism to make decisions on traffic signal control strategies based on queue lengths of each lane obtained from intersection detectors. Experimental results conducted in both the real-world dataset scenario in Zhengzhou and the public dataset scenario in Hangzhou demonstrate that the algorithm outperforms other baseline experiments.

To identify and differentiate common partial discharge faults in 10 kV switchgear, a simulation and detection platform for switchgear partial discharge faults was established. A partial discharge identification method based on ultrasonic signal characteristics was proposed. The transient impulse quantities of ultrasonic signals during partial discharge were extracted, and their probability density distribution was used for type discrimination. A comparative study was conducted on the probability density distribution charts of transient impulse quantities of ultrasonic signals generated by four different partial discharge models. The results indicate that with the increase in voltage level, the curves shift towards higher transient impulse quantities and exhibit a 'multi-peak' pattern. Curves generated by the same type of fault overlap in height, while curves generated by different types of faults show significant differences. This approach can be used to discriminate the type of partial discharge in 10 kV switchgear.

In order to facilitate the wireless power and signals transmission in scenarios involving multiple receiving devices, this study introduces a method for such transmission. This method, which is based on the principle of inductive coupling, is designed for use with multiple transmitters and receivers. A signal transmission coupling mechanism is incorporated into the existing energy transmission structure. The system’s energy transmission topology is based on the LCC-S composite resonant network topology. Digital signals are modulated and demodulated using the binary amplitude modulation (2ASK) method, which allows the data stream being transmitted to be modulated to the appropriate carrier frequency band for data transmission. The feasibility and effectiveness of the proposed method are ultimately confirmed through experimentation.

Maintaining the timeliness of information from multiple monitoring nodes in event-reporting networks is a challenging task. Based on the research on age of information(AoI) and aggregate routing, the relationship model between AoI and aggregate routing in event-driven network. An aggregated routing algorithm (STAR) is proposed, which takes the node AoI and the number of nodes involved in the data routing path as the reward function, and the routing node of Steiner point is selected as the routing node of information reporting. The good performance of this algorithm under different network environment parameters was verified by using the event AoI as a performance index of the target node.

Based on the multimodal nature of data, a wavelet-based adaptive multimodal data compression algorithm is designed. The algorithm adaptively adjusts the classification of the data under a given correlation threshold, fits the correlated data using least squares estimation, abstracts the featured data into a matrix, and removes spatial and temporal correlations from the data using wavelet transforms. Theoretical analysis and simulation experiments show that the new algorithm can effectively remove the multimodal correlation between data and the spatial and temporal correlation of the same data, and the new algorithm effectively improves the compression ratio and reduces the energy consumption of the network.

In large-scale multiple-input multiple-output (MIMO) systems, the computational complexity of linear data detection poses a significant challenge. This paper addresses this issue by focusing on enhancing detection performance through the optimization of the circuit structure based on the Neumann series approximation method. While using a limited number of Neumann series terms for approximate matrix inversion proves to be an effective method, the associated low complexity requires improvement. This paper proposes a novel architecture that enhances the cell distribution structure of Gram matrix computing cells. This improvement results in an objectively low-complexity approximation inversion of the pulsating array, with simulation results on FPGA supporting the theoretical findings. The inversion in the approximation matrix is done on Xilinx Virtex-7 FPGA for a system with 128 BS antennas for receiving the data of 8 single-antenna users; Slice LUTS and Slice Registers account for only roughly 19.53% and 27.17%, respectively.

The detection of weak signals has a wide range of applications in engineering. In this paper, a combined algorithm of wavelet transform and correlation detection is proposed to address the problems in current detection methods, namely the wavelet-correlation detection algorithm, to detect weak signals. By building a Simulink model on Matlab and implementing simulation through programming, the reliability of the algorithm is verified. The detection performance of the algorithm for weak signals in fixed noise and color noise environments is analyzed, and compared with the cross-correlation detection method. The results show that the wavelet-correlation detection algorithm can distinguish weak signal features from strong noise backgrounds. Taking the 5.2943dB input signal as an example, the wavelet- correlation detection algorithm improves the relative error of amplitude from 1.89% to 0.35%. Signal-to-noise ratio increased by 8.7638dB.

We propose a ghost imaging (GI) denoising method based on connected-region labeling (CRL) technique, named CRLGI, to reduce the noise interference from the environment, and improve the quality of imaging. In the scheme, a maximum labeled connected component is searched by using line-by-line scan algorithm in each binary speckle pattern, and the area whose number of pixels is smaller than that of the maximum labeled connected component is deleted to eliminate the noise in every speckle pattern. Then these optimized speckle patterns are illuminated on an unknown object. A bucket detector without any spatial resolution is used to collect the light transmitted through or reflected from the object. Finally, the image of the object can be reconstructed by the correlation between the optimized speckle patterns and corresponding bucket values. The simulations show that the proposed scheme has a better denoising effect, especially in the noise varying complex environment. For Gaussian, salt & pepper, speckle, Poisson noise environment, the images reconstructed by CRLGI scheme are improved than those recovered by traditional GI scheme.

This paper addresses the digitization of power standards in the power industry through the development of knowledge graphs and the application of intelligent technologies. We explore the transformation of standard management, leveraging technologies like cloud computing, big data, and AI, and discuss the strategies of international standardization bodies like ISO and IEC towards machine-readable standards. The core of our research is the construction and application of a power standard knowledge graph, aimed at enhancing the management of standardized knowledge. This involves stages from data collection, incorporating OCR technology, to data management with an emphasis on data quality and security. A significant contribution of this work is the development of the PLM-based Knowledge Graph Completion Model, particularly focusing on the KG-SBERT model, which integrates semantic and structured features from pre-trained language models and knowledge graphs. Our experiments demonstrate the superiority of KG-SBERT over traditional KGE models in knowledge graph completion tasks, highlighting the effectiveness of prompt information in improving model performance. surpassing traditional knowledge graph embedding techniques in link prediction accuracy, attaining an MRR of 47.67, and Hits@5 and Hits@10 scores of 25.15, 81.54, and 93.91, respectively. This study underlines the critical role of digital transformation in the power industry, demonstrating how advanced technologies can significantly enhance the management, security, and accessibility of power standards, with broad implications for the industry's future development.

With the continual advancement in network attack technology, network security defense methods based on a single model are difficult to effectively detect complex and unknown types of attacks. Therefore, an Improved Grey Wolf Optimization based Transformer (IGWO-Transformer) is proposed in this paper for multi-model ensemble network intrusion detection. IGWO-Transformer utilizes multiple sub-models to extract features from network flow sequences from different perspectives, followed by an ensemble prediction with Transformer. To address the issue of slow convergence in the vanilla Grey Wolf Optimization (GWO), which leads to the ensemble model's tendency towards individual models, a Cosine Convergence Factor (CCF) and a Dynamic Weight (DW) strategy are introduced to enhance the algorithm's search capability and convergence speed. Experimental results on the UNSW-NB15 dataset indicate that compared to single-model prediction methods, the proposed IGWO-Transformer significantly improves prediction Accuracy by 7.62. Extensive comparative and ablation experiments validate the effectiveness of the proposed method.

This research paper proposes a design for a UAV platform that carries a signal beacon, frequency inverter, and specially designed target for the Ka-band measurement and control antenna. It presents methods for calibrating the Ka-band antenna, including phase calibration, zero range calibration and photoelectric axis mismatch by designing the UAV flight path at different heights. Results confirm that UAVs with a high loading capacity and stable hovering ability can achieve better elevation positioning and thus get a higher calibration accuracy compared to traditional calibration methods. This method reduces the safety hazards associated with climbing the calibrating tower and provides a better way for Ka-band measurement and control equipment under extreme geographical conditions like forests and mountains.

The laboratory, as a research base for university students, is becoming increasingly important to manage. With the development of technology, intelligent laboratory management systems have become a trend. Therefore, how to implement a fast and accurate laboratory management system has become a key issue. To solve these problems, this paper constructs a preliminary laboratory personnel counting system based on the improved YOLOv9 target detection algorithm. The following research work has been done in this paper: The important downsampling part ADown in YOLOv9 is improved, and the convolution module SPD-Conv for low-resolution image detection is combined, which improves the detection effect of YOLOv9 in low-resolution scenes. A laboratory personnel entry and exit counting system is designed. By dividing the entrance line, the statistics of the laboratory entry and exit personnel are achieved. When the number of laboratory personnel returns to 0, an alarm reminder can be given.

As the information age rapidly evolves, the encrypted traffic in the network exhibits an exponential growth trend. Encrypted traffic classification technology has become the key to ensure information security and optimize network operation. However, the traditional traffic classification methods based on port number and payload cannot meet the increasing user privacy protection requirements and the demand for fast real-time classification. Meanwhile, current deep learning-based traffic classification approaches suffer from issues like inadequate data processing, limited feature extraction, and imbalanced sample category distribution, leading to poor classification performance. To solve these problems, we propose a residual network traffic classification model based on Convolutional Block Attention Module (CBAM) and Haar Wavelet Downsampling (HWD), aiming to fully exploit the intrinsic features between traffic bytes. And we validate the accuracy and effectiveness of the proposed method using the publicly available ISCX VPN-nonVPN dataset. The experimental results demonstrate that our model performs better in terms of accuracy, precision, recall and F1-score compared with existing methods, which has important theoretical significance and potential for practical application.

Remote photoplethysmography (rPPG), which estimates heart rate from facial video signals captured by inexpensive RGB cameras, has been shown to be effective at estimating heart rate in stable scenarios. Region of Interest (ROI) selection is a crucial aspect of mainstream rPPG heart rate estimation methods, as improper settings can introduce significant noise. In this paper, we reconsider ROI selection strategies from the perspective of mask smoothness and propose an ROI selection strategy called Temporal-spatial smoothing mask. This strategy first selects masks using the proposed YCbCr-LUT for spatial smoothing, and then temporally smooths the masks using Gaussian filtering. Experiments on the publicly available datasets UBFC-rPPG show that the proposed method effectively improves heart rate estimation performance.

In this paper, a LoRa network protocol based on CSMA/CA access mechanism and asynchronous downlink mode is proposed for the communication problem in the LoRa network according to the star network model. The protocol aims to improve LoRa network performance, reduce the risk of node conflicts, and achieve stable and efficient data transmission. In this paper, the network parameters of the protocol are determined, and the protocol mechanism, backoff algorithm, and asynchronous downlink mode are designed. The performance improvement of the improved protocol in terms of normalized throughput, communication collision rate, average communication delay is verified through simulation analysis. In the case of 150 nodes competing for the channel at the same time, this protocol has a better performance compared to the original protocol, which can improve the throughput, reduce the communication collision by 11% and reduce the delay by 83ms.

A compact, low coupling multi-input multi-output antenna is presented in this paper. The overall dimensions of the antenna are 28mm × 22mm × 1.6mm, and the antenna is made up of two radiating elements. The front of the antenna is composed of a "knife and fork" patch and a rectangular microstrip feed line; The rear face of the antenna is formed from a defective ground structure, which is etched on the ground. The antenna, on the other hand, employs a defective ground structure with two symmetric rectangular slots etched into the ground and has a common T-shape ground structure to achieve high antenna isolation. In conclusion, the terminal antenna's impedance bandwidth measures 3.07 ~ 13.49GHz, its relative bandwidth is 125%, its in-band isolation surpasses 16dB, and its ECC is less than 0.004. Because of this, the UWB-MIMO antenna is appropriate for UWB-MIMO wireless communication and has good radiation efficiency and a small ECC.

This paper introduces HetGNN, a novel machine learning framework for wireless network resource allocation, blending graph neural networks with long short-term memory networks to adeptly manage graph-structured and temporal data. An adaptive resource allocation algorithm, rooted in Actor-Critic reinforcement learning, dynamically refines strategies, ensuring efficient interaction between agents and their environment. Comparative analyses with benchmark algorithms on a simulation platform highlight HetGNN's superiority in enhancing spectrum and energy efficiency, alongside user experience. The study paves the way for advanced applications in 5G/6G networks, emphasizing the integration of federated and transfer learning for future intelligent wireless network developments.

In order to suppress the interference of narrowband signals such as WiMAX band, international satellite band, WLAN band, X band and International Telecommunication Union (ITU) band to ultra-wideband communication, this paper proposes a miniaturized ultra-wideband (UWB) antenna with five-trapped wave characteristics, which consists of a wrench-type radiating patch and an improved open-rectangle-slotted grounding plate with a FR4 dielectric substrate. The overall size is 29mm×20mm×1mm, and the antenna is realized by opening C-slots, C-slots and inverted U-slots on the antenna radiating surface, etching U-slits on the feedline and adding symmetric C-branches next to the feedline, respectively, for the five frequency bands of 3.14-3.71GHz, 4.14-4.84GHz, 5.04-5.99GHz, 7.23-7.84GHz, 8-9.14GHz, and 4.14-4.84GHz, respectively. 9.14GHz for five frequency bands with trapped wave characteristics. The experimental results show that the designed antenna operates in the frequency band of 2.82~13.17GHz, and the return loss in the working frequency range is less than -10dB except the above five trap bands, and it has good performance and omnidirectional radiation characteristics, which can be applied to a variety of wireless communication system fields.

To address the inability of traditional home switches to identify fault locations and be controlled remotely, this paper introduces a multi-function intelligent switch monitoring system based on NB-IoT. The system automatically records household electricity metrics, such as maximum current, voltage, and power consumption, offering insights into consumption habits. It uploads these parameters to the OneNET platform via the M5311 communication module, facilitating data exchange between the main control module and the cloud platform. Consequently, users can monitor their electricity usage and detect faults via mobile phones, remotely manage switches within the household, promoting efficient fault resolution. The use of cost-effective STM32 for data collection, analysis, and fault detection significantly lowers the application costs, while intelligent control reduces labor expenses. Its suitability for widespread household use and potential for mass adoption indicate broad application prospects.

As one of the most important modules in the wireless communication system, the radio frequency power amplifier’s performance directly affects the performance of the entire wireless communication system. Research on power amplifiers has been developing in the direction of broadband, integration, and high performance. Based on the continuous theory and filter matching theory, this paper proposes a continuous class F power amplifier, which expands the impedance matching space by introducing a correction factor. Compared with the traditional class F power amplifier, it effectively increases the operating bandwidth. A 0.5-2.5GHz ultra-wideband power amplifier is designed and implemented. The filter is used as the output matching network, which reduces the loss of the 50Ω transmission line and improves the transmission efficiency of the system. The design results show that in the 0.5-2.5GHz frequency band, the saturated output power is as high as 41.5dBm, the gain is about 12dB, and the power added efficiency (PAE) is 58%- 64%. This structure reduces the design difficulty and circuit complexity, has the advantages of wide frequency band, high efficiency, etc., and can well meet the needs of current wireless communication systems.

In aeronautical telemetry systems, diversity is often used to overcome multipath interference in frequency-selective fading channels. Generalized Time-Reversed Space-Time Block Coding (GTR-STBC) is a promising diversity scheme, which differs from other diversity schemes in that it dynamically allocates power to multiple transmit antennas based on feedback parameters from the receiver. However, the original system did not offer a solution for scenarios involving multiple receiving antennas, and it required a significant number of iterations during the computation process, leading to high computational complexity. In this paper, we propose a dual-receiving-antenna algorithm based on Maximum Ratio Combining (MRC) for GTR-STBC systems. Compared to single-receiving-antenna systems, the proposed algorithm improves the bit error rate performance by 0.6-1.2 dB. In scenarios with single receiving antennas, the proposed algorithm eliminates the need for iterative computation, significantly reducing computational complexity while maintaining negligible performance loss.

In direction-of-arrival (DOA), most of the existing popular deep learning-based methods consider uniform linear array (ULA) systems. Although they show higher performance than parametric methods, the high accuracy of the estimation still needs to be improved. In recent years, coprime arrays have been widely studied due to their advantages over ULAs. In this paper, we propose a deep framework for dual-signal DOA with coprime array, considering it as a multi-classification problem. The entire framework consists of the feature network and the classifier network. First, we propose a preprocessing method suitable for coprime arrays, where the extended covariance matrix is used as the input processing object. Next, the feature network extracts feature from the data. We introduce a lightweight channel attention mechanism to help improve the performance of the deep model. Finally, the classifier network performs DOA estimation.

Atmospheric turbulence causes refractive index fluctuations, and atmospheric molecules absorb and scatter light wave energy, causing the power density of the laser to reach the target to decrease, restricting the application performance of advanced optoelectronic systems. Precision optoelectronic equipment has problems such as long development cycles and difficulty in quality control. Digital twins have low cost to establish digital models, can be quickly iterated, and have great potential in system analysis and diagnosis. According to the laser atmospheric transmission physical model and performance evaluation requirements, the laser transmission model is constructed and integrated in the digital space, and the C++ language, Qt (Q Toolkit) platform and multi-threaded parallel processing technology is used to develop and test the software system. Provides practical software tools for laser transmission performance evaluation and system parameter optimization design. With the help of the software system, the functional relationship of the refractive index structure constant, wind speed, absorption coefficient and extinction coefficient profile with height changes as well as the energy intensity distribution of the target surface spot can be calculated and visualized.

This study aims to evaluate the accuracy of the atmospheric precipitable water vapor (PWV) inversion using the Global Navigation Satellite System (GNSS) technology and its temporal-spatial distribution characteristics in the central and western regions of the Chinese mainland. By comparing PWV data from 15 GNSS stations with ERA5 reanalysis data and radiosonde data, this paper analyzes the precision and consistency of GNSS-based PWV inversion, discusses the influence of latitude, elevation, and climatic factors on the accuracy of PWV inversion, and finally reveals the temporal-spatial variation patterns of PWV. The results indicate good consistency between GNSS PWV and radiosonde PWV, demonstrating the high reliability of GNSS in PWV measurement. The PWV inversion accuracy is closely related to the geographical location of the sites, with the mean absolute error and root mean square error increasing with latitude and decreasing with elevation. The temporal-spatial analysis reveals a growth trend of PWV at 0.12mm/year, along with significant annual, semi-annual, and 1/3-annual cyclical signals, and the spatial distribution closely related to topography and climatic conditions. The study confirms the high reliability of GNSS-based PWV inversion, analyzes the significant influence of site geographical characteristics on accuracy, and lays a solid foundation for the extensive application of the GNSS technology in climate change monitoring and atmospheric water vapor research.

In multi-sensor tracking, tracking targets stably and accurately is one of the primary tasks. Aiming at the problems of outlier interference during measurement and unstable tracking filtering in complex scenarios, this paper proposes a robust maneuvering target tracking algorithm. The algorithm combines the probabilistic data association algorithm with the interactive multi-model algorithm, and utilizes the extended Kalman filter for the filtering operation. The accuracy of the algorithm's state estimation during target maneuvering is verified through simulation experiments, which proves that the algorithm can meet the needs of target tracking in complex scenes.

Electric power transmission lines are integral components of the power system, and their digitalization and intelligent construction are important directions for promoting the transformation and development of energy and power. This paper presents a three-dimensional digital management platform for transmission tower based on We3D software, which can comprehensively display the three-dimensional model of transmission lines and their surrounding geographical information. By integrating the model with data interfaces, it enables simple and rapid access to model information and sensor data. This facilitates comprehensive perception and intelligent analysis of the operation status of transmission lines, further enhancing the timeliness and intelligence level of transmission line inspection, and improving the efficiency of transmission line inspection.

As a crucial component for controlling automobiles, the instrument panel primarily achieves information transmission through data visualization, enabling drivers to intuitively understand the current driving status. With the further development of digitization and intelligence, the design of automotive instrument panels has received widespread attention. By utilizing digital twin technology for the design and simulation of automotive digital instruments, the display and operation of these instruments have become more efficient and precise. This paper employs Java for development and Micro EJ as the development tool. Using digital twin technology, the external design of the automotive instrument is gradually improved. Based on the functional requirements of the instrument, the operation interface is optimized to enhance operational efficiency. Simulations are conducted to verify the quality and performance of the instrument. A digital simulation system for automotive instruments is established to enable virtual operations. By leveraging digital twin technology, this paper explores and practices effective methods for the design and simulation of automotive digital instruments. It has developed an efficient and precise digital instrument panel for automobiles, demonstrating the importance and practicality of digital twin technology in the design and simulation of automotive digital instruments. This provides new ideas and methods for the development of smart watches or other products.

Deep groove ball bearings are crucial components in rotating machinery, and their failures can lead to equipment shutdown and significant economic losses. To address the issue of simulating bearing failures, this article proposes an innovative method to establish a mathematical twin dynamic model that successfully simulates the fault characteristics of various components of deep groove ball bearings. Vibration and acceleration data are simulated using Simulink, and experimental comparisons are conducted to demonstrate the accuracy and practicality of the simulations. Additionally, a digital twin platform is constructed using a web-based front-end to enable data interaction and twins, providing powerful support for the research and prevention of deep groove ball bearing failures. Identify relevant articles in literature searches, great care should be taken in constructing both.

Adaptability of sensor network systems is an important technology that enables them to automatically adjust their working parameters according to changes in the environment, in order to achieve more efficient and reliable performance. Sensor network systems typically require real-time collection and processing of large amounts of data, which takes up transmission time. To solve the problem that some sensitive events are not distinguished in the transmission process, this paper provides an adaptive method for data transmission in sensor networks. The priority of nodes is determined through lookup tables and statistical data, different priorities are set for different services, and different execution parameters are given according to the size of priorities, so that devices with higher priorities can achieve priority transmission. This paper provides stronger differentiated service mechanism support for IEEE 802.15.4 standard.

To address the problem of lacking a global perspective in traditional methods for network resource adjustment and optimization, a resource allocation method for a power communication network is studied based on the operation situation in this article. Firstly, a power communication network situational awareness model is constructed. Secondly, a network resource model, a network business model, and a situational evaluation model are built. Finally, a power communication network resource allocation method based on a genetic algorithm is proposed, which adjusts the weights of network links to change data forwarding strategies and thus adjusts the situational values of each node, achieving resource allocation and dynamic optimization of the power communication network. Finally, through simulation experiments, it is demonstrated that the proposed method improves network situational awareness by 17%.

The high proportion of new energy sources integration causes issues such as load imbalance and high complexity in resource management. Low resource utilization has emerged in the power communication network. It is urgent to provide unified scheduling and on-demand resource allocation. Hence, the dynamic resource orchestration mechanism for a power communication network is studied based on NFV-SFC. First, the resource orchestration architecture based on NFV-SFC is constructed to complete the scheduling of virtual network functions. An NFV-SFC-based dynamic resource orchestration model is established to evaluate the performance of dynamic resource configuration and a dynamic resource orchestration mechanism based on a simulated annealing algorithm is proposed. Numerical results show the algorithm’s effectiveness.

The ever-growing demand for wireless communication services, coupled with the proliferation of diverse wireless devices, poses significant challenges related to spectrum efficiency and interference management. The specific features and technologies associated with 6G are still evolving, the overarching goals involve pushing the boundaries of wireless communication to meet the demands of future applications and services. This research aims to explore and advance the capabilities of smart spectrum sensing and wireless coexistence technologies to address these challenges. The study will focus on developing intelligent sensing mechanisms, adaptive spectrum allocation strategies, and cooperative coexistence protocols, with the goal of enhancing the overall performance, reliability, and coexistence of next-generation communication systems. This research contributes to the advancement of communication systems by addressing the pressing challenges of spectrum scarcity, interference, and coexistence. The proposed smart spectrum sensing and coexistence technologies have the potential to enhance the efficiency and reliability of next-generation wireless networks, fostering the deployment of emerging technologies such as 5G, IoT, and beyond.

Gesture recognition systems are currently receiving increasing attention as a popular and practical means of human-computer interaction. Human-computer interaction gesture recognition systems provide new strategies for social interaction with virtual environments. The objective of this paper is to provide an effective gesture recognition system by improving the dynamic time adjustment algorithm. In this study, the proposed technique is applied to a video dataset and high accuracy is obtained. Also, the system is applied to an open database and the performance is compared with other commonly used algorithms. Experiments show that the proposed algorithm has good performance in both precision and accuracy, reaching 98.74% and 96.32% respectively.

This paper investigates the impact of phase deviation between the transmitter and receiver on the correlation value when capturing in the GMSK modulation mode, and proposes a multi pulse joint capture algorithm that adds different phases to the captured pulse. This algorithm adds different phase offsets to multiple captured pulses during transmission based on the original GMSK related capture. At the same time, quantize the sampling signal of the receiver, fit a set of captured correlation values to the data, and find the phase shift position with the highest correlation value. And it is believed that this is the best capture point, further reducing the capture error during capture and improving the accuracy of capture time.

Deconvolved beamforming is a robust high-resolution beamforming method which yields narrow beam and low sidelobe. This paper describes the basic principles of Conventional BeamForming (CBF) and compares twelve different beamforming algorithms based on deconvolution, thus highlighting their commonalities and differences. Simulations and experiments comparing the accuracy of each algorithm on two-point coherent sources and one-point source with 56- element Multi-Arm Spiral Array (MASA), the influence of different signal frequency and distance between sensor and source on the identification performance of acoustic source is analyzed, the results suggest that CLEARN-PSF generally provides an improved spatial resolution and is more robust to any sources.

This paper proposes an optimized algorithm combining multi-band downsampling with orthogonal mapping to reduce the complexity of the receiving implementation process. Firstly, the feasibility of combining multi-band downsampling with orthogonal demodulation is discussed based on the spectral changes of the signal after downsampling. The algorithm is optimized from two perspectives: sampling rate design and spectrum shape reconstruction based on IQ dualichannel signals. Simulation tests show that the improved algorithm can effectively demodulate and restore the original amplitude and phase baseband signals, and can effectively reduce the amount of data processing and algorithm complexity.

In real-world application scenarios, communication in multi-agent systems faces numerous challenges, especially issues related to noise interference. Taking autonomous driving systems as an example, when multiple autonomous vehicles interact and communicate with each other in the same environment, they inevitably encounter interference from external environments and noise from internal systems. These interferences can originate from complex road conditions, electromagnetic interference from surrounding electronic devices, or even signal attenuation caused by changes in weather. These factors can negatively affect the quality of communication between vehicles, thereby impacting the decision-making and coordination ability of the autonomous driving systems and increasing the risk of accidents. To address the issue of noise interference in real data, this paper proposes an adaptive feature fusion communication algorithm based on the VQVAE architecture. In this work, data is first passed through an encoder to obtain the original continuous vector. Then, this continuous vector is processed through a linear network layer to learn the parameter tensor hα. Subsequently, the obtained original continuous vector is discretized to get a discretized result. Next, the learned hα vector is multiplied by the discretized result. The purpose of this operation is to control the degree of discretization based on the learned hα parameter. At the same time, we calculate the value of (1-hα) and then multiply it with the original continuous vector to retain a part of its continuity. Finally, these two parts are added together and sent to the receiving agent (i.e., decoder). Compared with manually setting the degree of discretization and the original AE, VQVAE models, experimental results on three public datasets demonstrate that the algorithm proposed in this study can effectively improve the performance of multi-agent communication.

With the rapid growth of multimedia data, multimodal retrieval has become an important research field. Multimodal retrieval is a retrieval task involving multiple media types (such as text, image, audio, etc.). With the explosion of data in various fields into the Internet, data is presented in various forms such as video, pictures, etc., the single module of data retrieval can no longer meet the needs of information development, and the demand for multi-modal data retrieval is increasing. In order to resolve the problem, by searching and reading literature, this paper analyzes different research methods such as shared representation learning, deep learning multimodal fusion, and hash method, which are needed in the process of multi-modal retrieval, and sorts out and sums up the basic ideas of researchers to solve these problems. Finally, the future development direction and application prospect of multimodal retrieval are prospected. This paper hopes to provide reference and inspiration for the subsequent research and application of multimodal retrieval and promote the development of multimodal retrieval technology.

At present, the degree of correlation between various types of data in backbone communication networks is low, and there is no automatic calculation and correlation function between data and data. Faced with a large amount of feedback data on power grid operation, it is difficult to automatically associate defects, regulations, contingency plans, real-time data, and equipment attribute data, and there is a lack of effective application and sharing methods for troubleshooting knowledge. The perception of important business channel faults, analysis of business overload links, and troubleshooting and disposal of fault points rely heavily on manual judgment, requiring manual retrieval of relevant information from different systems. Due to the varying problem-solving abilities of employees and the lack of experience and knowledge accumulation, it is difficult to guarantee the processing time of on-site faults. To address the above issues, research on fault diagnosis and inference technology for backbone optical communication systems based on knowledge graphs will be carried out. A knowledge graph will be established for the correlation of operational faults between the large power grid and backbone optical communication systems, as well as a deduction and analysis model for the impact range of operational faults in backbone optical communication systems. A solution for early warning of operational risks and hidden dangers in backbone optical communication systems will be formed, providing effective support for improving the information and communication security capabilities of the power grid.

Power communication network is a complex network structure composed of multiple communication nodes connected by a specific topology through transmission links. Identifying and protecting important nodes in power communication network in advance is an effective way to reduce network vulnerability. This paper presents a way to analyze the importance of nodes comprehensively by combining the topology of power communication network with the power service. Topology structure is evaluated by degree centrality, compact centrality and Katz centrality. Business importance is evaluated according to service quality and security indicators. Entropy weight method is used to determine the contribution weight of each indicator according to the concept of entropy in information theory. Finally, through the relative change of network efficiency, the destruction-resistance of the network is analyzed to verify the effectiveness of the proposed method.

In this paper, we address the channel estimation in downlink multiple input multiple output with filter bank multicarrier (MIMO-FBMC) systems, which serve different users simultaneously. Since there exists intrinsic imaginary interference in FBMC, the pilot design and algorithm are the key for the low-overhead channel estimation in downlink MIMO-FBMC systems. To tackle this issue, we propose a low-overhead pilot design scheme by utilizing the characteristic of the imaginary interference factor, and subsequently, a channel estimation algorithm based on minimum mean square error (MMSE) criteria is presented. It is shown that, the proposed MMSE algorithm is robust to correlation matrix mismatch, and each user can perform the MMSE estimation independently based on the designed downlink pilots, only by partial priori knowledge of MIMO channels. In comparison with traditional methods, the devised channel estimation scheme can attain desirable results at a lower pilot overhead. The proposed channel estimation scheme has been validated through numerical simulations, with regard to minimum bit error rate and mean square error.

The development of cross-chain technology has opened possibilities for interoperability within the blockchain transactions. However, security concerns regarding cross-chain transactions have remained a key issue in this field. This paper introduces an innovative TrustRank algorithm aiming at enhancing cross-chain security by optimizing reputation-incentive mechanisms. Specifically, the study concentrated on the reputation-incentive mechanism of the TrustRank algorithm, evaluating the intrinsic value of nodes based on their transaction history within a single chain and their significance as notaries in cross-chain interactions. In the algorithm proposed in this paper, TrustRank propagation weights are introduced to establish relationships between nodes. Through experiment and performance evaluation, the effectiveness of the TrustRank improved algorithm in enhancing cross-chain security is demonstrated. The results illustrate preventing malicious activities and attacks, as well as in the security of cross-chain smart contracts.

Nowadays, the target recognition of harmonic radar mainly adopts specific frequency points as the target characteristics, and the difficulty lies in the weak adaptability to the external actual environment, and false alarms and false negatives often occur in high-intensity clutter environments, making it difficult to effectively detect the target. In this paper, the harmonic radar echo signal dataset is established by near-field radar equipment, the harmonic radar echo data is processed by CNN, and the neural network parameters are iteratively trained by the Adam optimization algorithm to achieve effective recognition of PN junction and metal junction targets. Experiments show that the accuracy of the harmonic radar target recognition method based on CNN is higher than that of the traditional Constant False Alarm Detection (CFAR) technology and the LSTM harmonic radar target recognition method, and it also reduces the occurrence of false alarms.

Alongside the swift progression of modern maritime technology, the stability and reliability of shipborne communication terminal receivers have become increasingly crucial for the normal operation of ships. This study revolves around shipborne communication terminal receivers, capitalizing on finite element tools to analyze the first six natural vibration frequencies of the receiver, namely 94.596Hz, 141.84Hz, 147.01Hz, 158.65Hz, 179.67Hz, and 222.17Hz. Additionally, the impact of external excitations during practical usage on the internal core components of the device is examined. The findings reveal that under the operational state of the internal chips within the receiver, the maximum velocity along the X-axis response spectrum is 4.424mm/s, representing a 26.27% reduction compared to the ship's inherent vibrations. Furthermore, the maximum velocity along the Z-axis response spectrum for the chips is 4.539mm/s, indicating a 24.35% decline compared to the ship's vibrations. This research aims to provide insights and establish a scientific theoretical foundation for the structural design and optimization of shipborne communication terminal receivers.

Testing machine learning programs is challenging and costly using existing methods. Among them, metamorphic testing reduces the cost by verifying that metamorphic relationships are satisfied between test cases and outputs. The core of metamorphic testing is the metamorphic relationship, and the currently used Failure Detecting (FDR) fails to fully reveal its key properties in terms of error detection. Therefore, this paper proposes the concept of metamorphic relationship responsiveness, which comprehensively responds to the characteristics of metamorphic relationships by expressing the error detection rate of each metamorphic relationship as a multi-dimensional vector, thus providing more possibilities for machine learning research. And the optimized K-means algorithm is used for clustering analysis of metamorphic relationships, which mines out the association between metamorphic relationship construction requirements and error discovery ability. Experiments show that the metamorphic relationship reactivity can well support the cluster analysis and mine effective knowledge.

To effectively automate the selection of communication links and reduce the overall communication risk, a method for the automatic selection of power 5G communication links based on risk balance has been proposed. Initially, with the objectives of minimizing global risk and maximizing service channel availability, a model for the automatic selection of power 5G communication links is established, considering channel bandwidth, delay, and packet loss rate as constraints. The model evaluates service risk and network service risk balance among other indicators. Communication services are ranked from lowest to highest risk, and the model prioritizes lower-risk services for link selection. The communication link automatic selection model is solved using a quantum genetic algorithm, obtaining the lowest global risk and highest service channel availability, thus completing the automatic link selection process. Experiments have proven that this method can effectively determine the risk levels of various communication services, automate link selection, reduce overall risk, enhance service channel availability and bandwidth, and reduce channel delay and packet loss rate, achieving superior risk balancing effects.

The study proposes a transmission rate optimization method in unmanned aerial vehicle (UAV) networks based on non-orthogonal multiple access (NOMA) technique. In order to maximize the system and the rate, the study proposes a joint optimization function of resources that optimizes the associativity of the users, the bandwidth allocation of the system, the downlink power control, and the location of the UAV. Since the function solution is a complex mixed integer nonconvex optimization problem, the study proposes an iterative algorithm based on alternating and combinatorial optimization. Under an optimized user-association scheme, the optimization problem is decomposed into three sub-problems, and the Lagrangian relaxation method, successive convex approximation and logistic function approximation are sequentially applied to solve the sub-problems, respectively. Simulation results show that the algorithm proposed in this paper outperforms the original NOMA communication scheme in terms of system bandwidth and communication rate.

Vehicles periodically send beacon to broadcast their driving status to other vehicles. Frequent broadcasting of beacon leads to intense channel contention and congestion. Loss or delay beacon have a negative effect on traffic safety. This paper first detects the channel busy ratio (CBR) to estimate the current channel load. Then, it uses the Autoregressive Integrated Moving Average Model (ARIMA) and Kalman filtering to predict the CBR for the next moment. Finally, it adjusts the transmission power of vehicles through proportional-integral control based on the ideal and predicted value of the CBR. Simulations demonstrate that the proposed algorithm can reduce beacon transmission delay and improve the packet delivery ratios.

The Doppler rate-of-change plays a crucial role in characterizing the velocity of satellites within satellite communication systems. Accurate measurements of this parameter are essential to ensure reliable data transmission and optimal system performance. This research paper presents a novel method for accurately measuring the Doppler rate-of-change for HDR, and the problem of accurately validating Doppler frequency for HDR simulation has been effectively resolved by converting dynamic quantities into two static measurements. The proposed approach leverages the principles of spectrum analyzer superheterodyne frequency sweeping, zero intermediate frequency digital filtering, and oscilloscope roll mode measurement. By separately measuring the maximum Doppler frequency shift and Doppler scanning period, the method enables the precise and efficient calculation of the Doppler rate-of-change parameter. The key contribution of this method lies in addressing the issue of traceability of extreme value and dynamic index parameter measurement in satellite data transmission baseband testing equipment. This method provides a simple means for calibrating Doppler frequency dynamic parameters, Verifying the precision of simulating Doppler frequencies in the baseband equipment, which has been experimentally verified. Furthermore, the method for measuring dynamic Doppler frequency can be extended to various fields, including radar and unmanned aerial vehicle speed measurement.

With the demand for satellite communication in Chinese and abroad markets, the satcom-on-the-move (SOTM) based on ships, vehicles, airplanes, and other applications has gradually moved from the military field to military-civilian integration. Satcom-on-the-move can realize the real-time alignment with the target satellite in the inertial space through the gyro-stabilized platform and servo control according to its own position and carrier move information, and ensure the smooth uplink and downlink of satellite communication. This paper mainly reviews the system composition of the satcom-on-the-move, the impact of gyro drift, and beacon maximum tracking. A gyro drift compensation method based on beacon feedback is designed, and the effectiveness of the method is verified through testing. This method effectively improves the ability of the satcom-on-the-move to recover instantly after losing the lock.

New feeder automation principles such as fast-acting Multi-Agent and longitudinal difference protection rely on terminal-to-terminal horizontal peer-to-peer communication, while the current encryption and authentication system for power distribution automation is only established in the vertical communication dimension between the master station and the terminal, which cannot adapt to the security business requirements of horizontal encryption and authentication. Therefore, this paper researches the Multi-Agent inter-terminal authentication technology based on quantum encryption service engine and quantum authentication key, and constructs the quantum key negotiation mechanism for the group horizontal conversation between Multi-Agent terminals and the key auto-trigger updating technology based on the lateral information interaction rules between "terminal-terminal" in the typical feeder automation mode.

With the rapid development of power big data and intelligence, the complexity of intelligent distribution network is getting higher and higher, and the accompanying data security problems are also increasing, which seriously affects the healthy development of intelligent distribution network. In order to solve the problem of data security in power distribution network, this paper improves the Paillier encryption algorithm based on Chinese residual theorem and CRT system enhanced decryption technology, and analyzes the effectiveness of this algorithm by comparing examples. The results show that the improved Paillier encryption algorithm proposed in this paper can better fuse the communication security data of the distribution network, and has high operational efficiency and low communication load, which can meet the data security protection requirements of the intelligent distribution network, and has good practicability and economy.

This paper investigates the autonomous follower robot in the context of human-machine integration. A robust and efficient robot decision module is proposed to recognize, locate, and re-identify the target while handling situations of lost targets. First, the DBSCAN clustering algorithm is employed to recognize and locate the target based on point cloud data acquired by the single-line lidar. To identify the target pedestrian, the location information from Ultra-Wide Band (UWB) is utilized. Sensor data fusion with nonlinear least squares optimization enables the extraction of coordinate information in the lidar coordinate system. Subsequently, coordinate transformation is applied to obtain the position information in the robot's coordinate system. To enhance the robustness of the mobile robot's following process, a multimodal behavior controller is proposed based on the data from two sensors. This controller allows the robot to switch between different modes, facilitating the retrieval of the target even under external disturbances. Experimental validation demonstrates the effectiveness of the proposed multi-sensor fusion method in reducing the tracking error of the mobile robot by 55%.

With the development of the Internet of Things and computers, the demand for location services continues to increase. Indoor accurate positioning technology has attracted widespread attention. Millimeter wave radar is very suitable for indoor systems because of its low cost, small size, and all-weather characteristics. This paper proposes a human body tracking and identification algorithm (MHA) based on millimeter wave radar. This algorithm has high tracking accuracy and no visual impact. On the basis of millimeter wave radar, it first makes up for the angle measurement accuracy of millimeter wave radar, add a signal processing chain position estimation module, obtain sparse point cloud data and filter it, form a time-related trajectory, and propose an algorithm for accurate target identification. Evaluation and demonstration with various algorithms in various actual environments showed that the median position error of 5 people was 10cm and the recognition accuracy was 95%.