This PDF file contains the front matter associated with SPIE Proceedings Volume 12626, including the Title Page, Copyright information, Table of Contents, and Conference Committee listings.

At present, in the first-generation and second-generation relay satellite systems applied in orbit in China, the tracking signal received by relay satellite antenna is the modulation signal with the code rate of 1Mbps~300Mbps sent by the user satellite. With the development of satellite communication system, part of high dynamic and low bit rate spread spectrum users put forward the demand for relay transmission. Based on the SNR analysis of 1 kbit/s ~ 300kbit/s low speed spread spectrum signal received by relay satellite, a synchronous demodulation scheme is proposed to capture and track the low speed spread spectrum KSA user signal and demodulate the error signal. In order to capture low speed spread spectrum signal in diagonal tracking system, a carrier frequency offset search algorithm based on FFT is proposed, which reduces the two-dimensional search of code phase-carrier frequency offset to one-dimensional search. The algorithm of signal tracking and Angle error signal extraction is introduced in detail. Finally, the proposed algorithm of Angle error signal demodulation is simulated and tested. The test results show that the proposed method is feasible and correct for relay satellite to capture and track Angle error signal of low speed spread spectrum user. In this paper, theoretical analysis and engineering practice are closely combined, and the research conclusions can be used as the basis for the signal acquisition and tracking and error signal demodulation method and parameter design of the third generation relay satellite system to be developed in China.

Secure access of massive terminals in PIoT is the key to realize the secure transmission of power information. However, the power system mostly adopts the certificate-based terminal access solutions, which need to query and verify the legitimacy of the certificate during authentication. The heavy burden of certificate management limits the application of low-cost and resource constrained terminals in PIoT. The SM9 algorithm takes user’s ID as the public key, omits the certificate management. The key escrow also meets the supervision needs of the superior and subordinate in power system. Based on the analysis of the security threats faced by huge amounts of terminal secure access, a three-layer framework is constructed for PIoT. The access layer sets up authentication nodes to manage the secure access processes of terminals. Aiming at the low efficiency problem of signature verification when a large number of terminals access, a SM9-based aggregation signature scheme is designed. The authentication node is responsible for aggregating multiple terminals’ signatures into one signature. Compared with other identity-based aggregated signature schemes, the proposed scheme is efficient in terms of computational performance.

Automatic classification and recognition of underwater acoustic communication signal modulation plays a key role in the field of underwater communication and confrontation, and it is a necessary combat capability in modern naval warfare. However, the recognition method based on Gaussian white noise environment established on the basis of traditional theory is still difficult to recognize in the background of underwater impulse noise. Modulation identification faces enormous challenges. In response to this problem, this study proposes an ensemble learning classification algorithm based on the fusion of convolutional neural network feature extraction and artificial feature extraction. Under small sample conditions, the simulation results show that the recognition rate of the ensemble learning method after feature fusion in the mixed Signal samples with multiple signal-to-noise ratios is improved to 99.7%, and it is robust to underwater impulse noise interference.

When a recommendation system is preferred to use for predicting the ratings, it is common that many users-to-items ratings cannot be found. And what makes the prediction even trickier is that these missing ratings are missing not at random (MNAR), meaning that there are other factors contributing to why some ratings cannot be found. In order to make better predictions, common approaches to reducing the prediction errors for those missing ratings are using imputed errors and weighting the ratings that are observed with the observed propensities, but each of them is not perfect enough (since they still show bad performance with bias or heavily influenced by the variance of propensities in estimation). To generate a better recommendation system algorithm that could produce a satisfactory outcome, a new estimator combining the imputed errors and propensities in a doubly-robust way is designed to perform an unbiased estimation and meanwhile reduce the influence caused by the propensity variance. In addition, joint learning with error imputation and rating prediction based on the estimator is adopted to acquire a more reliable performance, which shows a better result than the dataset without doubly-robust joint learning.

To determine the order of fractional regularization and achieve a good denoising effect, we described the image complexity from three aspects: the appearance of the image gray level, the spatial distribution of the gray level, and the appearance of the target object. It is composed of five factors, respectively the entropy of information, energy, contrast, pertinence and edge ratio. Then the automatic selection of order is realized, and the alternating direction multiplier method is used to solve the model. The experiment result shows that the improved model not only achieves the self-adaptability of the order but also has a good effect in removing noise and preserving texture.

Nowadays, with the low-altitude airspace being open, the environment faced by radar is becoming more complex. Gliders, paramotors, and unmanned aerial vehicles (UAVs) are used more extensively, and the aerial targets may run into hundreds or even thousands. The ubiquitous noise, clutter, and interference also cause infinite uncertainties. In view of the increasingly complicated on-spot situation, this paper proposes a tracking algorithm of LSS-Target based on the adaptive Kalman filter, which utilizes MATLAB software to conduct the simulation analysis of the system algorithm and adopts the system building module of FPGA for simulation verification. Based on that, the research and development of the general radar system for tracking small targets and trajectory detection is carried out. This system can adjust the calculation of noise covariance according to the background environment, optimize the tracking algorithm adaptatively, and realize the processing of target tracking and the display of target information in real-time. Through the simulation experiment, the radar system based on the adaptive Kalman filter algorithm has high accuracy in target tracking and fine effectiveness in the target recognition, which can be widely applied to the tracking and detection of small targets.

Atherosclerotic plaque can cause severe stenosis of the arterial lumen. Blood flowing through the narrowed artery may have different flow characteristics, and produce different forces on the plaque surface and arterial wall. The turbulence and force field in the lumen may have serious effects on vascular endothelial cells, smooth muscle cells and circulating blood cells. At the same time, different plaque characteristics will affect the flow field in the blood vessel, thus affecting the stress distribution on the vessel wall. This paper mainly focuses on the influencing factors of different plaque characteristics. The finite element models with different patch characteristics are established for numerical simulation. The effects of plaque size (degree of stenosis), plaque spacing and plaque relative rotation angle on the stress distribution of vascular wall were calculated.

Tactile rendering is cutting-edge technology in the field of human-computer interaction. Tactile rendering technology applied to multimedia terminals has broad application prospects in medicine, education, e-commerce, entertainment, military, and other areas. Multimedia terminal tactile rendering provides tactile interaction between human fingers and multimedia terminals, opens up a new sensory channel for communication between human beings and the virtual world, enhances the realism and immersion of human-computer interaction, and has important analysis and research value. This article provides a comprehensive review of the rationales and research status of three main multimedia terminal tactile rendering methods, introduces the driving methods and common databases, and analyzes the application prospects and future development trends of multimedia terminal tactile rendering.

In order to efficiently solve the vehicle routing problem with time window (VRPTW), a hyper-heuristic algorithm based on reinforcement learning was proposed. Firstly, the performance of the underlying heuristic algorithm was evaluated, and then a multi-armed bandit (MAB) algorithm was used to select the low-level heuristic algorithm. At the same time, a simulated annealing-based acceptance rule was used to determine whether to accept the solution obtained by each lowlevel heuristic to ensure the diversity of solutions. Experiments was carried out on some VRPTW benchmark instances, and the results show that the proposed algorithm is effective and stable.

With the rapid development of Internet information technology, more and more information pours into the Internet. Information extraction technology came into being, in which event extraction technology is the focus and difficulty. Event trigger word extraction is the first important stage of event extraction. Aiming at the problem of multiple triggers in an event mention, an event trigger extraction model based on semantic enhancement is proposed. In the pre-processing stage, by optimizing the MLM mask mechanism and integrating external knowledge such as element embedding information and Chinese military event terminology, the model can obtain more reliable Chinese military language representation. Then BiLSTM is used to mine the context features of Chinese military news long texts. Finally, CRF is used to assist label correction to achieve event trigger word extraction. Based on the data set constructed in this paper, the effectiveness of the event trigger word extraction model is verified. This method will provide some reference for domain event trigger word extraction.

The increasing use of fine spatial and temporal data in hydrological modeling has resulted in a prohibitively computational demand and run time. The serial computing method adopted by most modeling routines has largely hindered the application of high-fidelity distributed hydrological models. This study proposes a parallel simulation method for hydrological models based on container and automated deployment tools, i.e., Docker and Kubernetes, to achieve parallel simulation of hydrological models on a container cluster. Specifically, the Soil and Water Assessment Tool (SWAT) is used as an example to verify the efficiency of the proposed method by comparing it with the traditional serial simulation method. The experimental results show that the combined application of Kubernetes and Docker can significantly improve the simulation efficiency of the SWAT model. A maximum speedup of 3.6 times was achieved on a Docker cluster consisting of 10 virtual worker nodes. Due to the remarkable encapsulation ability of the container technology, the proposed parallelization method can also be applied to other environment models in the future, and the simulation efficiency can be further improved by extending the Docker cluster.

Remote sensing data processing involves many steps. This paper uses process flow to represent these processing steps. This paper combines remote sensing image metadata (RSIM) with case based reasoning (CBR) technology, proposes a parameter value adaptive method based on RSIM, and constructs a remote RSIM case model. It makes the representation of RSIM more intuitive and convenient. This method consists of two parts: case building and case similarity measurement. Case building refers to building RSIM into RSIM cases according to case models, and then classifying and saving each case. Case similarity measurement refers to the similarity measurement of metadata in the production process and each case under the flow. Finally, the parameter values will be configured according to the highest similarity. The method is tested in an actual remote sensing production system, and the experimental results show that the method has good applicability in the parameter value adaptation of remote sensing process production.

Sonar is a general term for various underwater acoustic devices, which are used to perform underwater target detection, positioning, identification, tracking and underwater communication, navigation, measurement and other tasks. All sonar devices including active sonar and passive sonar both require a signal processor to execute underwater acoustic signal processing. Beamforming is the main and core component of the sonar signal processor, and its processing speed affects the real-time performance of sonar. Thus, this paper first implements a parallel sonar beamforming algorithm to improve the processing speed of the beamforming algorithm. The proposed algorithm was carried out several experiments, and the experimental results demonstrate that the parallelism of multi-core CPU can bring better speedup ratio and efficiency than the traditional serial CPU algorithm.

Current trajectory data grows rapidly in a dynamic and streaming form, and unreasonable data organization causes the problems of skewed data storage and high overhead, as well as slow retrieval speed and page lag during visualization. To achieve effective spatial data organization, this paper proposes a data storage model with multi-level spatio-temporal organization. Spatially, the trajectory data is partitioned based on Hilbert curve, combined with pre-partitioning mechanism to solve the storage skewing problem of distributed database HBase; temporally, borrowing from the organization of spatio-temporal cube, the spatio-temporal hybrid coding is constructed by using the method of slicing by day and minute system coding to solve the retrieval of trajectory data into maps. The experiment proves that the organization model can effectively improve the data storage and retrieval efficiency, enhance the overall effect of trajectory visualization, and provide effective technical support for data mining and analysis.

In order to provide reliability and safety assessments for safety-critical systems, Model-Based Safety Analysis (MBSA) methods have been developed, of which AltaRica3.0 is one representative method. However, there are still several limitations. The most important one is its inability to verify system behaviors when the system has stochastic behaviors. As a widely used safety analysis and evaluation software, SBIP is an emphasis on formal modeling and statistical analysis of safety-critical systems exhibiting stochastic behaviors. In this paper, we present a transformation approach that integrates the advantages of AltaRica3.0 and SBIP, realizes the function of conducting multiple analyses by modeling the system only once, reduces the modeling cost of safety-critical systems, and improves the reliability and safety analysis efficiency of safety-critical systems.

Aiming at the burst traffic, this paper studies how to perform elastic capacity expansion before the burst traffic based on the current access traffic value in a period of time when there is little or no historical data.In this paper, a resource demand prediction model is constructed to improve the degree of fitting between the predicted and actual burst traffic.A resource feedback model is proposed. It can evaluate the resource utilization of the service node at the current moment through the resource performance indicator detection value obtained from the cloud data center in real time, and then calculate the feedback factor. Finally, the dynamic mapping between the predicted value of load demand and resource scaling volume is realized. Experiment shows that the accuracy of the proposed method is higher than that of the traditional methods.

With the explosive growth of neoconiosis worldwide, medical institutions worldwide generate a large number of chest radiograph images to be annotated for neoconiosis diagnosis every day. Moreover, due to the privacy of patient information, it is not possible to pool medical data from multiple medical institutions to jointly build a fast and accurate medical image annotation system for the task of new coronary pneumonia annotation. To this end, we propose a federal learning-based automatic annotation method that enables multiple medical institutions to design and develop an accurate and robust annotation system for neocoronary pneumonia without sharing data. Experimental results show that the proposed federated learning-based automatic annotation method (93.37% accuracy) is able to protect the privacy of medical image data across hospitals and accomplish a higher accuracy rate of medical image annotation compared to the automatic annotation method constructed by aggregating data (95.29% accuracy).

In 2020, the global spread of Coronavirus Disease 2019 exposed entire world to a severe health crisis. This has limited fast and accurate screening of suspected cases due to equipment shortages and and harsh testing environments. The current diagnosis of suspected cases has benefited greatly from the use of radiographic brain imaging, also including X-ray and scintigraphy, as a crucial addition to screening tests for new coronary pneumonia disease. However, it is impractical to gather enormous volumes of data quickly, which makes it difficult for depth models to be trained. To solve these problems, we obtained a new dataset by data augmentation Mixup method for the used chest CT slices. It uses lung infection segmentation (Inf-Net [1]) in a deep network and adds a learning framework with semi-supervised to form a Mixup-Inf-Net semi-supervised learning framework model to identify COVID-19 infection area from chest CT slices. The system depends primarily on unlabeled data and merely a minimal amount of annotated data is required; therefore, the unlabeled data generated by Mixup provides good assistance. Our framework can be used to improve improve learning and performance. The SemiSeg dataset and the actual 3D CT images that we produced are used in a variety of tests, and the analysis shows that Mixup-Inf-Net semi-supervised outperforms most SOTA segmentation models learning framework model in this study, which also enhances segmentation performance.

At present, the measurement of irregular workpiece is mainly carried out by coordinate measuring machine, but this kind of equipment is expensive and inconvenient to operate, and can’t be applied to soft objects. This paper proposed a 3D surface reconstruction method based on binocular vision to imitate the structure of human eyes, and then measured various sizes on the point cloud model. The reconstruction method based on binocular vision is fast, efficient and lowcost. Then, the 3D coordinates of the feature points in the world coordinate system are solved by the reprejectImgeTo3d function to generate the point cloud. The final point cloud model has rich surface texture information and can completely reflect the 3D size information of the object.

The purpose of this paper is to improve the efficiency of video recognition process and save computational data. According to the encoding and decoding of video, we can know that not every part of the information of each frame is valid in video classification. There are a lot of invalid information, which takes up computing space. Therefore, this paper proposes an efficient video classification and prediction method based on reinforcement learning intra prediction. The process of eliminating time redundancy is further added, and then the video frames with low value are ignored, which further improves the computational efficiency. This method extracts the key frames from the marine biological video, and then focuses on the key image areas of the key frames, so as to reduce the network computing overhead.

In order to analyze the bridge strain signal under explosive load, the sym N wavelet basis function is selected by wavelet analysis method to conduct multi-layer decomposition and reconstruction of the corresponding strain signal, analyze the main characteristics of each frequency band signal, and discuss the frequency band range of the measured strain signal. The results show that the measured strain signal is mainly concentrated in the low frequency part, and the high frequency part is caused by various interference factors; The highest frequency of effective strain signal is mainly below 5KHz, which can provide theoretical reference for strain gauge selection during bridge strain measurement under explosive load.

With the rapid development of automation technology, the processing of various kinds of shaped parts has received the attention of the majority of scholars, such as the identification of automotive door panel weld joints. Automotive door plate weld joint features are small targets, interference, and uneven distribution. The previous solder joint recognition method is affected by the small target and many interferences, and these disturbances lead to inaccurate solder joint identification. To address the problem of inaccurate solder joint recognition, a solder joint recognition method based on the improved YOLOv5 algorithm is proposed. Firstly, the dataset of solder joints is constructed, and then the YOLOv5 network structure is improved by adding a tiny target detection layer，which makes the accuracy of small target detection improved. The accuracy of the improved YOLOv5 for solder joint detection is improved from the original 91.7% to 98.9%, which can be extended to such small target application scenarios.

Coherent wind lidar can detect atmospheric wind field information in real time and accurately. It plays an important role in military aviation applications, numerical weather prediction, meteorological climatology research, wind energy resource utilization and other fields.Coherent wind lidar is based on Doppler effect to realize wind speed measurement. Therefore, the estimation accuracy of Doppler frequency shift directly determines the performance of coherent wind lidar system.In this paper, after the FFT transform of the signal, the spectrum correction effect of the energy centrobaric correction method, the ratio correction method and the spectrum correction algorithm based on sinc function fitting is verified by simulation. Finally, the spectrum correction algorithm based on sinc function fitting is selected to correct the frequency of the measured Doppler signal. The correction effect of the algorithm is obvious, and the purpose of reducing the error and improving the analysis accuracy is achieved.

For the existing face mosaic scheme, all faces in the picture are coded, which destroys the face information of non-target people in the picture, which is not conducive to the subsequent research of the character information in the picture. This paper proposes a face mosaic processing scheme combining retainface and facenet, which realizes face detection with faster speed and high accuracy, can quickly perform mosaic processing, and uses the conditional judgment function to make the program perform face mosaic processing on and only on the target person.

As an important technology to promote office automation, document detection and recognition can improve the efficiency of business processes and user experience, make enterprise business more intelligent, and have very broad application scenarios. In this paper, a document detection and recognition system based on DB detection model and CRNN recognition model is built to detect and recognize document images using 3.64 million samples from the image dataset intercepted by ICDARD 2015 and Chinese corpus, and display the document information in the corresponding table in real time. The test results show that the system effectively improves the model inference speed while ensuring the accuracy of document recognition and detection, and completes the document information entry efficiently and quickly.

This paper proposes a target recognition technique based on target relaxation frequency parameters. These parameters are calculated from the frequency response observed by multi-frequency magnetic induction. Firstly, we prepare multifrequency response and class labels of various materials. The least-square optimization method calculates the prepared materials' relaxation frequency characteristic parameter matrix. Secondly, the matrix is used in the linear regression method to calculate the regression model's undetermined coefficients. Then the coefficients help us to obtain the class label of the target. Finally, we compare the target label with the prepared class labels to identify the substance type of the target. The results show that the technique proposed in this paper can accurately identify and classify the substance type of the target, and the multi-frequency magnetic induction technology based on relaxation frequency can be applied in substance identification.

Flood control is one of the most important operations of Yellow River control. The dam bank danger in flood season every year has the characteristics of time-varying, multi-dimension and complex situation, which poses a certain threat to people's life and property safety. For timely, flood control engineering data quickly and implement management interface of information visualization, effectively reduce the labor intensity of artificial working risk, this paper puts forward a kind of based on three-dimensional library of Cesium Huang Heba shore the realization method of 3D visualization platform adopted the aerial drones to create dam shore entity 3D model, Combined with WebGIS technology and front-end technology, the system's 3D expression is realized. Through interface design and system function development, the system has three-dimensional model display, with the functions of dam bank cruise and measurement tools, the platform has strong sense of reality and good interaction effect, which realizes the three-dimensional visual display of the Madu control project in Zhengzhou, Henan Province.

The ubiquitous WiFi makes it a location service infrastructure and attracts more attention now. We use the cubic spline interpolation method to interpolate the attenuation factor n, and a wide range of indoor corridor environments test results show that the cubic spline interpolation method is more conducive to achieving the precise distance. We also improved the classical trilateration positioning algorithm with a priori knowledge of the specific positioning environment. The improved positioning algorithm significantly reduces positioning errors and improves positioning accuracy. The positioning method is transferred to the embedded positioning node to make distributed positioning with lower calculation cost.

Band selection is an important method for hyperspectral image(HSI) dimension reduction, which can greatly reduce the complexity of HSI analysis. Linear prediction algorithm is an unsupervised band selection algorithm with good classification effect. This paper uses Schmidt’s orthogonalization to improve it. Experiments prove that the optimized algorithm is consistent with linear prediction algorithm in band selection but much more efficient and the complexity of the optimized algorithm is in proportion to the number of selected bands.

Semantic segmentation is widely used in remote sensing data extraction and classification. Existing semantic segmentation networks focus on capturing contextual information in many different ways, simply fusing features at different levels, and ultimately improving the accuracy of semantic segmentation. However, low-level semantic features lack spatial context guidance, and high-level semantic features tend to encode large objects with coarse spatial details, making segmentation results prone to losing fine details. In this paper, we analyze the advantages and disadvantages of different levels of feature maps, and enhance the feature representation from two aspects to solve this problem. On the one hand, inspired by the architectural idea of atrous spatial pyramic pooling (ASPP), we adjust the structure of ASPP module and add the attention module to ASPP, and a new Attention-ASPP(AASPP) module is constructed in this paper. On the other hand, feature information such as boundary contours is enhanced by channel attention modeling, thereby improving local detail representation. Comprehensive experimental results show that our model framework achieves excellent segmentation performance on two public datasets, WHU building dataset and ISPRS Potsdam dataset.

For the data of virtual scene, there are often heterogeneous data from different databases. If the same object is not processed in different databases, it will have a great impact on the subsequent data synchronization. After the synchronization of heterogeneous information, the differences between these heterogeneous information will be eliminated, which will bring great convenience to the subsequent data synchronization. Therefore, this paper is based on the synchronization of heterogeneous information data in remote control and unified description. The utilized heterogeneous data synchronization is feature layer synchronization. Feature sets represented by each data are extracted from each database and are synchronized into feature vectors. Then the synchronization of the feature vectors is carried out. The experimental results show that the method in this paper can accurately warn the abnormal behavior between data synchronization and realize the function of remote control stop. Through the research of virtual control problem and according to the result of data synchronization, warning is carried out to remind the remote operators to control the virtual scene position and posture in real time and to reduce the occurrence of collision accidents.

At present, our country's plastic industry is developing rapidly. When plastic products bring great convenience to people's lives, waste plastic waste also causes huge pollution to the environment. Sorting, in view of the problem that various colors of plastic bottles are mixed in the sorting of waste plastic bottles, and the sorting is more labor-intensive, which leads to the increase of recycling costs. This paper studies the denoising process and color feature extraction before image recognition of plastic bottles, using HSV The model replaces the RGB model for color feature extraction. The recognition rate of the HSV model is 96.67%, and the recognition rate of the RGB model is 95.00%, but the recognition speed of HSV is increased by 16.43%.

The generation of radio frequency thermal plasma contains rich and complex physical field distribution. Correct understanding of these physical field distribution has a guiding role in the application of radio frequency thermal plasma in the industrial field. Radio frequency (RF) thermal plasma technology can be widely used in the industrial field. To obtain more accurate information of the plasma physical field, a three-dimensional numerical simulation was used to study the distribution of the RF thermal plasma physical field. The mathematical model of plasma was established based on three-dimensional numerical simulation. The electromagnetic field equation was set. The relationship between electric field and magnetic induction was obtained. The thermodynamic equation of the plasma physical field was obtained, and the inlet and outlet boundary conditions were set. A plasma physical field distribution simulation algorithm is designed to obtain a physical field simulation distribution result. In the simulation analysis, the voltage, current and electron temperature are simulated and analyzed respectively. The experimental results verify that when the pressure increases, the current and voltage show a certain downward trend, and when the discharge power increases, the current and voltage show a synchronous increasing trend. The electron temperatures at 100 Pa and 205 Pa are between 1.07-1.10 and 1.04-1.07, respectively. The difference between the simulated electron density and the actual electron density is less than 1 × 1019, which indicates that the method has high accuracy. The physical field distribution simulation method provides technical support for the practical application of the plasma in the radio frequency heat, and has great application value. Besides, The research results have important guiding significance for optimizing and controlling the practical application process of RF thermal plasma.

The construction of urban rail transits has formed a cluster effect on the surrounding passenger flow, and the peak passenger flow demand of some lines has exceeded the design expectation, so it is urgent to further improve the transport capacity of the lines. However, since the existing line has been built and put into use for many years, how to improve the transport capacity under the existing constraints has become a difficult problem. According to the operation characteristics of rail transit lines, this paper points out the bottleneck of capacity improvement for existing lines, and puts forward different technical solutions for capacity improvement according to different types of bottleneck points. It is expected to provide valuable technical support for the improvement of transport capacity for existing lines.

A great deal of energy is wasted and lots of traffic jams occur since drivers repeated seeking the available lots to parking vehicle as lacking of information exchange between drivers and the vehicle park spot. In order to addressing the drawbacks of information islands, an improved ant colony algorithm (IACA) is proposed to allocate parking slots for autonomous valet parking vehicles. Compressed global pheromone and defined thresholds are the major improvements that avoids the reduction of global optimization ability and population diversity as falling into local optimal solutions due to the positive feedback effect in the late stage. In order to verify the effectiveness of the presented algorithm, two common test scenarios are designed and three comparison algorithms are developed, including first coming first serve (FCFS), normal ant colony algorithm and immune algorithm. The results show that all the four algorithms could complete the parking slots allocations. However, the cost of proposed algorithm is 13.2% lower than that of FCFS and the convergence speed is 94% higher than that of the normal ant colony algorithm, which means the proposed algorithm has strong generality and better global search ability.

In order to locate the underwater sound in the complex acoustic environment, a localization algorithm based on the modal amplitude correlation is improved. The received acoustic field data was converted from hydrophone domain into modal domain. The modal amplitudes affected slightly by environmental mismatch were selected, and the corresponding modal amplitudes of the copy were extracted from copy of the acoustic field which was calculated by simulation model. The MVDR processor was used to match the two to realize the underwater sound source location, and the location effect was compared with the Bartlett processor. Finally, the accuracy of the MVDR method of underwater sound source location based on modal amplitude correlation was verified by the experiment at sea.

By identifying the working mode of the communication radiation source, its behavioral intent can be comprehended. Tactical data link is a kind of classic communication radiation source with a variety of working modes. Aiming at the typical data link TADIL A, a method for extracting one-dimensional time-frequency features of signals using short-time Fourier transform (STFT) is proposed. Then, one-dimensional Convolutional Neural Network (CNN) DenseNet-1D is used for training and testing to complete the task of identifying different working modes. The experimental results illustrate that the working mode recognition based on the physical layer signal is feasible.

Multiple sensors are often used to work together in vehicle driving, but how to effectively fuse the data of each sensor is a difficult research point. The attention mechanism can assign different weights to each target, allowing more computing resources to focus on key targets, greatly improving computational efficiency and accuracy. In this paper, ResNet34 and SalsaNext are used as encoders of dual-stream networks to extract general features of images and point clouds respectively, and a cascade residual attention fusion strategy is proposed, which is used between two-stream networks to fuse features from two modalities at different encoding stages. Experiments on the SemanticKitti dataset show that this fusion strategy has better performance than single-stage fusion and PMF fusion structures.

One of the important research directions in information extraction is event extraction(EE). It aims at recognizing event types and event arguments from natural language texts, which is an important technical basis for artificial intelligence application that serves for the information work in business, science and technology, military and other fields. Currently, data annotation samples of the relevant field based on encyclopedia and news data are relatively rare and lack relevant datasets. Therefore, there are only a few public research on the event extraction in the relevant intelligence field. By integrating universal information extraction, the event extraction method which use the pre-trained model for the relevant intelligence field can handle the problem of rare data annotation samples for the event extraction in the relevant field. By expanding training samples automatically, the context information of encyclopedia and news data is learnt effectively to extract relevant events from encyclopedia and news data. Compared with other event extraction methods, using precision, recall and F1 value as assessment indicators, in event recognition tasks, the value of F1 enhances 1.26%, and in argument recognition tasks, the value of F1 enhances 1.58%. The method can significantly boost the extraction performance in small samples shown in the experimental results.

In the processing, manufacturing, and production of modern fields, rolling bearings, the most basic module of most mechanical equipment, have a key role that cannot be ignored. This paper proposes three fault diagnosis model architectures for rolling bearings based on the deep convolutional neural network. Three models were tested on the industry-common Case Western Reserve University Dataset (CWRU). The original vibration signal acquisition and processing module mainly uses the vibration signal window translation method to complete the segmentation of overlapping signals and uses the Inception network to efficiently complete one-dimensional signal preprocessing. Finally, we use t-distributed stochastic neighbor embedding (t-SNE) to reduce the dimension and visualize the learned fault data distribution.

This paper presents a design method of basic mathematical simulation model of damage effect. Based on this method, the basic components and software framework of damage effect simulation are realized. This method is composed of coordinate system definition and coordinate conversion components, plane normal vector simulation basic components, maneuver route change simulation basic components, 2D situation display components, and so on.

This paper presents a simulation algorithm for target to ground fire distribution. Taking battalion as an example, the algorithm for battalion level target to ground fire distribution includes the calculation model of optimal strike relationship, the cost function of missile control fire distribution and the solution method of linear programming. Based on this algorithm, the software is implemented and applied to related projects. Combined with the actual project, good results have been achieved.

Load forecasting is the use of historical data to build a model based on the operating characteristics of the power system, economic conditions, social conditions and natural conditions, and the use of this model to make reliable estimates of the trend of load changes in future periods. Load management is of great importance to the development of many tasks in the power system. Not only does it require comprehensive consideration of many aspects of power dispatch control, operation planning and marketing operations, but it also requires scientific forecasting of power loads in conjunction with actual practice. Short-term load forecasting is an important basis for the power supply company's scheduling plan, helping the staff to monitor the working status of the power system in real time, and is of great significance to the power supply company in reducing costs and the power plant in making power generation plans. However, the load forecasting models established by current research workers do not meet the more stringent forecasting requirements of power supply companies in terms of both accuracy and stability. The main objective of this paper is to investigate the automatic scheduling of short-term load resources for power grids based on a weighted plain Bayesian algorithm. Based on the periodicity and regularity of the historical load, this paper divides the fluctuating intervals of the sequence according to the trend of load changes, extracts the characteristics of the fluctuating intervals to represent the load intervals as electricity consumption patterns and performs density clustering. Based on the clustering results, similar load sequences are selected from the aggregated clusters to determine the reference sequences, and then a load forecasting model is established for load forecasting by combining time series and regression ideas to achieve ultra-short-term load forecasting. The test results show that the method has a good forecasting effect and provides a forecasting method based on electricity consumption patterns for power management departments.

COVID-19 epidemic is not over. The correct wearing of masks can effectively prevent the spread of the virus. Aiming at a series of problems of existing mask-wearing detection algorithms, such as only detecting whether to wear or not, being unable to detect whether to wear correctly, difficulty in detecting small targets in dense scenes, and low detection accuracy, It is suggested to use a better algorithm based on YOLOv5s. It improves the generalization and transmission performance of the model by changing the ACON activation function. Then Bifpn is used to replace PAN to effectively integrate the target features of different sizes extracted by the network. Finally, To enable the network to pay attention to a wide area, CA is introduced to the backbone. This embeds the location information into the channel attention.

Constant False Alarm Rate (CFAR) is an effective means to reduce the probability of false alarm in signal detection. In this paper, the target characteristics and radar echo characteristics based on dense false targets or chaff diluted interference are analyzed, and based on this, an improved algorithm of Cell Averaging(CA)-CFAR is proposed, the number of range gates of the protection unit is changed, and the data based analysis is carried out, which significantly improves the detection probability of the signal.

To solve the issues with industrial traditional area measurement instruments equipped with infrared sensors, which are bulky and compute slowly, a planimetric measuring method based on machine vision is developed. The area of the plane is calculated through three processes: 3-D reconstruction by monocular camera calibration, top view by perspective transformation and image segmentation to obtain the proportion of the object to be measured. This paper enhances the calculation of the world coordinate points in the 3-D reconstruction process and lowers the measurement error from more than 1% to 0.28%. It solves the drawbacks of the conventional measuring device, which is huge and slow, and can correctly measure the planar area with less expensive hardware, which has certain application value.

Aiming at the problems of low detection accuracy, poor real-time and robustness in vehicle target detection in the field of autonomous driving and other fields of existing target detection algorithms, this paper proposes a model based on YOLOv5s and introduces an attention mechanism to fuse multi-scale features. The detection algorithm of YOLOv5s model adds a layer of CBAM(Convolutional Block Attention Module) before the SPPF of Backbone of the YOLOv5s model, and uses ACON(Acrivate or Not) to replace ReLU as the activation function of the network, which improves the detection accuracy and can be effectively applied to vehicle detection in complex scenes.

In order to solve the problems of low detection accuracy and poor robustness of fruits under natural conditions, an apple detection model based on improved yolov3 is proposed with apples in natural environment as the research object. Firstly, the ordinary convolution inside the residual network in DarkNet53 is replaced with a depth-separable convolution to improve the detection speed of the model while reducing the number of parameters in the network; secondly, the channel-space attention mechanism (Convolutional Block Attention Module, CBAM)module is added to the process of feature pyramid upsampling to further strengthen the model feature extraction and improve the recognition accuracy of overlapping apples; finally, the introduction of CIOU Loss complete cross-comparison loss function, which integrates the influence of the overlapping area, aspect ratio, centroid distance and other factors on the bounding box regression, makes the bounding box regression stable while the target detection accuracy is also improved. The experiments show that the improved model achieves a MAP (Mean Average Precision) value of 91.30% and an F1 value of 86% under the test set, which is 4.2 percentage points and 5 percentage points hight than the original model respectively.

This article presents, for the first time, a custom-built passive RFID tag with large ferrite-core antenna for underground detection. Based on the equivalent circuit of tag, the parameters affecting the sensitivity of the ferrite-core tag antenna are studied. Then, through the theoretical analysis of the interactions between these parameters and sensitivity, the design of ferrite core and antenna coil winding are given in detail. Finally, the working distances of the two fabricated ferritecore tag coil antennas are measured with two different reader antennas in a simulated underground environment, and the results show that a detection depth of more than 2 meters can be realized by the presented passive RFID tag with large ferrite-core antenna.

With the continuous development of modern network and communication technology, in view of the people to the location service especially rising demand based on the location of the indoor positioning services, as well as the existing indoor positioning method is complex and positioning accuracy is not high question, this paper proposes a fusion from coding and bluetooth stack location fingerprinting indoor localization algorithm has high precision, The algorithm using the stack from coding network first determine the coarse position of target, and use the network to predict location for the secondary position, finally use the fingerprint matching algorithm for accurate positioning of the final, average position precision can be up to 25 cm, finally through the simulation analysis this ask the performance of the proposed algorithm, this algorithm is verified indoor location can reduce the computation complexity, It can also improve the positioning accuracy.

Segmentation of SAR ship target is fundamental to ship recognition. Due to the ship target motion and SAR image quality, the azimuth ambiguity and side lobe effect in the ship target SAR image will affect the segmentation accuracy. This paper proposes a SAR image segmentation algorithm based on the Radon transform, which first extracted the target azimuth angle and the direction orthogonal to the principal axis through the Radon transform, and then conducted segmentation in the area. Experimental results show that the proposed algorithm has a good segmentation effect for medium and high resolution SAR ship target.

Wireless backscatter is a novel passive communication technology, which enables battery-free devices to harvest energy from environmental RF signals and communicate with other devices via backscattering environmental signals. Wireless backscatter is ideal for the IoT. However, the existing wireless backscatter has several drawbacks, including a high bit error rate, expensive deployment costs, and high signal-noise ratio requirements. We proposed a distributed wireless backscatter system to address the aforementioned issues. Our design introduces multiple input multiple output (MIMO) system into wireless backscatter. With MIMO, backscatter communication can work at the same frequency as Wi-Fi. Moreover, the proposed distributed protocol increases the opportunity for backscatter. Extensive simulations show that our design significantly increases system throughput.

As the environment of special targets is complex and constantly changing, better requirements are put forward for rapid and high-precision target detection and recognition. In this paper, the improved YOLOV7 algorithm is adopted. First, Kmeans is used to match the new anchor coordinates, and multiple detection scales are added to improve the detection accuracy; Secondly, the attention mechanism module is integrated into the feature extraction network Darknet-53 to obtain important features; Then, taking advantage of the lightweight technology of Ghost module, Ghost BottleNeck composed of Ghost modules is introduced to replace the Neck module in YOLOV7, which greatly reduces the parameters and computation of the network model; Finally, IOU_ Nms is modified to DIOU_ Nms is used to optimize the loss function. experiments show that the accuracy and real-time performance of the algorithm are improved.

Ultra-dense network (UDN) can provide extremely high throughput and data rates. However, the dense deployment of access points (APs) causes severe interference. In order to reduce interference and effectively allocate network resources while ensuring the user's quality of service (QoS), this paper proposes a UDN resource joint allocation scheme based on cooperative computing and genetic algorithm. This algorithm solves the problem of the slow convergence of the traditional genetic algorithm and can also prevent potential split-brain problems for overlapping APs. The simulation results show that the algorithm can guarantee the system's performance while avoiding the high computational complexity problem caused by global resource optimization.

Due to the terrain, the frequent floods of the Yellow River have posed a great threat to the cities along the Yellow River. At present, although the danger alarm system and video surveillance system are widely installed along the Yellow River Basin, there is a lack of linkage between the two, which wastes monitoring resources. In addition, the push rules of different types of alarm signals are relatively complex, and the coupling between rules and system implementation is high, which is not conducive to rule modification and expansion; moreover, there are many cities along the Yellow River, if each city customizes its own system, it will lead to more overhead. Therefore, this problem can be solved by improving the hierarchical scalability of the system. So, based on the Drools framework, this paper uses the rule engine to realize the unified processing of different rules and the linkage integration of sensors and cameras, and designs a recursive model to simplify the development process, so as to realize the accurate positioning of alarm information at all levels of provinces, cities and counties, and the hierarchical processing of various alarm information.

With the rapid development of rail transit technology in China, this paper aims at the problem of train environmental monitoring and designs carriage monitoring system platform. Using Zigbee we can monitor the temperature humidity and gas concentration real- time. Matlab is used to establish a mathematical model related among temperature humidity and smoke in order to make people feel comfort and guarantee the safety.

NDVI (Normalized Difference Vegetation Index) time series usually contain a lot of loud noise, which limits its further application. However, the existing filtering and reconstruction methods cannot effectively remove continuous loud noise, which is particularly obvious in cloudy areas. This paper proposes a Spatial-temporal Kriging improved Savitzky Golay filtering algorithm (TSK-SG) based on Spatial-temporal Kriging. By combining the quality factors in MODIS VI products to generate reference data, a Spatial-temporal Kriging variogram model is established and interpolated using the adjacent Spatial-temporal information. Finally, the fitting result is obtained by iterating Savitzky Golay (S-G) filtering based on the quality weight. The NDVI time series curve reconstructed by this method can effectively suppress noise and has a better spatial reconstruction effect, which can better reflect the phenological characteristics of different types of crops.

With the development of remote sensing products towards the direction of civilianization and popularization, work-flow customization plays an important role in the production of remote sensing products. The traditional customized work-flow model has large time cost, complex operation and high requirements for professional knowledge. The work-flow recommendation system can improve the construction efficiency of remote sensing work-flow to some extent and assist users to design high-quality remote sensing work-flow models. However, most of the existing remote sensing work-flow modeling methods ignore the logical structure characteristics of the work-flow, leading to large errors in the calculation results of similarity. Difficult to make work-flow recommendations effectively. Therefore, this paper proposes a customized recommendation algorithm for remote sensing work-flow based on logical structure. By focusing on logical structure, the reliability of similarity calculation between work-flow is improved, so as to find similar work-flow to help users recommend the next modeling node. Firstly, the work-flow model needs to be preprocessed: the user converts the constructed work-flow model into a process structure tree by using Petri net workflow, and uses the path table generation algorithm based on logical structure to convert the model information into data information and store it in the database for subsequent data processing; Then the flow data in the flow tree set was converted into a path table according to certain rules, and then the longest common subsequence similarity of each data in the path table was calculated to obtain the similarity calculation results based on the logical structure characteristics, the most similar work-flow in the work-flow library is found and the recommendation is made for the user. The method proposed in this paper is evaluated experimentally on the real data set, in terms of recall, precision and F1-score, which shows that the method proposed in this paper can effectively improve the recommendation efficiency and meet the actual needs of users.

Homography estimation is often an indispensable step in computer vision tasks that require multi-frame time-domain information. However, when we estimate the traditional homography matrix, the rotational and translational terms are often difficult to balance. In this paper, based on the 4-point homography parameter matrix, we reproduce the Synthetic COCO dataset (S-COCO) and the Photometrically Distorted Synthetic COCO dataset (PDS-COCO). Then, we use the Darknet in YOLOv3 as the backbone to design a deep network for 4-point homography estimation. Experiments show that compared with existing main one-stop methods, our proposed deep learning network achieves the best performance on the S-COCO dataset and excellent performance on the PDS-COCO dataset.

In the gun sound location system, how to use more information to quickly locate is a very important issue. In the current study, the muzzle wave has been used for single-point positioning to determine the position of the sniper, but it lacks the ballistic positioning of the flying missile. Tracking and localization using the Mach waves generated by the bullet's flight can also estimate the location of the sniper. The trajectory after the Kalman filter is more in line with the actual bullet flight trajectory, and the position of the sniper can be well resolved by using the least squares estimation. Finally, the CRLB of the proposed method is derived, and the effectiveness of the proposed method is verified by simulation.

This paper is mainly based on the artificial intelligence technology to automatically detect the visual inspection content of the top, bottom and both sides of the car body of the EMU. Specifically, image recognition is used to complete the fault identification and classification of the bottom, top and both sides of the car body of the EMU, and improve the detection efficiency. Specifically, the overall architecture of the omnidirectional fault detection system for multiple units is designed around the application of the maintenance operation process of the bottom, roof and both sides of the car body of the multiple units. The functional modules of the fault detection system for the bottom of the multiple units and the fault detection system for the roof and car body of the multiple units are analyzed and designed respectively, and the overall scheme design of the omni directional fault detection system for multiple units is completed.

Aiming at a series of problems such as low evaluation efficiency of traditional algorithms, cumbersome process operation, high use cost and no complete algorithm inspection process system, this paper proposes an algorithm flow technology inspection processing system and method based on remote sensing applications, mainly including satellite product processing, algorithm processing, sample feature selection, sample extraction, and algorithm processing results comparison, It mainly realizes the functions of input, output and parameter adjustment required by different function functions, and designs an interface interaction system that adapts to the input and output elements required by different function modules. It realizes the modular, plug-in and componentized design of interface interaction of function modules, and obtains the evaluation system of algorithm credibility.

Numerical simulation is the default method for the study of shock wave evolution process, but it requires thousands of CPU cores and takes a long time for computation. Moreover, when the initial and boundary conditions change, the simulation needs to be recalculated. Therefore, numerical simulation cannot meet the needs of rapid evaluation and real-time display. In this paper, an encoder-decoder network structure based on ConvGRU-ODE is proposed to accelerate the simulation of shock wave evolution process, Given the experiment parameters, the proposed method can calculate the pressure effect field at each time by only using the calculation parameters. The algorithm only takes 120 seconds to calculate the shock wave evolution process of a single scene, which is significantly faster than the numerical simulation. The average relative error of pressure effect field in important areas is less than 10%, which is acceptable in engineering practice. The algorithm has good adaptability and can be applied to different scenes.

P91 steel is widely used in high temperature components of ultra supercritical thermal power units, whose high steam parameters can improve the thermal efficiency. Creep cavities can be used to evaluated the creep damage state of the P91 power plant steel. In this study, industrial computed tomography (CT) tests were conducted on the crept P91 steel sample. It was found that the digital radiography (DR) and CT images using 12 W target power exhibited clearer internal defect morphology than that of the images taken with 80 W target power, because the relatively low target power kept the focal spot small. In addition, combined the CT images along radial and axial directions, the creep cavities of P91 steel were judged as long strip volume defects with ~50 μm diameter along the stress direction, demonstrating that the industrial CT testing method could be applied to defect detection of other power plant materials.

The main content of material management in electric power enterprises is to allocate and optimize storage materials. For warehousing goods, ensuring the storage location of goods can make the highest space utilization, the lowest energy consumption, and the lowest cost, which is of great significance to the study of goods location-allocation. Through the research on the state of the automated warehouse, storage location assignment and optimization, and the optimization of in-out warehouses, this paper establishes the optimization model of storage location assignment and the method of dimension reduction. Through the discretization of dimension step and displacement, the algorithm is suitable for discrete problems. Through reverse mutation, dimension exchange, and other operations, the initial state is optimized. The algorithm of storage location assignment is formed, so that the algorithm can get better results in the locationallocation problem. Adjusting and improving the algorithm will make it more effective in the problem of location assignment. It can solve a series of problems existing in the storage system, which is to achieve the requirements of a better location, higher efficiency and faster response, and also improve the picking efficiency.

The Starlink is a giant Low Earth Orbit (LEO) constellation. Firstly, this paper uses the interface provided by Satellite Tool Kit (STK) and Matlab programming to simulate, calculate the coverage and verify the visibility algorithm. In addition, satellite constellation can provide reconnaissance, surveillance, communication, navigation and other capabilities while carrying sensors. Based on the satellite orbit prediction and sensor coverage, the visibility of the satellite to the target area is to calculate the time period when the satellite passes through the specified area in the prediction cycle. Finally, according to the calculation of instantaneous ground coverage area by sensors with different shapes and attitude angles, whether the satellite coverage area overlaps with the target area can be abstracted as judging the geometric relationship of polygons on the plane.

The fixed window algorithm used in the traditional vehicle flow virtual loop detection has inherent shortcomings and limitations. The traditional algorithm can achieve a good detection effect when the driving vehicles strictly follow the lane. However, in real life, the phenomenon of vehicles driving across lanes is relatively serious, which often leads to the simultaneous detection of vehicles on the detection lines of two adjacent lanes, thus causing false detection. Aiming at the shortcomings of traditional fixed window algorithm in traffic flow detection, this paper proposes an improved dynamic window algorithm, which can effectively overcome the shortcomings of fixed window detection by setting two detection lines on multiple lanes across the entire road. The experimental results show that the improved algorithm has achieved good results in traffic flow detection, and has good robustness and stability.

Aiming at the problems of low recognition accuracy and inexplicable features extracted by neural networks in signal modulation recognition algorithms based on deep learning, this paper proposes a CLSTM recognition algorithm based on IQ signal features. The algorithm first uses a 2×1 convolution kernel to extract the phase feature of the signal, then uses a multiple convolution layer to extract the temporal feature of the signal, and then outputs the result of the convolution layer to the LSTM module to further extract the temporal feature, finally, the recognition result is output by the fully connected layer and the recognition effect is improved by optimizing the network structure. The experimental results show that on the RML2016.10a public dataset, when the SNR is higher than 0 dB, the recognition accuracy of the algorithm reaches 89.21%, which is 15.02% and 9.85% higher than the classical CNN2 and CLDNN network algorithms.

In the field of Internet of Things middleware, this paper analyses the challenges to build application support platform and reviews the current research work. The paper presents a solution combining hardware and software on the basis of hybrid middleware architecture. The solution is designed to enrich interface types of access gateway, improve the flexibility of middleware architecture, and enhance user programming performance through graphical programming. Finally, the process of developing and deploying applications with application support platform is introduced through test experiments, which shows the practicability of application support platform.

In order to solve the problem of high leakage rate of the access layer network traffic detection method of power IoT, a Monte Carlo method based on the access layer network traffic detection method of power IoT is designed. Combining techniques such as higher-order modulation, improving the power IoT architecture, calculating the degree of randomness of the data blocks obtained quantitatively, extracting access layer features and security requirements, resetting the power window values again, and optimising the network traffic detection model based on Monte Carlo methods. Experimental results: The mean values of the leakage rate of the power IoT access layer network traffic detection method in the paper are: 1.355%, 17.003% and 5.223% for DoS attack, client poisoning and SSH scenarios respectively, indicating that the advantages of the designed power IoT access layer network traffic detection method are more obvious after the full integration of the Monte Carlo method.

In view of the low contrast and large colour deviation of the image captured by the camera in foggy weather, it is proposed to modify the parameters of atmospheric light and transmittance through dual channels, refine the transmittance by the tolerance mechanism, and obtain the dehazed image in reverse according to the atmospheric scattering model function. Combined with the parallel operation and high-speed characteristics of FPGA (Field Programmable Gate Array) hardware platform, the proposed algorithm is parallelized and simplified, and the implementation of dark channel dehazing hardware based on FPGA is designed. After testing by MATLAB and ModelSim, the restored image is clear and distortion-free.

In order to reduce the following error of traditional pure pursuit algorithm, a Stanley compensatory pure pursuit algorithm is proposed. The simplified model of intelligent vehicle is established, and the front wheel angle compensation is used to ensure that the intelligent vehicle follows the expected path accurately. The simulation results show that the improved following control algorithm is effective, the maximum, average and standard deviation of the following error are reduced by 64.23%, 39.52% and 27.47% respectively. The simulation results show that the robustness of the improved pure pursuit algorithm is enhanced, and the intelligent vehicle can achieve stable, reliable and high precision tracking control effect.

With the increasing application of vehicle Ethernet in the intelligent networked vehicle network architecture, how to ensure the safety of Ethernet communication data between vehicle controllers has become an important research direction in the field of vehicle information security. Based on the Strongswan open source project, this paper has carried out the national security transformation of IPSec protocol, and implemented and verified it on the embedded development board with NXP I.MX8 as the core, proving that it can be applied to the Ethernet communication scenario between vehicle embedded controllers.

In order to ensure the safety and reliability of underground cable trench environment, the remote environmental monitoring technology of pipeline is studied in this paper. The feasibility of low-power wide-area Internet of Things NBIoT technology and local-area Internet of Things ZigBee/Lora/BLE technology in remote monitoring of complex environment in underground cable trench is analyzed in detail. It can meet the reliability and speed requirements of collected data transmission. The system can dynamically evaluate the service life of key hardware equipment in different working modes through data acquisition, so as to ensure the stable operation of power grid, and provide a reference for the future construction of utility tunnel. The acquisition and transmission of the terminal data are tested, and the optimal power parameters are determined in the test to ensure the real-time accuracy of the data transmitted between the terminal node and the coordinator. The real-time data and state monitoring of the PC terminal are tested, and the error within a reasonable range is calculated by using the relative ratio of the actual measured data information and the data information received by the monitoring system interface, indicating that the real-time state monitoring system of the upper computer can operate normally. The entire smart grid monitoring system can achieve the expected design goals.

With the rapid and in-depth integration of various emerging wireless technologies into all walks of life, the massive terminals in the "Internet of Everything" carry diversified business and bring higher requirements for users' network service experience. How to provide users with personalized services and improve users' experience and satisfaction in the dynamic network environment formed by terminal mobility and network state variability is the key to network switching research.For dense terminal scenarios, in order to avoid frequent network selection caused by network congestion and network load jitter caused by blind switching of dense terminals, a load prediction vertical switching algorithm based on GA optimization is proposed. Firstly, the probability of load state space is predicted by the Markov chain, and the predicted probability is mapped to the network load trend value by the load trend function. Then, the load trend value and important network parameters are used as the decision criterion, and the cost function is established to make a comprehensive decision so as to complete the network switching reasonably and accurately. A genetic Algorithm (GA) was introduced to optimize the weights of the decision criteria, and the optimal weight combination was obtained to minimize the switching times. Simulation results show that the proposed algorithm can effectively balance the load between networks, meet the needs of users with less switching, improve the ping-pong effect, reduce the blocking rate, and finally, effectively improve the quality of user experience.

According to the needs of the field of chemical defence support simulation, this paper proposes a design method of chemical defence support simulation support software, and realizes the corresponding software based on this method. The method includes chemical defence support simulation initialization and data receiving module, chemical defence support equipment simulation module, chemical defence support command simulation module, chemical defence reconnaissance simulation module, decontamination simulation module, smoke and fire spray simulation module The simulation data output and recording module of chemical defence support is composed. Based on this method, the simulation problem of chemical defence support can be solved.

The underwater multi-node is a device that can sample and process underwater acousto-magnetic signals. Due to the nodes contain digital, analog, and digital-analog hybrid interfaces, and different types of functional modules are mixed together, the test of the node circuit is not easy to achieve. In this article, I designed a set of test software for underwater multi-node circuits. It can automatically monitor the power supply of the node circuit and the status information of the interface, and can also quickly judge whether the function of the circuit and the interface are wrong.

The accurate estimation of the road adhesion coefficient can provide a judgment basis for the accurate decision of the vehicle active safety system. It can accurately reflect the interaction between tires and road surface. The decrease of the adhesion coefficient on wet road surface increases the probability of traffic accidents. The vehicle dynamics model and tire model are established in simulink, and the vehicle operating environment is established in CarSim. The proposed algorithm is verified by co-simulation.

Rice is one of the most important food crops, and the stability of its yield is crucial to national food security. Rice plant counting becomes critical in fine agriculture research and is closely related to rice yield. In this paper, we proposed a new multi-scale fusion based rice counting method (RPCNet), which consists of one feature extractor frontend and two feature decoder modules namely Density Map Estimator (DME) and Plant Object Recognizer (POR). Fineloss is introduced into DME to improve the network's ability to separate adherent plants. To verify the validity of our method, we conducted experiments on a high-throughput rice plant image dataset. Experiment results show that the MAE and RMSE of the proposed RPCNet are 2.6 and 3.4 respectively, which outperforms states-of-the-art methods. Results suggest that RPCNet can accurately and efficiently estimate the number of rice plants and replace traditional manual counting.

Autonomous driving technology, mobile robot technology and AGV vehicle technology in warehousing and logistics have been a hot topic of research in recent years, path planning plays a pivotal role in these technologies. A good path can improve the operation efficiency of the vehicle robot, reduce its energy consumption and improve its endurance. In this paper, the A-star algorithm is optimized and improved from its planning speed, the number of search points and turning points. Finally using matlab simulation verification, the final proof that the average planning speed increased by more than 20%, the number of search points reduced by more than 25%, the number of turning points reduced by more than 30%.

The increasing perfection of BeiDou positioning system has increased the market of application terminals, but positioning accuracy has become a major problem for users. According to the advantages of Kalman Filtering algorithm, such as time continuity and noise reduction, this paper briefly introduces the principle of pseudo range positioning, introduces, simulates and analyzes the error of least square method and Kalman Filtering method, and proposes a traceless Kalman Filtering algorithm to improve the least square method for adaptive noise estimation. Through MATLAB simulation, the advantages and disadvantages of the algorithm performance are compared, and the feasibility of the algorithm in improving the positioning accuracy is verified.

Aiming at the problem of many aeroengine monitoring parameters, large amount of data, and timeliness of data, a novel aero-engine Remaining Useful Life (RUL) prediction method based on Temporal convolutional network (TCN) was proposed. Firstly, the data were redivided by setting different sliding window lengths, and then the optimal parameter selection of the model was studied. Finally, the remaining useful life prediction results of this method and traditional methods were compared and analyzed. The results showed that: The different parameters affected the conclusion of the calculation of the model. When the sliding window length was 30, the batch_size was 64, the dropout was 0.1, and the kenel_size was 8, the model had good prediction results. The best deterministic correlation coefficient between the predicted value and the actual value was 0.86, and the predicted trend of change was basically consistent with the actual value. The root mean square error of the model was 19.85, which was parallel to Long short-term memory (LSTM) and Convolutional neural networks (CNN), and the result verified the effectiveness of the method in predicting the remaining useful life of the engine. Through the above research, it provided a new model reference for solving the problem of engine remaining useful life prediction.

Aiming at the problems of slow convergence speed and easy to fall into the local optimal solution in the path planning of mobile robots in the application of traditional ant colony algorithm, a strategy to update the pheromone increment according to the planned path and the historical optimal path is proposed, so as to improve the convergence of the algorithm. To avoid the algorithm getting stuck in stagnation, an adaptive pheromone volatility coefficient is introduced. It not only avoids the stagnation phenomenon in the search, but also improves the path convergence speed. Simulation results show that the improved algorithm has good convergence.

Some customer satisfaction measurement systems suffer from long response times. A customer satisfaction measurement system based on grey fuzzy theory is designed to improve the timeliness of this type of system. The hardware logic connects the output of TIM2 CH2 timer 2 of STM32 with the FSELECT pin input of AD9831 to switch the modulation frequency to send FSK signal; identify the customer type, extract the initial variables of customer satisfaction, use the top few principal components as the unrotated common factors, construct the satisfaction evaluation matrix based on grey fuzzy theory and optimise the system The function of the measurement management module was optimized. Experimental results: The average response time of the designed customer satisfaction evaluation system is 71.084ms, which indicates that the customer satisfaction evaluation system is more universal under the influence of grey fuzzy theory.

In order to further improve the accuracy and generalization performance of traditional Support Vector Machine (SVM) model and improve the problems of low accuracy and poor generalization of the constructed classifier, the penalty parameter C and the kernel parameter of the support vector machine (SVM) are improved using the sparrow search algorithm (t-GSSA) based on adaptive t distribution with golden sine improvement. The tGSSA-SVM breast cancer identification model was developed by performing the optimization search. The Wisconsin breast cancer data set (WBSCD) was applied for experiments and compared with the traditional SVM, PSO-SVM, and WOA-SVM. The results showed that the performance of the optimized tGSSA-SVM diagnostic model was improved compared with the existing methods in terms of accuracy, etc., and it was a better modeling algorithm.

The existing software implementation schemes of Convolutional Neural Networks (CNN) cannot meet the requirements of computing performance and power consumption. To further improve the energy efficiency of the deep neural network, improve throughput and reduce power consumption, a hardware accelerator based on a convolutional neural network was designed, and a verification platform was built for it. The platform has good reusability and can flexibly complete the verification work of the target chip under various modes and configurations, and perform performance evaluation and functional correctness verification on the chip. Through the board-level verification results, this design reduces power consumption by 12.36% compared with similar accelerators and improves hardware resource utilization by 13.87% while keeping the same conditions as clock frequency and bus bit width.

Aiming at virtual applications, a prototype of affine transformation for designing patterns is proposed in this paper. Due to the capability of generating fractal geometry with self-similar and micro-fine structure by simple affine transformation, the iterative function system (IFS) is utilized to model the prototype, which is comprised of prototype the affine transform, computing the parameters of IFS ,designing and implementing iterative algorithms, and generating basic fractal sets as well. Simulation experiments has shown the validity of the proposed method.

Metamorphic testing is an effective way to alleviate the problems of testing Oracle, for which metamorphic relations are key. Studies have shown that the complexity of metamorphic expressions is related to their error detection capability. Therefore, an effective method for calculating the complexity of metamorphic relations can improve the efficiency of metamorphic testing to a certain extent. To address this issue, we analyse the types of input and output components of metamorphic relations described by expressions used in numerical computation programs, define the complexity of metamorphic relations according to the scale complexity in the classification of complexity concepts, and give a method for calculating the complexity of metamorphic relations. The experiments show that the complexity of the metamorphic relationship derived from the metamorphic relationship complexity calculation method is roughly positively correlated with its error detection capability. The complexity of the metamorphic relationship can provide a reference for the selection of metamorphic relationships.

Abnormal prefrontal cortex neurons have caused a variety of mental and brain disorders, which seriously affects health of people all over the world. Researchers propose a series of modulation treatments. In this paper, we propose a new fuzzy modulation strategy for the abnormal prefrontal cortex rhythm, and design a fuzzy proportion integration differentiation modulation based on fuzzy inference. This study is helpful for various brain diseases caused by abnormal brain rhythm, and is significant for the treatment of brain diseases and life health.

With the development of network technology and people's growing need for a better life, IoT technology has been integrated into more and more people's daily lives. At the same time, due to the different needs of individual users and the differences in the function of networking equipment spawned a variety of Internet of things products. Hence an IoT platform that is convenient, efficient and easy to expand is particularly important. This article designs and implements an integrated IoT platform for many types of persistent connected IoT devices on the market. The platform supports concurrent access of large-scale IoT devices with multiple sources and protocols, and supports remote management of devices through instructions. Monitor the health of the equipment in real time and alarm in real time in case of abnormalities.

The traditional network security situational awareness method obtains the clustering results of network behavior analysis by using adaptive weight clustering algorithm, which leads to poor perception due to the lack of fusion of situational indicators. In this regard, a cloud computing-based network security situational awareness approach is proposed. By using Hadoop cloud computing network platform to collect security posture information and establish Bayesian network to fuse network security posture indicators and construct a posture perception model. In the experiments, the proposed method is verified for the perception effect. The analysis of the experimental results shows that the proposed method is used for network security situational awareness of network data, and the perception is less time-consuming and has a more excellent performance of network security situational awareness.

In the era of information explosion, language learners can easily obtain a large number of language learning materials. It is a common way to obtain learning materials of the target language learned from streaming media. However, streaming media platforms generally only inspect the compliance of the content in terms of politics and economy, while the inspection of language is relatively lacking. This makes it difficult for language learners to determine whether the text of the streaming media is suitable for their own language proficiency level. Therefore, it is necessary to research on the architecture of network text readability rating system based on streaming media. In this paper, PocketSphinx, SpeechRecognition and other add-on packages based on Python platform have conducted voice recognition on streaming media information collected from user equipment, and then conducted readability grade evaluation on the recognized text, and analyzed the language characteristics of the text, so that language learners can learn corresponding language knowledge.

Computer networks have brought a very great transformation to English teaching, this transformation has more advantages than disadvantages. In recent years, people have conducted in-depth theoretical and practical researches on the integration of information technology and college English courses, certain research results have been achieved. In this paper, we take the intelligent classroom formed after the organic integration of computer network and English teaching courses as an example and elaborate the design of the new model of intelligent teaching.

With the surge in the amount of data transmitted on the network, intelligent learning and other technologies have emerged to solve the problem of anomaly detection of streaming data in large data. For network security issues, based on the extraction of network traffic characteristics, network traffic classification or clustering is an important technical means to discover network faults and network attacks. In this paper, a distributed detection framework for detecting anomalous behaviors of encrypted network traffic is proposed. The intelligent router is adopted to obtain the encrypted network traffic monitoring stub, and then the neural network codec is used to adaptively learn the characteristics of the encrypted traffic and identify the abnormal behavior of the encrypted protocol traffic. A wide coverage traffic pattern extraction algorithm based on the network state sequence is designed to obtain the traffic patterns that represent the network conditions of the data center. Finally, the simulation test verifies that the model has frequent traffic patterns with priority. The performance of the network anomaly detection model is better than other detection methods, which improves the accuracy of detection and has a better recognition effect.

Starting from the demand analysis, based on the intelligent six axis robot combined with ultrasonic flaw detection equipment, the advanced robot technology and ultrasonic acquisition, processing and identification technology are used to build a flaw detection identification acquisition system for the wheel set robot of the EMU. The goal is to realize the automatic collection and transmission of the detection information of the wheel set defects of the EMU, the automatic early warning and reporting of abnormal conditions, the centralized management and resource sharing of the monitoring data, and to form the EMU wheel set flaw detector human system that integrates monitoring automation, intelligent control and management decision-making. Focusing on the application of EMU wheel set flaw detection robot in the EMU maintenance process, this paper designs the overall architecture and physical architecture of the EMU wheel set flaw detection robot system, and designs its functional modules and interface modules. Finally, the overall scheme design of the EMU wheel set flaw detection robot system is completed.

In the area of social network, different attributes have different effects on the structure of network. Most of the existing privacy protection methods for attributed networks ignore the situation which different attributes have different effects on the network structure. They protect the privacy of the attributes indiscriminately. In respect of the issues above, a differentially private discrete multi-attributed network releasing method is proposed. Firstly, a probability model of discrete multi-attributed network is structured and the correlation parameter between multiple attributes and network structure is defined. The factor with different effects of different attributes on network structure is added into the model. Then, the algorithm uses the correlation parameter to establish the partition model of metadata and divides the metadata into different groups. As the group has different network model and attribute between each other, the groups are independence. The differential privacy of discrete multi-attributed network is realized through sanitizing parameters of the model and allocating metadata using exponential mechanism. Finally, experiment on real datasets verifies that the algorithm can satisfy the characteristics of the discrete multi-attributed network. It can also improve the efficiency and data availability.

Elastic buffer is often used in the receiving module of the physical layer of serial high-speed interface, which is used to compensate the frequency and synchronize the phase of the recovered clock and the local clock in the physical layer receiver. Based on the protocol of PCIe4.0, an elastic buffer is designed for PCIe4.0 with half full implementation. On the basis of the protocol, the clock compensation function can be well realized, and the data interface of its module and block alignment module can also be combined. Through the simulation verification of VCS, the frequency and phase compensation between the recovery clock and the local clock can be realized, and the correctness of its data receiving function can also be proved.

In recent years, with the rapid development of network information technology, problems related to information security have also emerged, such as the leakage of personal privacy information, and the monitoring or tampering of important information. These problems affect social order all the time, and even threaten national security seriously. Cipher is a common method to solve information security problems. Ballet series block cipher algorithm is a lightweight cipher algorithm using Lai-Massey and ARX structure. Aiming at the Ballet-128/128 algorithm, Verilog HDL is used to implement the algorithm in FPGA, and two design schemes are proposed. Firstly, a scheme is proposed to complete the key expansion module first, and then perform 46 rounds of encryption/decryption iterative calculation on the algorithm. This scheme adopts the method of Finite State Machine to effectively reduce resource consumption. Secondly, on this basis, the encryption/decryption operation process of the algorithm adopts a 46-stage Pipeline structure design, which can realize the encryption/decryption operation of multiple groups of data, and further improve the operation efficiency of the algorithm. Finally, in the Quartus II 13.0.1 environment, the Cyclone III series EP3C40F780C6 chip is used for engineering implementation. The engineering implementation results of the two design methods are consistent with the standard vector, and the operating efficiency is effectively improved. In the end, the throughput rate of the Finite State Machine design of the algorithm is 0.45Gbps, and the throughput rate of the Pipeline structure design optimized on this basis can reach 24.75Gbps, which can quickly encrypt/decrypt a large amount of data, satisfying most of the encryption/decryption systems.

The Jiles-Atherton model is a key theory for researching the hysteresis loop of ferromagnetic materials. In application, its theoretical parameters are difficult to determine. This paper proposes a new method for calculating the parameter k based on its definition in the Jiles-Atherton theory and the hypothesis of pinning energy. The influence of the maximum magnetization intensity (Mm) on the calculation results of the theoretical parameter k value is discussed by using the data reported in the literature, The value of (Mm) is determined, Then, the reliability of the calculated k value is verified by the measured initial magnetization curve.

In recent years, as people pay more and more attention to the protection of cloud data, reversible data hiding in encrypted image (RDHEI) has become the focus of research. This paper proposes a hiding scheme based on pixel prediction and group marking. First, the cover image is pre-processed, including block, grouping, pixel prediction, etc. Then, the embedding method and position are selected according to the parameters, and the marked map is generated. Finally, we encrypt the image and embed data in the encrypted pixels according to the marker map to obtain an image for transmission. For the receiver, separable data recovery is performed based on the key. The correct rate of secret data extraction is 100%, and the cover image can be recovered without distortion. This method does not require auxiliary information during decryption, but the decryption quality is good. At the same time, the method provides two embedding methods to provide a high embedding rate. Therefore, this method has advantages in image decryption quality, embedding rate, and separability.

At present, the network connection scale in the communication and transmission system is expanding with the increasing traffic, it is not easy to process the data, which is inefficient and difficult to ensure the quality of service. In order to increase the system transmission capacity and make the processing more efficient, a lossless information embedding algorithm based on data reconstruction is proposed, according to the characteristics of large variance and discrete distribution of different image, text, and voice data in the communication transmission system. Combined with the data characteristics of each group, interval mapping is carried out for segmented data, which is reorganized by certain rules to increase the redundancy of this segment and expand the data embedding range. The proposed algorithm realizes the lossless embedding of information and the original carrier data and embedded information are recovered losslessly. Simulation results show that the algorithm expands the redundancy range of segmented data and increases the capacity of information embedding, which is suitable for different data transmission systems.