This PDF file contains the front matter associated with SPIE Proceedings Volume 12258 including the Title Page, Copyright information, and Table of Contents.

The multiple-sets splitting feasibility problem (MSSFP) is the extension of splitting feasibility problem. This problem has been widely used in visual neural network and fuzzy image processing system. In this article, we offer the regularization methods of the inertial relaxed CQ algorithm, where the choice of parameters is not related to the operator norm. Under certain conditions, the weak convergence of the sequences is proved. In addition, the convergence of the algorithm is verified by numerical experiments.

Porosity, as one of the important reservoir physical parameters, plays an important role in reservoir evaluation. Considering the actual work needs, finding a low-cost and efficient method to obtain high-precision porosity has become an important topic of reservoir evaluation. Due to the complex nonlinear mapping relationship and timing characteristics between logging parameters and porosity, a model of deep learning is proposed to predict the porosity of carbonate reservoir according to the existing logging data. Firstly, on the basis of core analysis and geological and logging data, data preprocessing is carried out for carbonate reservoir logging data, including core depth homing, logging data standardization and normalization. The second step is to establish the prediction model of reservoir parameters by using proper learning samples. The third step is to evaluate the predicted effect of porosity model and modify the model by using superposition diagram method and error statistics method. The calculation demerit of the final model is compared with the traditional results.The comparison results in the last step show that the prediction results of reservoir parameters by neural network are more accurate than those by traditional methods.

Mispronunciation Detection and Diagnosis (MDD) is one of the key components of the Computer Assisted Pronunciation Training (CAPT) system. The construction of the current mainstream MDD system is an automatic speech recognition (ASR) system based on DNN-HMM, on which a large amount of labeled data is required for training. In this paper, the self-supervised pre-training model wav2vec2.0 is applied to the MDD task. Self-supervised pre-training uses a large amount of unlabeled data to learn common features, and only a small amount of labeled data is required for training in subsequent applications. In order to utilize the prior text information, the audio features are combined with the text features through the attention mechanism, and the information of both is used in the decoding process. The experiment is conducted on the publicly available L2-Aritic and TIMIT datasets, yielding satisfactory results.

Many real-world networks are suitable to be modeled as heterogeneous graphs, which are made up of many sorts of nodes and links. When the heterogeneous map is a non-attribute graph or some features on the graph are missing, it will lead to poor performance of the previous models. In this paper, we hold that useful position features can be generated through the guidance of topological information on the graph and present a generic framework for Heterogeneous Graph Neural Networks(HGNNs), termed Position Encoding(PE). First of all, PE leverages existing node embedding methods to obtain the implicit semantics on a graph and generate low-dimensional node embedding. Secondly, for each task-related target node, PE generates corresponding sampling subgraphs, in which we use node embedding to calculate the relative positions and encode the positions into position features that can be used directly or as an additional feature. Then the set of subgraphs with position features can be easily combined with the desired Graph Neural Networks (GNNs) or HGNNs to learn the representation of target nodes. We evaluated our method on graph classification tasks over three commonly used heterogeneous graph datasets with two processing ways, and experimental results show the superiority of PE over baselines.

We introduce the AT-GCN (Adaptive Threshold filtering Graph Convolutional Neural network model). AT-GCN is a recommendation model based on graph structure. Compared with the commonly used graph structure recommendation model, AT-GCN can effectively solve the problem of edge representation and information transfer, and improve the recommendation effect. In the experimental part, several groups of experiments were carried out on AT-GCN, and the above conclusions were finally verified by the experimental results.

To avoid data overload, recommendation systems have been created. Due to the difficulty of data collection, the recommendation system faces a cold start and needs to introduce auxiliary information. In this paper, we use social recommendation to solve the cold start, and we adopt a graph convolutional neural network to aggregate high-order neighbors and sample the neighbors for the auxiliary recommendation system. Ultimately our model achieves impressive results on classical datasets. Compared to the baselines, we achieved a higher accuracy rate.

Cellular-based Internet of Things (IoT) network has huge potential, i.e., enhancing security, promoting smart cities, and so on, while the interference will become very serious with the increasing densification of cells. Reconfigurable intelligent surface (RIS) has tremendous potential to alleviate interference. In this paper, we use RIS to improve the system performance via reducing the interference from neighboring cells and users in the same cell. We establish an optimization problem to maximize data rate with the constraint of the power of base stations (BSs). Due to the nonconvexity of the optimization problem, it’s difficult to obtain the optimal passive beamforming and active beamforming directly. Thus, we propose the fractional programming (FP) method to approximate non-convex objective function to tractable forms. Then, we utilize an improved block coordinate descent (BCD) algorithm to find suitable passive and active beamforming for the presented two convex functions. Simulation results demonstrate that the BCD method has better system performance than the non-RIS and random phase schemes.

With the gradual improvement of space launch sites image communication system construction, Visual command is increasingly demanding video image retrieval. The previous keyword retrieval method can not meet the requirements of new generation space mission network communication, content-based retrieval need to be developed. In view of the problem that video information is unstructured and cannot be quickly previewed, this paper studies the video key frame extraction algorithm and propose a video key frame extraction algorithm based on convolutional neural network.

Short videos on the Internet are growing exponentially, and the number of videos uploaded every day is huge; people also involve a lot of video data in real life. People can retrieve and view all kinds of videos, but it also brings a lot of problems. On the one hand, the accumulation of a large number of videos makes people unable to find the videos they want quickly, and the repeated scenes in the videos will also waste people's time and energy; on the other hand, a large amount of video data also brings enormous pressure to storage. Aiming at the problems of inaccurate selection of key frames and how to select video frame features in existing video summarization models, this paper proposes a multi-feature-based video summarization generation model (DME-VSNet), which extracts multiple features of video frames. Including importance score, image memory strength and image entropy. Aiming at the problem of inaccurate video shot segmentation, this model proposes a video shot segmentation algorithm based on TransNet network, which divides the original video into several short shots through shot boundaries; the model inputs the above three features into the proposed The video frame score is obtained in the MLP architecture, and the key frame is selected by the score to generate a video summary. The effectiveness of the video shot segmentation method based on TransNet network and the overall model based on convolutional neural network is verified by comparative experiments. The experimental results show that the evaluation results of the video summaries generated by the three features are better.

In this paper, we propose the recommendation algorithm PNCF for neural networks. We designed a pre-training task for a distributed representation of embeddings based on many-to-many information. We used the word2vec technique in natural language processing to implement the embedding of users and items. We also constructed a brand-new video website tagauthor pre-training dataset. The code in this paper was implemented in PyTorch and is publicly available on GitHub (github.com/jannchie/ PNCF).

Text style transfer is one of the controllable text generation tasks, which can convert text style attributes. The mainstream method is to separate the content and style of the text, and then combine the content vectors with other style vectors to generate. However, implicit expression cannot completely separate meaning and style, separation and recombination may also lead to a decrease in the naturalness and fluency of the text. Therefore, we propose a new idea, which first encodes the text into a latent representation, and iteratively optimizes the latent representation with the smallest changes to achieve style transfer. Introducing noise enhancement before generating results improves the robustness of the generated system and reduces the occurrence of individual results with large errors. Experiments show that our optimization method based on noise enhancement performs well on two public datasets, Yelp and Amazon. Result has excellent performance in three indicators: content preservation, transfer strength, and fluency.

In the post epidemic era and the rapid development of science and technology finance, bank credit card marketing has been greatly impacted. This paper proposes a new deep learning model DeepAFM (Deep Attentional Factorization Machine), which is used to predict potential credit card users of bank, so as to provide an effective basis for bank precision marketing. The model uses factorization machine and embedding layer to decompose the parameter matrix into low dimensional parameter matrix; The Attentional Mechanism is introduced to learn the weight of cross features and extract important features; A fully connected depth network is introduced to realize the mining of higher-order cross features. Finally, through the comparison with other algorithms, the results show that the expression ability of DeepAFM model is better and the automatic mining of important data is more accurate.

Experience-based book recommendation systems suffer from blurred boundaries and unguaranteed quality. In China, there is no reasonable and credible book recommendation system for elementary school students based on big data and expert systems. The basis of book recommendation is text classification. In this paper, we first construct a hierarchy dictionary to encode books according to the textbook read by pupils in different grades. We then classify extra-curricular books into lower grade, upper grade, and non-elementary classes. Four machine learning methods and a deep learning algorithm are used to classify the text. The model metrics are evaluated by Accuracy, Recall, F1-score. Among them, both deep learning and machine learning have a good performance on binary classification tasks, with the accuracy of the LR method reaching 98.5%. This reflects the value of our construction of a hierarchical vocabulary. In addition, accuracy for the triple classification task generally achieves 75%.

With the vigorous development of the Internet, the number of commonly used software has also increased rapidly. The security and reliability of software have become important challenges that researchers must deal with. Fuzzing is a way of detecting vulnerabilities by providing unintended inputs to the target software and observing the final running results. During these years, fuzzing has proven its effectiveness in software security testing. But a large number of fuzzing tools rely solely on runtime information while testing software. Achieving static vulnerability prediction for programs in advance can greatly improve the efficiency of fuzzing. Vulnerability prediction aims to obtain the possibility of vulnerabilities in different parts of the program. The existing vulnerability prediction methods are relatively simple for feature extraction between basic blocks. We design a novel model combining self-attention mechanism and convolutional neural networks, which can extract and integrate the internal information of functions. Compared with the state-of-the-art V-Fuzz, our recall can be improved by 9.7 percentage points, and the accuracies of K-100~K-1000 can be higher than 90%.

In recent years, deep neural network has been widely used in image recognition, natural language processing, computer vision and other fields, but it is prone to overfitting during network training. To solve this problem, this paper uses TensorFlow2.0 framework to construct multilayer perceptron deep network for Fashion-MNIST dataset, and uses dropout algorithm to solve the overfitting problem in the process of network training. The research results show that the dropout algorithm is applied to the deep neural network, which can make the deep neural network model have strong generalization ability and can effectively solve the overfitting problem of the training network. The research on overfitting problem has important practical significance for reducing the identification error of deep network.

Software quality prediction technology is the main method of early prediction and control of software quality. Generalized regression neural network (GRNN) can better map the nonlinear relationship between software metrics and software quality elements, but the prediction accuracy of the software quality prediction model based on GRNN is low. To improve the accuracy of the quality prediction model, we use the improved cuckoo search (CS) algorithm to optimize the smoothing factor of GRNN, solve the problems of insufficient population diversity and slow convergence speed in the later stage of the cuckoo algorithm, and propose a software quality prediction model based on the improved CS algorithm to optimize GRNN by introducing Gaussian disturbance function, to improve the accuracy of predicting the number of software defects. Finally, the paper uses the public promise data set for simulation experiments and verifies the model by comparing it with the GRNN model optimized by the CS algorithm and the standard GRNN model.

With the development of big data, artificial intelligence and other technologies, data-driven aviation equipment fault diagnosis and prediction technology has gradually become a research hotspot in the aviation field. Many typical intelligent algorithm models have been applied to this field. However, limited by the airborne embedded computing environment, there are still some problems in the deployment of intelligent prediction models represented by deep neural networks on aircraft. This paper summarizes and analyzes the research and application of typical deep neural networks such as convolutional neural networks in the field of aircraft fault diagnosis and prediction. Facing the airborne embedded environment, the current difficulties in deploying the deep neural network algorithm model in the airborne environment are analyzed. The development direction of the application of fault prediction and diagnosis algorithms represented by neural networks in the future is discussed.

Aiming at the high cost of existing IT operation and maintenance services, an intelligent scheduling system based on task duration was designed. By studying the separable task scheduling in a homogeneous system environment, for a homogeneous star network, comparative analysis of the relationship between the busy state to the idle state, the continuous transmission of transmission tasks, the release time of mixed timing constraints and the time of receiving tasks, is used to This builds a separable task scheduling model. The genetic algorithm is used to optimally solve the scheduling model. Considering the existence and universality of heterogeneous platforms in the actual parallel and distributed system environment, the task scheduling situation of the processor in the heterogeneous star network environment is researched and analyzed. . According to the mixed timing constraints, a task scheduling optimization model in a heterogeneous system environment is constructed, and the genetic algorithm is improved to solve the problem. Due to the enlargement of the problem search space, the local search operator is used to converge the algorithm. The results show that compared with several existing scheduling algorithms, the algorithm proposed in this paper can obtain a better scheduling scheme, so the designed intelligent scheduling system is practical.

Aiming at the security problems of existing windows, such as low intelligence and inability to monitor and control windows in real time, this paper designs an intelligent window system based on OneNET cloud platform. In this system, STM32 embedded single chip microcomputer is used as the main control unit, and the data acquisition module is used to upload the detected environmental data such as temperature, humidity, whether it is raining or not, whether someone is standing outside the window, etc. to OneNET cloud platform through ESP8266 Wi-Fi module, so that users can monitor the status of the window in real time on the intelligent terminal device, and at the same time, the window status can be adjusted in real time by issuing operation commands through software. The system has the advantages of low cost, simple and convenient design, fast running speed after testing, real-time monitoring of window status, anti-theft, child safety protection and intelligent control of windows.

Machine learning, as the core of artificial intelligence technology, has been rapidly developed in recent years, and has made breakthrough progress in many fields. Similarly, machine learning has been widely used in the field of economic management. Unlike other fields, data in the economic field is often complex and disordered. This complexity and disorder limit the use of some machine learning methods, but it gives neural network a huge space to play. The largest advantage of neural network is that there is no requirement on the structure of the input data. However, previous work has applied neural networks directly, without making specific improvements based on the structure in economics. In the actual economic forecast and decision-making, although there are many influencing factors, the weight of each factor is not the same. Previous neural networks put all the data into the network and then got a result without considering the different weights of each factor. We propose a new neural network with different weights forecasting and local connections, which can apply different weights to each factor to get more accurate and practical results. We use our proposed method to forecast the sales volume of Haier company, and the results show that our method is significantly better than the previous method.

Steel beam is a kind of basic component widely used in machinery and civil engineering industry and its application has been widely studied home and abroad. In this paper, the neural network toolbox in MATLAB software was used to predict and analyze damage identification based on the changes of yield strength, elongation and tensile strength of steel beams with different thickness in accelerated corrosion experiments. The results show that, on the premise of selecting appropriate training samples, the BP neural network method had a great effect on the damage identification of steel beams, and its average error was about 3%, which could meet the requirements of the damage identification of steel beams in adverse environment.

The quality of facility layout is related to the operational efficiency of the entire system. Based on bibliometric methods, 859 related papers on the theme of facility layout included in the Web of Science core dataset from 2002 to 2021 were analyzed. First, through the visualization tools of Web of Science, the issue of facility layout is analyzed from the annual publication volume, journals and research directions. Next, based on the Citespace software, the cooperation and citation relationships among countries, institutions, and authors are analyzed from the macro, meso, and micro levels, and the cooccurrence analysis of keywords and nominal terms is carried out to generate a visual knowledge map, and to solve the problem of facility layout. The research status, hotspots and trends are reviewed. The results show that the problem of facility layout has been a hot research area. However, there are many and scattered situations in China's facility layout research, and it is necessary to strengthen cooperation. In addition, there are few studies in the fields of digital facility layout, smart factory, and digital twin, which need to be focused on.

In order to obtain more accurate online information of boiler temperature field and achieve the purpose of real-time measurement and monitoring of flame temperature field distribution in the furnace, an algebraic reconstruction-neural network algorithm (ART-NN) based on optical tomography measurement was proposed. The algorithm combines the advantages of Algebraic Reconstruction Algorithm (ART) and BP neural network. Using this algorithm in the case of adding random errors, a variety of classical temperature fields are numerically simulated. The results show that the stability and reconstruction results of the ART-NN algorithm are better than those of traditional algorithms such as ART and TSVD under the same error level. Optical tomography temperature field measurement provides an efficient method.

GPS technology has the characteristics of high precision, high sampling, real-time, and simultaneous determination of three-dimensional coordinates of points, which can not be compared with other monitoring technologies. It plays a very important role in deformation monitoring. Starting from the composition of GPS positioning system, this paper expounds the three components of GPS positioning system, as well as the association and coordination between each component of the work; then the GPS deformation monitoring mode and several error sources in the monitoring process are introduced. The advantages and disadvantages of GPS technique in deformation monitoring are analyzed and its application trend is predicted. GPS positioning technology is applied in all aspects of our life, creating a lot of social and economic value for us.

Target detection technology is a typical application in the military field. It can quickly and accurately find and identify all kinds of enemy vehicle targets in the battlefield, and respond to all kinds of battlefield targets more quickly, which has become the key to improve the battlefield situation. Because the battlefield environment is very complex, the traditional target detection algorithm is not ideal when detecting targets in complex scenes. Therefore, a target detection algorithm for armored vehicles of military tanks based on improved YOLOv3-tiny is proposed, which realizes the automatic detection of military targets in complex environments by deep learning. Firstly, based on YOLOv3-tiny algorithm, ResNext residual network is added to replace the original feature extraction network, which better improves the problem of missing and false detection of small targets and optimizes the convolution network structure. Then, the dense network is introduced, and the features of different layers are fused to realize feature reuse, which improves the efficiency of extracting better features of target vehicles, strengthens the network's ability to learn features, and improves the detection effect. Experimental results show that the recall rate and precision rate are increased by 4.62% and 3.79% respectively, the average precision rate is increased by 4.32%, and the frame rate can reach 78 frames/s.

Gait recognition is widely used due to its advantages of long-distance recognition, unnecessity of active participation of the recognized person, etc. In recent years, many gait recognition models based on deep neural networks have achieved relatively high accuracy. However, many studies have shown that deep neural networks are vulnerable to adversarial attacks, and the recognition of deep neural networks can be made wrong by the addition of small perturbance to the input samples. Therefore, it is very important to explore the robustness of neural networks for gait recognition. Since the structure and parameters of the gait recognition model are often difficult to obtain in practical applications, a half-white-box adversarial attack method based on GAN was proposed in thesis. The difference between the adversarial examples generated by using this method and the original examples is difficult to be distinguished by the naked eye. Through experiments, it is found that the input of adversarial examples has a large impact on the output of the gait recognition model. To make sure that the generated adversarial perturbance can be easily implemented in the physical environment, we changed the model structure and increased the input data based on the above method, and proposed a new method. In this method, the generator of GAN can generate adversarial perturbance in specific shapes. The experimental results show that even without knowing the network structure and parameters of the target model, the accuracy of the gait recognition model in the face of adversarial examples will decrease, indicating that there are problems with the robustness of even very advanced gait recognition models.

Nowadays, Convolutional Neural Network (CNN) based image recognition is a popular research direction. This study uses the Fashion-Mnist dataset, which is more challenging than the Mnist dataset. aims to add Long short-term memory (LSTM) to the structure of CNN to create a hybrid model of CNN and LSTM, called CNN+LSTM model. This model is used to complete and optimize the image classification problem on Fashion-Mnist dataset. The final image classification accuracy of the obtained model is 91.36%, which still needs to be improved, but the accuracy results are better compared to the accuracy of other models.

The end-to-end speaker recognition model has made a great breakthrough in the research field and practical application scenarios. However, in practical application, we often suffer the diversity of noise, and the interference of noise will affect the performance of speaker recognition and classification model, and the model usually degrades in unseen scenes with noise. In this paper, a speaker recognition model combined with front-end enhancement is proposed. The front-end enhancement model (DTLN or CRNN) is combined with the back-end speaker recognition model (Res2Net-GhostVLAD) to improve the robustness of the model against noisy scenes, and the generalization capability of the model is increased by the data augmentation method (SpecAugment). Our proposed method was trained on VoxCeleb and AISHELL datasets and tested on VoxCeleb datasets. The test results show that the proposed method significantly improves the performance of the speaker recognition model in noisy scenarios, and the relative improvement of different front-end models is 9% and 13%, respectively.

This paper proposes a new model-based gait recognition method. Different from other methods using 3D (3-dimensional) keypoint information and skeleton information, we directly stack the 2D (2-dimensional) keypoint heatmaps in the gait sequence in the time dimension, and input it into the network structure based on 3D-CNN (3-dimensional-convolutional neural network). Then, through the gait analysis on the two dimensions of time and space, the effective gait features are finally obtained. Compared with other model-based methods, this method is more clear, concise and elegant in the process of feature extraction. The test of CASIA-B dataset shows that in the model-based gait recognition method, we have competitive performance.

Aiming at the TCP Incast problem in Software Defined Networks (SDN) data center networks, a new congestion control algorithm based on deep reinforcement learning is proposed in this paper. By the way of continuously interacting between the agent and the environment to obtain the optimal strategy, the SDN controller plays the role as agent and many-to-one model of data center networks plays the role as environment. Under the condition that queue length of the buffer of the bottleneck switch does not exceed the preset congestion threshold, data transmission rates of multiple servers are trained using deep reinforcement learning algorithms, and finally the optimal transmission rates which will not cause congestion are acquired. The simulation results show that the algorithm proposed in this paper can effectively avoid congestion in many-to-one scenarios in SDN data center networks and can improve the overall performance of networks.

The Internet of Things devices has rapidly increased and been widely used in recent years. The era of the Internet of Everything is quietly coming, which puts forward higher requirements for the research on network traffic classification in the Internet of Things environment. However, traffic in the network layer and link layer is often ignored. This paper proposes a network traffic classification and feature extraction tool that covers multiple layers of network protocols to convert the original network traffic into digital features. With the features, two deep neural network models constructed were trained, and evaluation of their multiple indicators proved the effectiveness and superiority of our proposed intrusion detection system for IoT. It can achieve a classification accuracy of 98% and 97% of detection rate.

In this paper, the main purpose is to study how to reduce PAPR in the OFDM-IM system. According to some methods used in the OFDM system, have a further exploration analysis in OFDM-IM system. Firstly, it is about the OFDM-IM system introduction and the effect of PAPR on the OFDM-IM system. Secondly, SLM and PTS are studied in many aspects in the direction of OFDM, like the choice of phase and subcarrier grouping patterns. When phase factor is {1, -1}, the method used in OFDM is borrowed in PTS to improve the calculation time of results in the OFDM-IM system. In addition, the comparative analysis of PAPR under the combination of SLM-PTS and SLM-PTS methods is also studied. It can be concluded that the new method combines the characteristics of both, and its running time is less than that of traditional PTS and the performance of PAPR is less than that of traditional single SLM.

Timing attack is a side channel attack method. Elliptic curve cryptography (ECC) is one of the most important publickey cryptography. In this paper, a new timing attack on the Elliptic Curve Digital Signature Algorithm (ECDSA) based on Hidden Markov Model (HMM) was presented. Precisely speaking, the Grover algorithm was used to retrieve the parts of the ephemeral key, and the Koblitz Curve K-409 which was recommended by NIST was attacked successfully. The experiment results showed that the attack could recover almost all the key bits in a few minutes by collecting only once timing dates, and was easy to experiment at a high success rate.

MOOC platform is one of the most important data sources of educational big data, so the correlation analysis of MOOC learning behavior data has become a research hotspot in educational data mining and learning analysis. The purpose of this paper is to study the MOOC recommendation algorithm based on the learning process sequence modeling and quantitative analysis. Aiming at the problem of frustration caused by dropping classes in MOOC, this study improves the recommendation feature model, and proposes an adaptive process recommendation method. Based on the data modeling of MOOC learning process and quantifying the learning status, it realizes multi-feature adaptive trade-off recommendation, and achieves Reduce the purpose of dropping out. First, the traditional recommendation characterized by interest is improved, and a new feature model is adopted to reflect the learner's satisfaction needs and reduce frustration. Secondly, the influence of various similarity distances such as time distance and knowledge distance on learning features is considered to improve the recommendation accuracy. Finally, the recommendation results are evaluated. The experimental results show that when k1 is 10, the recall of MRSS reaches 0.42, and the accuracy rate is the best.

This paper builds an OFDM-IM simulation system, analyzes the system bit error rate (BER) and the system signal-to-noise ratio (SNR) changes under different activated sub-carrier values. The conclusion is that when only BER is considered, activating a small number of sub-carriers will get better results. Then it analyzes the change rule of the optimal solution number of activated sub-carriers with the total number of sub-carriers, and obtains a general mathematical formula description, and combines relevant data such as the proportion of the optimal number of activated sub-carriers, and the theoretical transmission speed optimization rate, Compared with the OFDM system, the conclusion is that when the subcarriers carry more information, the addition of the IM system does not significantly improve the performance of the system, but increases the complexity of calculation and coding. In the above case, the optimal number of activated subcarriers is almost equivalent to the total number of sub-carriers, so adding IM modulation at this time does not make much sense.

For the purpose of better measuring the influence of musicians and exploring the similarities and characteristics of various genres, we need to optimize the Musician Influence Model properly and represent musical features in an appropriate way. The previous model was established based on the correction of followers’ and influencers’ number of each musician and the correction of total number of actual musicians according to genre proportion, which did not take the change of genre proportion over time and indirect influence of musician into consideration. To better optimize it, I innovatively introduced “direct number”, “indirect number” and “strong number” of followers and influencers, and all of them was corrected by proper functions. Through the standardization of data and the frequent use of entropy weight method, the relationship between the data became more accurate and reliable. The advantages a musician should have to be more influential can be assumed after obtaining the ranking table with Bob Dylan, The Beatles, Radiohead, Avril Lavigne and Alan Jackson being the top five. Besides, I selected six parameters - danceability, energy, valence, Tempo, acousticness and instrumentalness, to characterize the music. And I selected 6 representative genres based on the analysis of the proportion, musician number and average popularity of all genres. We draw some conclusions by analyzing the musical characteristics radar map of each genre.

Our country is a developing country dominated by agriculture, and the trade of agricultural and sideline products accounts for a considerable proportion in China's national economy. Soil moisture content is one of the basic parameters reflecting soil fertility and crop growth in agricultural production activities. The detection of soil moisture content plays an extremely important role in improving soil utilization and increasing crop yield. At the same time, thermal infrared technology has the characteristics of wide acquisition range, high acquisition accuracy, fast information transmission speed and less affected by interference factors. It provides a new method for in-situ detection of soil moisture. In this paper, infrared thermal imaging technology is used to detect the soil in the measured area in situ, which not only avoids the influence of traditional drying method on the sample in the sampling process, but also improves the detection accuracy. The combination of thermal imaging technology and thermal inertia model makes it suitable for the detection of soil moisture content under most environmental conditions, including bare soil, sandy land and low vegetation coverage.

To improve the capability of mining the deep features of the target in the recognition process of the high-resolution distance image (HRRP) of the ship targets, based on multiple perceptron (MLP), the R-MLPs model is proposed. The model is constructed by recurrent neural network (RNN) and MLP, and the signal is encoded by RNN using the characteristics of the ship targets HRRP time-series signal. The encoded signal is further processed using an MLP with a spatial gate mechanism to extract more robust features. To verify the superiority of the R-MLPs model, we also use RNN, LSTM and gMLP models for experimental comparison under different signal-to-clutter ratio conditions and low number of network layers. The experimental results show that R-MLPs can identify targets effectively and accurately, and have better robustness at low signal-to-clutter ratios and fewer layers.

Aiming at the characteristics of small targets, many interferents and inconspicuous features of spore images of wheat powdery mildew, a weight adaptive feature fusion model is proposed based on SSD network structure to improve the accuracy of spore detection. Firstly, a feature fusion path is constructed to recursively fuse features of various scales from deep to shallow, and at the same time, a layer of feature matrix is added to enhance the utilization of deep and shallow features by the network; Secondly, a hybrid attention module is proposed, which redistributes the weights of features adaptively to enhance the ability of extracting network context information. Finally, the k-means algorithm is used to set the shape of the prior box, which effectively improves the problem that it is difficult to manually adjust the hyperparameter of the neural network. The AP of powdery mildew spores was 91.17%, Compared with the classical SSD detection method, it has been greatly improved.

One of the urgent problems to be solved in UAV cluster warfare is to find a suitable routing algorithm to adapt to the characteristics of UAV Ad Hoc network topology. Based on the further development of cellular network, the wireless Ad Hoc network does not need preset infrastructure, no control center, and can be temporary networked, which is more in line with the needs of unmanned aerial vehicle Ad Hoc network. In consideration of the premise of distributed deployment of unmanned aerial vehicle (UAV), matlab is used for simulation of opportunistic routing algorithm. As for multiple constraints Qos routing problem, tabu search algorithm is adopted to optimize the opportunistic routing algorithm. The experimental results show that, after intelligent algorithm is optimized, along with the increase of the time consumed for algorithm, other constraints impose effective impact on the algorithm, enabling the opportunistic routing algorithm make the optimal choice according to multiple constraints. Then genetic algorithm is used for comparison, and the conclusion that genetic algorithm can get faster convergence is drawn.

To effectively help second language (L2) Chinese learners to produce tones correctly in computer assisted language learning (CALL), tone recognition of continuous speech is necessary. Because of the complex tone variation in continuous speech, this paper proposed TAM-BLSTM tone recognition model. Firstly, the generation model, target approximation model (TAM) is used to simulate fundamental frequency (f0) from original f0 contour in the unit of prosodic words, and the TAM parameters for each Chinese character are derived. Then BLSTM model with attention mechanism is set up with input feature of the TAM parameters and basic acoustic features, such as statistical f0 parameters, vowel duration, to solve the problem of tone detection of Mandarin continuous speech. Finally, the trained tone detection model is applied to the tone error detection of the L2 learners. The experimental results with Biaobei corpus show that the accuracy of the feature set combined with TAM parameters is 2.3% higher than that of using basic acoustic features alone, and the overall accuracy of ATT-BLSTM network model is higher than that based on ATT-LSTM.

In recent years, how to identify the most influential nodes has become the forefront of network science. Considering the influence of community structure and neighbor nodes within the second order on node influence diffusion, this paper proposes an influence maximization algorithm based on community structure (IMCS). Firstly, the CPM algorithm is used to detect the community of the network to obtain the community structure of network. Then, select the nodes belonging to multiple communities in the community and some potential nodes in each community to form a candidate node set. Finally, use the improved Prob-Degree algorithm to screen all seed nodes. Experimental data show that compared with Prob-degree, CoFIM and DD, the algorithm proposed in this paper has relatively good overall performance in Oregon network, and there are seed node intervals with relatively good performance in other networks too.

Aiming at the problem that traditional ViBe algorithm is prone to appear ghost area in the process of moving target detection, an improved ViBe algorithm was proposed. In the background model initialization stage, a time background model is added for each pixel, which uses the temporal characteristics of the pixel; at the same time, the original background model is improved, and the image used for modeling is determined by comparing the Hamming distance between frames. The coexistence of dual-background models is realized. In the foreground detection stage, the decision mechanism of the foreground detection is improved, that is, the Euclidean distance between the current pixel and the median of the temporal background model sample is used to supplement the decision; Finally, a cross-updating strategy is designed to rapidly update the dual-background model in the background model updating stage. Experiments show that compared with the traditional ViBe algorithm, the algorithm in this paper has a better effect on ghost suppression.

Improving the visual recognition and positioning accuracy in the outdoor environment is an important way to improve the picking efficiency of fruit picking robots. With the rapid development of artificial intelligence, the convolutional neural network algorithm has gradually become an important research direction for machine recognition and localization. It can automatically extract target features, with high recognition accuracy, high speed and strong robustness. This paper takes pears as the research object, and proposes an improved pear recognition and localization algorithm based on the yolov5 model. The generalization ability of the model is improved by preprocessing and data enhancement of the data set, and an improved k-means clustering algorithm is proposed to realize the optimal calculation of the initial anchor frame. Compared with the original yolov5 model, the fitness and best recall rate of the improved algorithm in recognizing pears are increased by 6% and 9%, respectively.

Packing technology is commonly used in malicious software. With the increasing awareness of software publishers on their own intellectual property protection, the phenomenon of packing benign software is becoming more and more common. This phenomenon leads to a high false positive rate in traditional machine learning-based malware identification results. Traditional researches on malware detection based on machine learning focus on improving the identification accuracy of malware, and there are few researches on reducing the false positive rate. This article focuses on this issue. We select the data set that labels whether benign software is packed or not, and use a variety of machine learning algorithms to conduct experiments. Finally, we obtain the method with the lowest false positive rate. The experimental results show that the comprehensive index of the Extra-Trees algorithm is optimal.

The elevator dispatching system is an organic part of the elevator transportation, which can guarantee the smooth, safe and effective operation of the elevator. China's traditional elevator scheduling system is mainly through the button to call the elevator. The traditional elevator group control method based on the minimum waiting time can not meet the overall demand of the elevator due to its single evaluation index, which leads to a series of problems such as low efficiency of elevator group control and poor joint mobility. In view of this situation, an intelligent elevator scheduling system based on image recognition and voice recognition is designed in this paper. Image recognition, voice recognition and Internet of Things are combined to increase the communication between elevator and passengers, as to improve the efficiency and comfort of passengers and reduce the energy consumption of elevator.

At present, there are few researches on the detection of steel coil end face defects, and the existing detection methods cannot take into account the accuracy and speed of the problem. An improved Faster-RCNN and improved YOLOv3 detection algorithms are proposed. Since Faster-RCNN loses the underlying information during continuous sampling, it is difficult to detect small defects on the end face of the steel coil. It is proposed to use multi-scale feature fusion to improve the optimization. At the same time, due to the poor performance of the Faster-RCNN backbone network, the detection rate is slow. It is proposed to use A better-performing residual network alternative. In view of the inaccurate positioning of YOLOv3, it is proposed to use GIoU to replace the IoU in the original YOLOv3. Likewise, to further improve the YOLOv3 detection rate, a residual network is used instead of Darknet-53. Experimental analysis shows that the detection rate of the improved Faster-RCNN_R50 reaches 12.47Fps, the original Faster-RCNN is 8.35Fps, an increase of 4.12Fps, the detection accuracy (AP) of Faster-RCNN_R50 reaches 92.1%, and the original Faster-RCNN is 81.2 %, an increase of 10.9%; the improved YOLOv3_R50 detection rate is as high as 22.87Fps, the original YOLOv3 is 19.07Fps, an increase of 3.80Fps, the YOLOv3_R50 detection accuracy rate reaches 89.5%, the original YOLOv3 detection accuracy rate is 68.7%, an increase of 20.8% . The conclusion shows that in the absence of real-time detection requirements, the improved Faster- RCNN_R50 has the highest accuracy, and if there is a real-time detection requirement, the improved YOLOv3_R50 has the fastest speed, while also ensuring a certain detection accuracy.

When using UV-Vis spectroscopy to detect water quality parameters, the scattering of suspended matter in water will cause the overall spectral curve to rise nonlinearly, which will affect the accuracy of the experimental results. Aiming at the problem that the spectrum is easily interfered by the light scattering of suspended matter, a total light scattering compensation method based on Mie scattering theory is studied to compensate for the interference of spectral turbidity. The extinction spectrum of , and then differentiated from the original spectrum, to achieve accurate compensation for turbidity interference, and this method does not require prior data support, which can improve the detection accuracy of organic matter content by UV-Vis spectroscopy.

The characteristics of the ARMA model are analyzed according to the characteristics of the short-term flow data of the local area network in this paper. The time prediction model of network flow is established on the ARMA method. The prediction parameters of the ARMA model are determined and the model is simulated According to the short-term flow prediction. The comparison between the simulation results and the measured data of NetFlow shows that the model can accurately predict the short-term flow behavior trend of the local area network, which can provide reference and reference for the analysis of network traffic behavior.

Building target is an important target on the battlefield. In the process of system simulation, it is necessary to analyze the vulnerability of various typical building targets, this paper presents a damage tree modeling and vulnerability analysis method for typical building targets, including target composition and functional characteristic analysis, system structure functional damage tree analysis, target vulnerability model establishment, vulnerability characteristic analysis and so on, which can be used for damage modeling and vulnerability analysis of related typical building targets.

In the energy-saving technology of TWDM-PON, in order to avoid the delay deterioration caused by the failure to wake up the ONU receiver in time, the ONU receiver usually needs to remain open, which is bound to cause unnecessary idle energy consumption. For this reason, an energy management mechanism based on pipelined cycle polling (PCP-EMM) is proposed in this paper. Through pipelined cycle polling, ONU can obtain the authorization information of the next polling cycle in advance, so as to solve the problem that ONU cannot be awakened in time because of not knowing the arrival time of burst data in the process of downstream transmission. In addition, the thought of modularization at both ends of ONU and OLT is introduced, according to which each module will be closed or awakened adaptively through the effective management of module energy state, so as to reduce the energy consumption of the network. The simulation result shows when the network load is less than 90%, PCP-EMM can effectively reduce the energy consumption of the system while meeting the delay limitation of the service. And the highest energy-saving rate is up to 53.77%.

In order to study the spatial distribution characteristics and causes of Heifangtai landslide in Gansu Province, the sentinel- 1A images from September 2017 to November 2020 were used as the data source to extract surface subsidence information in the study area using SBAS technology, and the high coherence point D1 of the landslide in Dangchuan village was selected, the subsidence was analyzed by combining irrigation, rainfall and temperature data. And the BP neural network was used to predict the point. The results showed that: (1) the area identified by SBAS technology was mainly spread in Xinyuan village, Fangtai village, Zhuwang village, Chenjia village and around the tableland. (2) In February and March, due to the large temperature difference, the landslide of Dangchuan started to settle as the temperature increased and caused the permafrost to melt; The amount of irrigation and rainfall increases from June, when the loess tableland starts to sink and landslides occur frequently; After October, the landslide in Dangchuan Village produced a frozen stagnant water effect, and there was a tendency for the subsidence to increase. (3) The prediction result of BP neural network shows that the subsidence rate of D1 point will surpass 60 mm in 2022, which is important for the early identification and prevention of the area.

The application purpose of space rapid response launch system based on data link refers to integrate the command information system, sensor system, the launch platform, and platform-mounted command and control system, test and launch control system, measurement and control system and other elements, to realize end-to-end information interconnection and application interoperability, so as to expand the situational awareness scope, command and control efficiency and action coordination ability of a single launch unit. This paper has carried out the research on the launch efficiency improvement methods based on data link application system, distributed decision making and other key technologies, the application work modes and workflows for the typical launch mission scenario, so as to improve the effectiveness of joint command and control at the campaign level and operational coordination at the tactical level.

In network data transmission, the data to be encrypted is not always at the same level. The access structure is embedded in the ciphertext, resulting in excessive communication and computing overhead. Therefore, a secure sharing method of network data transmission is proposed based on multi-layer encryption algorithm. According to the relationship between users, type servers and service providers, the logical structure of data transmission is established, and the data sharing contract is called on the basis of secure communication. By designing the key synchronization mechanism, both sides of the transmission can use the same key for encryption and decryption to ensure the forward security of the data. The key distribution scheme is established based on multi-layer encryption algorithm. The ciphertext information of the next layer is directly embedded in the ciphertext of the previous layer, which realizes the jump transmission of ciphertext and ensures the security of data privacy. In the final stage, both the master node and the slave node will receive the confirmation information, and then check the signature of the confirmation information to complete the secure sharing of data transmission. The test results show that this method reduces the communication overhead and computing overhead, and improves the efficiency of encryption and decryption.

Deep learning methods for multi-parametric MRI hold the greatest promise for automated computer-aided diagnosis of prostate cancer, including classification and segmentation. In this work, we propose a new model (MC-DSCN) for classification and segmentation simultaneously. MC-DSCN contains three components: the coarse segmentation component based on the residual U-net with attention blocks, the classification component based on the stacked residual blocks and multi-parametric fusion mechanism, and the fine segmentation component that incorporates the information about lesion location (cancer response map, CRM) arising from the classification component. Extensive experiments are performed to demonstrate that the proposed method could effectively transfer mutual information between segmentation and classification components and facilitate each other in a bootstrapping way, thus outperforming the methods designed to perform only one task.

Since signal transmission is performed on a 220V/330V low voltage power line, there is a series of problems such as distribution transformer blocking the carrier signal, more load access leading to significant signal reduction, inherent pulse interference of power line. Therefore, the channel transmission characteristics of power network must be considered when power lines are used to transmit carrier communication signals with high frequency and low energy. On the premise of analyzing the necessity of power line carrier communication performance index evaluation, the channel input impedance characteristics, channel noise and interference characteristics, channel attenuation characteristics are studied. On this basis, the power line carrier communication equipment performance index evaluation system based on fuzzy comprehensive evaluation method is studied.

With the development of the Internet, users’ personal information has become one of the key factors for Internet platform. Due to the privacy concern, users are often reluctant to provide their personal privacy information to Internet platform. ELM model is an important model to analyze consumer behavior. Based on ELM model, this paper studies the willingness of users to provide their private information. The results shows that privacy collection method, privacy protection statement, and privacy protection technology of the website have a negative correlation with the willingness of users to provide privacy. This research is helpful for enterprises to master user information, analyze user behavior, and find the key factors affecting users’ willingness to provide private information.

In order to solve the problem of storage and management efficiency of massive urban monitoring data, HDFS distributed file system is used to store video unstructured data, HBase attribute information table is researched and designed, HBase distributed database is used to store attribute information, HBase global retrieval information row key and MapReduce workflow are studied, retrieval is constructed, and video monitoring data retrieval is realized. Finally, based on the research of unstructured data retrieval technology and distributed data processing framework, we get the massive urban monitoring data storage and retrieval scheme.

In recent years, there have been many natural disasters in China. After the disaster, communication equipment was damaged to varying degrees, which severely hindered the establishment of communication links between disaster relief departments and disaster sites, which caused great inconvenience to disaster relief tasks. The classification of picture compression technology and commonly used encoding methods was introduced, the international compression standard JPEG algorithm was selected as the picture compression algorithm in this article, the implementation steps of color mode conversion, transformation, quantization, and encoding in the JPEG algorithm was introduced, and finally the Java package was used to implement this compression algorithm. The current status of Beidou satellite navigation system was expounded. A communication program based on Beidou short message communication technology for image transmission was written. The experimental environment and experimental steps were detailed. The feasibility of the JPEG image compression algorithm was verified, which laid the foundation for the improvement of the subsequent experimental process.

The main type of communication system can be divided into base-band transmission and pass-band transmission. For baseband transmission, polar NRZ coding is utilized among common PCM coding schemes to achieve the best error performance when the SNR stays the same. Matched filter receiver is adopted to carry on error analysis. For pass-band transmission, signal-space is introduced to simplify the procedure with a few examples. Then a new Euclidean space is mentioned, which is sort of different from the former, to directly map sending bits with their actual position on the space. 2D and 3D space, related to the 4-ary system and 8-ary system respectively, is depicted to illustrate the point. At last, a potential application for pixel maps is considered to qualitatively underscore the error impact on the practical system.

In order to meet the demand for flexible acquisition of video images and real-time processing of images, the speed and efficiency of image processing are improved. This paper is an embedded system based on the ZYNQ-7000 series chip which integrates ARM+FPGA structure for image acquisition and real-time processing. An improved sharpening algorithm that can be used for edge detection in the image processing module is proposed, and the algorithm is integrated into the hardware through the HLS tool that comes with Vivado, which realizes the flexible acquisition and real-time processing of video image data. By studying image processing algorithms commonly used in embedded acquisition and processing systems, an improved sharpening algorithm is designed based on threshold segmentation and image filtering and applied in experiments. In terms of hardware, it has the advantages of low power consumption, high performance and flexibility. In terms of algorithm, it not only improves the speed of image optimization processing, but also has a large development space and application value.

Aiming at the characteristics of underwater acoustic networks such as long delay, low channel utilization and low throughput, as well as the existence of temporal-spatial uncertainties and hidden terminals problem, this paper proposes a concurrent scheduling MAC protocol. The protocol exchanges information such as ID, location, level of the receiving and sending nodes, and the scheduling time of data reception by sending handshake control frames. so as to realize the concurrent transmission of data from multiple sending nodes to one receiving node, which not only avoids data collision to a great extent, but also effectively solves the problems of temporal-spatial uncertainties, hidden terminals and and high energy consumption and other problems in Underwater Acoustic Networks, and improve the fairness of node channel access. Network performance such as channel utilization and network throughput. The simulation results shows that the network throughput of CS-MAC protocol is better than ALOHA, R-MAC, and slotted- FAMA, etc. CS-MAC protocol outperforms ALOHA and LSPB-MAC protocols in terms of packet delivery rate and end-to-end delay.

In order to realize the construction of satellite communication system simulation platform, this paper introduces the architecture of satellite mobile communication system, analyzes the wireless transmission characteristics of satellite channel, studies the typical satellite channel model, and discusses the basic design of the system simulation platform. This research not only provides the construction method of satellite communication system simulation platform, but also provides reference for satellite communication system beam management and frequency planning design.

Accurately mapping the surface rivers is important in ecological environment monitoring and disaster prevention. The development of remote sensing technology and computer vision greatly improves the efficiency of this task. However, there are few methods that map the rivers from an image directly. The existing automatic river mapping methods usually had two successive stages: waterbody extraction and flow-path extraction, where the latter methods were very dependent on the waterbody masks generated by the former methods. Errors in waterbody masks caused breaks and redundancies in the extracted graphs. This paper proposed RiverMapper, which mapped the rivers step-wisely without dividing into two stages. Following the directions and actions predicted by the convolution neural network, RiverMapper walked along the rivers step by step and cropped the fixed-size image patches at each step for segmentation. Final river graphs were constructed by the waterbody mask patches and those tracks generated by RiverMapper. We applied RiverMapper on optical remote sensing images containing the Changjiang River and the Huanghe River. Without the degradation of the performance on waterbody extraction, RiverMapper outperformed other methods in terms of the local topological and geometrical similarity between the predicted and the ground-truth river graphs.

At present, the convolutional neural network is deepening in level and has a huge amount of computation, so it is difficult to realize application in scenarios with low computing capacity. Therefore, this paper proposes a method based on channel pruning and weight quantization to reduce the amount of computation and compress the image super-resolution to reconstruct the network model IDN. Experimental results show that the proposed method effectively compresses the model structure, greatly shortens the calculation time of the model and makes the model more lightweight under the premise that the performance indexes are basically unchanged.

Aiming at the characteristics of high-speed video signals and the stability requirements in some application scenarios, a digital data transmission technology based on FPD-Link III transmission technology is proposed. There is no intrusive design for digital video signals. The design adopts the serializer chip of the model DS90UB925Q-Q1DS90UH925Q-Q 1 and the deserializer chip of the model DS90UH925Q-Q 1, which can realize the non-intrusive acquisition of the video signal. The functions of each part are introduced in detail, and the overall functional block diagram and hardware principle block diagram are given. The practical engineering application shows that the design is simple and reliable, low cost, and has high promotion value.

Ammunition damage effect is an important basis for system simulation and ammunition power analysis. To solve the problem of ammunition damage effect analysis, it is necessary to collect relevant data, model, simulate and analyze ammunition. Based on the theoretical formula and test data, this paper proposes a typical ammunition modeling, power field analysis and damage effect analysis method, and constructs the corresponding typical ammunition damage effect analysis software, which includes database, parametric modeling module, power field analysis module, three dimensional visualization module. Compared with the corresponding test data, the credibility of the method and the availability of the software are verified.

In this paper, the outage performance for a full-duplex (FD) relaying satellite-terrestrial system is investigated. The system consists of a satellite source, a relay, and two users. Firstly, the satellite transmits the information of two users to the FD relay, and then the relay decodes and forwards the information to both users. Herein, the non-orthogonal multiple access protocol is used to incorporate the signal of two users. We investigate the system outage performance in the presence of residual self-interference of the relay and derive the exact expressions of outage probability for two users. Moreover, the accuracy of the results is verified with Monte-Carlo simulations.

With the rapid development of the Internet of Things, the problem of channel noise restriction has become increasingly prominent. In the application process, the SNR is too low. In order to solve the above problems, an adaptive digital filtering method of noise in Internet of Things communication channel is designed. Extract the network spread spectrum factor, extend the original signal to a higher bandwidth transmission, improve the communication channel noise detection mode, forward the received source signal to the destination and amplify, realize the design of adaptive digital filtering method. Experimental results: The signal to noise ratio of the self-adaptive digital filtering method of the Internet of Things communication channel noise in this paper is 2.903dB and 2.866dB more than the mean of the other two methods, indicating that the self-adaptive digital filtering method of the Internet of Things communication channel noise in this design is more effective.

In the wideband transmission digital beamforming technology, due to the aperture crossing phenomenon, we cannot simply replace the delay with digital phase shifting. Commonly used fractional delay methods include the windowing method, VFD filter and other methods, but the fractional delay filter often has a serious hardware resource consumption, which is not conducive to the realization of multi-channel beamforming on a single chip. In this paper, a time-delay beamforming method based on DDS parameter control is proposed, and the system structure of emission beamforming is further optimized on the basis of which to save more resource consumption.

Age of Information(AoI) is a novel metric to measure freshness of data in status update scenarios proposed by academia in recent years. Real-time applications need to transmit data packets for status update to the target node as soon as possible. However, due to the data density, the limited computing capacity of edge devices and the influence of the environment, the problems of intensive computation and high delay are caused. Mobile edge computing (MEC) is a new computing mode that extends cloud computing power closer to the user, where computing offloading and other technologies promise to solve those problems. We mainly studies the AoI optimization in MEC networks, in which data freshness and offloading strategy play an important role. Firstly, we propose the average AoI minimization problem for MEC network scenarios, and propose a multi-agent deep reinforcement learning(DRL) algorithm called Federated Multi-Agent Actor-Critic (Fed-MAAC). Federated learning is used to train agents to improve algorithm performance and data security. At the same time, we conducted experiments in gym, a popular simulation environment in reinforcement learning, and compared Fed-MAAC with baseline algorithm. The simulation results show that this algorithm is superior to other algorithms in average AoI optimization performance.

In recent years, with the rapid development of science and technology, the progress of fiber optic sensing technology is also relatively significant. In this paper, we design a circuit that can be used to amplify and detect weak signal photoelectric detection and processing by amplifying the differential input and zeroing the output, expanding the adjustable secondary, and doing a good job of power conversion, low-pass filtering and data acquisition, and then use the relevant software to carry out simulation experiments, showing that the designed circuit can effectively detect the nW-level optical signal The circuit is widely used because of its good overall linearity and anti-interference capability.

Due to its dynamic and changing network topology structure, AdHoc network can carry out wireless transmission of data, voice, video and other services in harsh conditions. Therefore, whether in the military field, or in the field of civil or commercial field, Ad Hoc network has great development prospects.The access protocol of Ad Hoc network wireless channel has always been the key and difficult point of scholars at home and abroad. System throughput, channel utilization rate and system delay all depend on the protocol used.Based on this, this paper mainly expounds the main problems facing Ad Hoc channel access, computer simulation experiment of CSMA protocol with monitoring function, computer simulation experiment of CSMA protocol with three simulation mechanism, hope that the computer simulation can be helpful through the simulation test for such researchers.

Abstract Meaning Representation (AMR) is a kind of semantic representation of natural language, which aims to represent the semantics of a sentence by a rooted, directed, and acyclic graph (DAG). Most existing AMR parsing works are designed under specific dictionary. However, these works make the content length of each node limited, and they mainly need to go through a very complicated post-processing process. In this paper, we propose a novel encoder-decoder framework for AMR parsing to address these issues, which generates a graph structure and predicts node relationships simultaneously. Specifically, we represent each node as a five-tuple form, containing token sequence of variable length and the connection relationship with other nodes. BERT model is employed as the encoder module. Our decoder module first generates a linearization representation of the graph structure, then predicts multiple elements of each node by four different attention based classifiers. We also found an effective way to improve the generalization performance of Transformer model for graph generation. By assigning different index number to nodes in each training step and remove positional encoding used in most generative models, the model can learn the relationship between nodes better. Experiments against two AMR datasets demonstrate the competitive performance of our proposed method compared with baseline methods.

Digital pulse-position modulation (PPM) is a time-domain version of frequency-shift keying that encodes each variable pulse to a particular position within a period. Although numerous essays have researched PPM applied to optical fiber, the PPM process with additive white Gaussian noise (AWGN) was not mentioned. In this paper, a relatively simple model was constructed for 2-bit message signals passing through the AWGN channel. While encoding the baseband signals, Gray Code was adopted to encode them in each period separated into four equal parts. Afterward, they all passed through the AWGN channel, and the constellation of baseband signals and received signals was performed in MATLAB. Then we applied the minimum distance identification method to find the estimated baseband signals. Finally, the PPM signal-tonoise ratio (SNR) was obtained based on previous probability errors. Executing the codes in MATLAB, we got the baseband signals, constellation, and SNR function that can all be directly shown in the figure. In general, our work filled the gap of the PPM method sending over the AWGN channel. This can also become the preliminary foundation for a broader range of applications using PPM to implement the whole communication process with AWGN.

The classical wavelet transform in image processing is easy to lose the details, and the edge of the image is often easily disturbed by the noise which leads to poor extraction effect. In order to overcome the shortcomings of the classical image edge detection methods, an edge detection method based on multi-scale hybrid wavelet transform is proposed. A series of operations such as B-spline filter, Gaussian filter, multi-scale detection and adaptive threshold removal with pseudo-edge contour are carried out to remove the complex noise of image and to obtain the precise edge. Due to the importance of underwater image edge extraction in underwater target recognition, this method is applied to the edge detection of underwater images. The experimental results show that the method based on multi-scale hybrid wavelet transform can suppress the image noise and extract the accurate edge contour of the underwater image, which proves the effectiveness.

Functional medicine imaging has been successfully applied to capture functional changes in pathological tissues of the body in recent years. SPECT nuclear medicine functional imaging has the potential to acquire information about areas of concern (e.g., lesions and organs) in a non-invasive manner, enabling semi-automated or automated decision-making for the purposes of disease diagnosis, treatment, evaluation, and prediction. To reliably identify that whether or not at least one hotspot or lesion presents in a whole-body SPECT image, in this work, we develop a group of CNN-based classifiers. Specifically, we first propose a preprocessing method that transforms each original SPECT file into the required form by deep learning model, including normalization, 3-channel construction, rotation and scaling, size standardization, and size adapting. Second, six different classifiers are constructed by fine-tuning parameters of the standard VGG-16 model. Last, a group of real-world SPECT whole-body bone scan files were utilized to evaluate the developed classifiers. Experiment results shows that our classifiers are workable for the 2-class classification of SPECT images, achieving a best value of 0.7641, 0.6678, 1.000, and 0.6574 for defined evaluation metrics Acc, Pre, Rec, and AUC, respectively.

Distribution network communication operation and maintenance platform rely on communication network. The distribution network substation needs to obtain distribution network operation data configuration through communication network. The reliability of communication network directly affects the protection performance. Therefore, there is a need to establish a communication network that is reliable and meet protection performance requirements. In order to improve the engineering application of distribution network communication operation and maintenance platform, it is necessary to quantitatively evaluate the data quality of the distribution network communication operation and maintenance platform. Based on the construction background of smart distribution network, this paper studies the architecture of distribution network communication operation and maintenance platform. Starting from the requirements of information interconnection, intercommunication, interoperability and plug-and-play, the description specification of distribution network communication operation is established according to the principle of artificial intelligence data analysis. The unified modeling of protection and control system is realized. By analyzing the business composition of regional protection and self-healing control system of distribution network, the size of network information flow is quantitatively evaluated, which provides a basis for the engineering application of regional protection and self-healing control system of the distribution network. Experimental analysis shows that compared with the traditional method, the comprehensive performance of this method is improved by 10%. The partial performance is improved by 40%, which can support the master-slave regional protection and control business system.

With the increasing development of the market economy, interpersonal communication has become increasingly frequent. With the continuous advancement of science and technology, communication methods have been increasingly developed. Computer video conference management system also emerged under this historical background and made great progress. The application field of video conferencing is expanding and has become a new and important method of information communication in today's society. This article will briefly analyze the basic principles and usage methods of video conferencing systems.

With the development of metro, the grade of automation(GoA) of metro operation system has been greatly updated. The information between onboard train control and management system(TCMS) and track side maintenance system has become more complex, and the integration of control and communication networks has become a consensus target in metro industry. The control information of traditional train control and management is based on multi-function vehicle bus(MVB) without wireless network, and the security of information transmission can be neglected. In the system of control and communication network integration, the security of information transmission has become a key element. This paper proposes a wireless network maintenance information transmission protocol for metro TCMS based on RSSP-I with advanced encryption standard(AES), which can defense the 7 threads of metro TCMS suitable for GoA4 vehicles.

A new round of technological revolution and industrial revolution with artificial intelligence as the core is affecting all fields of society. However, grid generation is still a great challenge for computational fluid dynamics (CFD) simulations over complex geometries even though CFD researchers have made great progress in this field after decades of effort. In order to improve the traditional advancing front method of unstructured grid generation, scholars at home and abroad have carried out many beneficial attempts and explorations in finite element mesh generation using artificial intelligence. Firstly, this paper reviews the development history and present situation of artificial intelligence technology. Then, the application status of artificial intelligence in grid generation, size field prediction and grid optimization under the background of the big data era is discussed, and the advantages and disadvantages of various methods are pointed out. Finally, the development trend of artificial intelligence for grid generation is described.

This research builds a time synchronization demand model, a clock network model and a hierarchical distributed architecture for low-carbon Operation Services in smart campus (cluster) , combined with the existing time synchronization network, beidou time service, 5G network time service, multi-source heterogeneous communication protocol time service and local time service multi-source synchronization, breakthrough precise remote time tracing technology, it can provide high-adaptability and high-precision synchronous time tracing source for new low-carbon optimized operation of Smart Park (cluster) carbon measurement, electric carbon fusion, carbon emission, carbon footprint monitoring, electric spot market and so on.

Most of the supports in the engine are hollow structures. The hollow supports are the necessary paths for the internal acoustic emission (AE) signals to propagate to the shell. In this paper, the variation trend of amplitude and energy of AE signals with different frequencies in the propagation process of hollow supports are studied by numerical simulation and experiment. The amplitude and energy attenuation coefficients are proposed to quantify the propagation characteristics of signals. Through the pencil lead-break experiments, the spectrum analysis and the time-frequency domain analysis based on wavelet transform (WT) of AE signals in the propagation process are carried out. The results show that the tail end of the hollow support can converge and enhance the signals, and validated through experiments. This study demonstrates the propagation characteristics of AE signals in the hollow support structure, which helps to understand the signal propagation in complex structures and optimize the placement of AE sensors to improve the received signal strength.

The smart measuring switch is mainly used to realize the normal connection and disconnection of the distribution line, provide short-circuit, overload and disconnection protection functions, and effectively control the accuracy of the electric energy meter in the electric energy metering box. Because the communication protocol and test method of smart measuring switch are not released, the interconnection and exchange effect of products is poor. In view of the above problems, this paper first analyzes the problems existing in the protocol extension based on DL/T 645, then puts forward the protocol extension design based on DL/T 698.45, and designs the protocol consistency and interoperability test cases, which provides a feasible technical reference for realizing the interconnection and exchange of smart measuring switches.

Distribution network communication operation and maintenance platform rely on communication network. The distribution network substation needs to obtain distribution network operation data configuration through communication network. The reliability of communication network directly affects the protection performance. Therefore, there is a need to establish a communication network that is reliable and meet protection performance requirements. In order to improve the engineering application of distribution network communication operation and maintenance platform, it is necessary to quantitatively evaluate the data quality of the distribution network communication operation and maintenance platform. Based on the construction background of smart distribution network, this paper studies the architecture of distribution network communication operation and maintenance platform. Starting from the requirements of information interconnection, intercommunication, interoperability and plug-and-play, the description specification of distribution network communication operation is established according to the principle of artificial intelligence data analysis. The unified modeling of protection and control system is realized. By analyzing the business composition of regional protection and self-healing control system of distribution network, the size of network information flow is quantitatively evaluated, which provides a basis for the engineering application of regional protection and self-healing control system of the distribution network. Experimental analysis shows that compared with the traditional method, the comprehensive performance of this method is improved by 10%. The partial performance is improved by 40%, which can support the master-slave regional protection and control business system.

Today, my country's economy is developing rapidly, and it has basically entered the information age in an all-round way. The application of information technology is more and more extensive, and it has effectively improved the efficiency of life and work. As one of the important parts of information technology, electronic communication system can provide greater convenience for communication between people. Therefore, contemporary people's dependence on electronic communication equipment is becoming stronger and stronger, and the development of key electronic communication technologies can also promote the sound development of my country's electronic communication system. This paper analyzes the application of key technologies in electronic communication and the measures for building network architecture models for reference.

In this design, a 0.35um BCD process is used to design a high power supply rejection ratio LDO with feed-forward ripple elimination. By analyzing the power supply noise interference, a feed-forward ripple elimination circuit is used to reduce the power supply noise on the output. In order to ensure that the LDO has a high power supply rejection ratio in a wide load range, the output voltage of the LDO is 1.8V, the input voltage range is 2.5V to 5V, the load current range is 0 to 20mA, and the quiescent current is less than 80μA. The simulation results show that when the input voltage is 5V, the PSRR at low frequency at no load is 114dB, and the PSRR at 10 kHz is 77dB; when the load is 20mA, the PSRR at low frequency is 104dB, and the PSRR at 10 kHz is 72dB.

In view of the traditional RC oscillator's low accuracy and inability to be adjusted, a new repairable and adjustable high-precision RC oscillator is proposed. The oscillator is composed of two inverters, comparators, controllable switches, capacitors and resistors. The comparator generates output voltage by comparing the capacitor voltage value and the reference voltage value. At the same time, it controls the working state of the switching MOS tube to complete the cyclic charge and discharge of the capacitor. The two groups of capacitors charge and discharge alternately, and then produce continuous oscillation waveform. The high-frequency clock is realized by alternating comparison between bilateral comparator and bias voltage. The advantage of this mode is that the oscillation frequency is only related to the charging time, but has nothing to do with the discharge time. There is no need to consider the influence of the discharge time delay on the frequency. In this paper, the coarse adjustment of the accuracy of the oscillator is realized by adjusting the charging current of the capacitor, and then the output of the oscillator is accurately adjusted by selecting the capacitance value of the capacitor.

With the development of electrode materials and noise reduction algorithms, the wearable electrocardiogram (ECG) monitoring equipment has been widely applied in our daily life. Since these monitoring devices are usually used in sport scenarios, the collected ECG signal is easily contaminated by different kinds of noises and artifacts, especially motion artifacts, and thus the noise reduction has become an urgent problem to be solved for the subsequent clinical application. In this paper, a motion artifact removal method of dynamic ECG is proposed based on variational mode decomposition (VMD) and adaptive filter, which fully explores the correlation property of the acceleration signal and motion artifacts. From experimental results we find that our proposed approach can effectively suppress the motion artifacts in the dynamic ECG signals.

A low power consumer capacitor-less low dropout regulator (LDO) with supper class AB CMOS operational transconductance amplifier (OTA) has been proposed in this paper. The OTA of this design is based on the combination of class AB differential input stages and local common-mode feedback which provides additional common-mode sensing techniques thus to obtain a high common mode rejection ratio (CMRR). Local common-mode feedback (LCMFB) is applied to various class AB differential input stages, leading to different class AB OTA topologies. The presented LDO is fabricated in a 0.18 um standard CMOS process. The circuit consumes a quiescent current of 1.37 uA, regulating the output at 1.0 V with maximum output current of 600 mA from a voltage supply of 1.2V. It achieved full range stability from 1mA to 100 mA load current at a maximum 100pf load capacitor.

In order to solve the problem that the main flap beam of the conventional adaptive beam forming algorithm cannot correctly point to the desired signal direction of the target under the array error, a robust adaptive broadband constant beamwidth digital beamforming method based on spatial response variation constraint is proposed. Firstly, the beamformer with constant beamwidth based on spatial response variation constraints have been designed. Secondly, for the array error, the relationship between the error vector norm between the real steering vector and the assumed steering vector and the array error matrix is derived, and an inequality optimization model is established. Finally, the proposed method is a non-convex problem, which is transformed into a convex programming model through matrix decomposition and the idea of changing elements, and is solved by the convex optimization toolbox. The simulation results show that the proposed method is more robust than some other methods.

At present, there is no effective monitoring method for drop insurance. Temperature is an essential indicator for the operation status of drop insurance. Therefore, this paper designs a temperature monitoring system for drop insurance based on the advantages of infrared temperature sensors and wireless communication technology. The system takes infrared temperature sensors based on thermopile technology, LPC824 processor, and ZigBee technology as the core and solar panel and lithium battery as the system power supply. Furthermore, the system has the advantages of real-time monitoring, stable performance, and low power consumption. It provides a practical and effective solution for online monitoring of the temperature of drop insurance and the alarm when the temperature is abnormal.

This paper introduces six innovative designs of comparators proposed in recent years. The original intention to propose these improvements mainly focuses on the indexes including speed, area, power and offset, which measure the performance of comparators. Several newly proposed techniques are effective in designs. An improved DOC (dynamic-offsetcancellation) could increase speed and save area; Floating inverter pre-amplifier (FIA) can realize maximizing energy efficiency and reducing noise; Importing on-chip inductors in regenerative comparators has been verified reaching attractive goals successfully. Modifying structures and transistor type is also a feasible approach. After optimization, dynamic bias latch-type improves its predecessor’s performance in energy and noise; N-parallel paths could be applied to improve driving capability of latch; A novel voltage comparator termed an edge-race comparator (ERC) outperforms the conventional edition---the edge-pursuit comparator (EPC).

In the process of research on ship path tracking control, considering that the rudder angle needs to be optimized and the rudder amplitude and speed are constrained, this paper proposes a model predictive control (MPC) algorithm. High-order nonlinear observers are designed to avoid the impact of environmental interference and solve the problem that the highorder state value of the system is not easy to measure. At the same time, the speed state of the ship and the total unknown items including model uncertainties and external interference are estimated. The prediction model in the article uses a separate ship model that considers the response system of the steering gear, which makes the ship motion control more in line with the actual situation and improves the accuracy. Finally, it is verified by Matlab simulation. The designed controller enables the ship to track the reference path under the time-varying disturbance of wind, waves and currents, and the rudder angle is small and smooth. The results show the effectiveness of the design.

With the development of space science, the research on the performance of space deployable antenna has become one of the hot topics. The finite element modeling method was proposed and the corresponding finite element model was established. The antenna fundamental frequency was 1.0291 Hz, the experimental result was 1.13 Hz, and the error was 8.9%, which proved the accuracy of the established finite element model. In order to study the influence of high and low temperature environment on the performance of the antenna in orbit, a thermal deformation analysis method of a spacedeployable composite antenna in orbit was proposed. After the analysis, the maximum stress and maximum strain of the antenna were 39.54 MPa and 1.002 mm after the temperature change from -50℃ to 70℃. It indicates that the performance of antenna in orbit is greatly affected by high and low temperature environment. The modeling and analysis method described in this paper can provide a reference for the design, analysis and optimization of deployable structures of the same type.

Synaptic devices are progressing towards large-scale integration, but artificial neuron device research is still in its infancy. NbOx is an artificial neuron core material, but the Vth, Vhold of the threshold shift of NbOx will shift within a certain range, which will be unfavorable for making oscillating neurons.In this paper, We explore the Vth and Vhold shift and electrical properties of NbOx volatile memristor in different temperatures. The results indicate that the window of I-V behaves more stable at low temperatures, and higher temperatures would make the window stochastic increasing and more prone to failure. Furthermore, this suggests that the reduction of the high resistance state of the volatile NbOx memristor is due to the reduction of the Schottky barrier.

Pulse Position Modulation (PPM) is an important modulation method in communication systems, which plays an important role in many branches. This paper is designed to investigate the basic principles and processes of PPM modulation, to derive the results of each branch of this project, to analyze the advantages and disadvantages of Pulse Position Modulation, and to determine the appropriate application direction for PPM and the scope for remediation. This project is based on the MATLAB software and uses a PPM signal with only four symbols to carry out modulate and demodulate it in four steps, thus completing this project. In the first part, PPM signals are generated in the form of a random sequence of Gray codes, resulting in a large number of points; Then, the generated signal passes through a Gaussian white noise channel, and the modulation points are placed on a constellation diagram of the PPM plane for observation. After that, demodulation is performed by using the minimum distance discrimination method, and some error points are listed. Finally, the SER vs SNR curve is plotted. This project can basically explain the working principle of PPM, help researchers to get familiar with the PPM modulation process, and make it easier to choose how to work with communication in practice. It will guide the research direction on how to improve the accuracy of PPM in the future to broaden the application of this method.

Pulse-position modulation (PPM) is a form of signal modulation in which M message bits are encoded by transmitting a single pulse in one of 2M possible required time shifts. It is primarily useful for optical communications systems, which tend to have little or no multipath interference. PPM is currently being used in fiber-optic communications, deep-space communications, and continues to be used in R/C systems. In this paper, we simulate a relatively complete process of PPM transmission. We generate the basic form of PPM using random sequences and then we visualize the impact of AWGN on the signal on the PPM real plain. Then we use the minimum distance discrimination method to demodulate the signal. Details about errors are collected as well to illustrate the point. At last, we imitate the PAM system to calculate the BER.

This paper introduces the basic knowledge about PPM and four PPM related applications, namely, PPM modulation and demodulation based blue-green laser communication system, PPM modulation based underwater optical communication system design and simulation, PPM based optical communication positioning accuracy enhancement, and atmospheric laser communication PPM modulation and demodulation system design and simulation. In all four applications, PPM plays a vital role and the applications reflect the various uses of PPM in different areas of communication. This article will provide further insight into the applications and knowledge of PPM.

This paper studies five innovative designs of comparators proposed these years. The dynamic bias comparator ensures that the pre-amplifier output nodes are only partially discharged to reduce the energy consumption. The comparator with a floating inverter amplifier (FIA)-based pre-amplifier realizes the stability of input common-mode voltage and reduces influence of the process corner, moreover, thereby greatly boosting gm/ID and reduce noise, offset and delay. The edgepursuit comparator (EPC) has unique ability to adapt energy cost automatically, it can provide a new idea for the design of comparator. Triple-latch feedforward dynamic comparator (TLFF) with minimized stacking achieved < 70-ps delay in a wide common-mode (VCM) and power supply (VDD) range, and with the increase of input voltage, its delay advantage is more obvious. Low-Power High-Speed Dynamic Comparator in the evaluation phase, the latch reduces energy consumption and delay by delaying activation and using small cross-coupled transistors

Automatic control system is the representative of social comprehensive technological innovation, which has the characteristics of digitalization, programming and flexibility, and provides more reliable technical support for the performance of mechanical production. Based on this, this paper takes the food packaging machine system as an example to optimize the existing automatic food packaging machinery control system, give full play to the advantages of automation ability, and promote the improvement of social productivity.

In order to solve the problem of low positioning accuracy of disaster relief personnel and rescue efficiency when a disaster occurs in a coal mine, this paper proposes a disaster relief personnel positioning system for coal mine based on Micro-Electro-Mechanical System (MEMS) sensor. The proposed system uses the MPU9150 inertial sensor to obtain measurement data and the CC2530 microprocessor as the main control chip for data acquisition and processing. In the proposed system, the Pedestrian Dead Reckoning (PDR) algorithm is used to determine the step size based on the fusion expression between the walking frequency variance and acceleration, and the quaternion method is used to estimate pedestrian orientation angle. In order to reduce the error caused by the drift of the accelerometer and gyroscope, the extended Kalman filter is employed to correct the original data. The experimental results show that the positioning error of the proposed system is less than 1.6 m in 100 m. Thus, the proposed method can achieve high accuracy in the disaster relief personnel positioning in an underground coal mine.

Electrostatic potential testing is an important parameter for characterization of the electrostatic source. A non-contact electrostatic potential test device is designed for zreo-point drift, interference effect, narrow frequency range and other common problems in the current non-contact electrostatic potential test device in this paper. The above problems such as zreo-point drift, interference effect and narrow frequency range are solved. The basic principle of non-contact electrostatic potential test is introduced. The hardware and software design and implementation process of non-contact electrostatic potential test device are described in detail. The test verification and application of the device are given. Finally the key technical problems solved in the design implementation process are summarized.

This article has researched on the new comparator design technical, including Four kinds of improving method. 1. A new energy-efficient ring oscillator collapse-based comparator named edge-pursuit comparator changes the comparison energy and adjusts the performance automatically without control, which decrease the unnecessary energy consumption. 2. A comparator with a dynamic bias pre-amplifier and compares it with current technologies in terms of energy consumption and input referred noise voltage. 3. A new type of low-power, high speed dynamic comparator, gm-enhanced, applies the separated gate-biasing cross-coupled transistors, the comparator becomes faster and have less consumption. 4.A floating reservoir capacitor called FIA is presented, which improves gm/ID to decreases the noise and reuses the current to achieve 7 times energy-efficiency improvement than the previous technology, SA latch. The cases above are studied step by step. Moreover, the rules and formular of the cases are derived to find their character, which will benefit to applying in the different requirement.

A communication system for remote communication is designed. The system uses FPGA as the main control unit and pulse position modulation PPM as the basic modulation mode. Aiming at the problem that frame synchronization can not be realized in PPM communication, this design adopts the way of adding frame head frame tail structure and inserting protection gap to ensure information synchronization. In addition, the four-phase clock synchronization extraction method is used in the synchronous demodulation of PPM signal at the receiving end, which effectively reduces the error rate of FPGA in the working process, and greatly simplifies the design of the whole system. In this design, the encoding of PPM will use Gray code mapping to reduce the bit error rate. Finally, the system achieves a faster communication rate, and the BER of the actual test is low.

The demand of aeroengine for non-contact telemetry information transmission is very urgent. Based on the analysis of the characteristics of aeroengine telemetry information transmission system, the rotor disk and stator disk suitable for nearfield communication under high centrifugal load are designed in this paper. In order to ensure the stable transmission of telemetry information under large angular displacement between rotor disk and stator disk, a loop-shaped antenna with centrosymmetric characteristics is designed and installed on stator for receiving telemetry information. The designed circular antenna is fabricated and measured. The measurement results show that the circular receiving antenna has good reflection coefficient in the range of 1-4GHz, and its performance can fully meet the needs of aeroengine telemetry information transmission.

GPS technology has the characteristics of high precision, high sampling, real-time, and simultaneous determination of three-dimensional coordinates of points, which can not be compared with other monitoring technologies. It plays a very important role in deformation monitoring. Starting from the composition of GPS positioning system, this paper expounds the three components of GPS positioning system, as well as the association and coordination between each component of the work; then the GPS deformation monitoring mode and several error sources in the monitoring process are introduced. The advantages and disadvantages of GPS technique in deformation monitoring are analyzed and its application trend is predicted. GPS positioning technology is applied in all aspects of our life, creating a lot of social and economic value for us.