DAS ( Distributed fiber Acoustic Sensing) is a new sensing technology. However, the effective signals in DAS data are covered by a variety of complex and strong noise with a very low signal-to-noise ratio, which seriously affects the subsequent signal inversion and interpretation. Therefore, an attention-guided deep network (ADNet: Attention-guided Denoising convolutional neural Network) is proposed for DAS exploration data intelligent denoising. Compared to traditional methods, an attention-guiding module is introduced into the network to generate an attention-characteristic map, so that the deep network focuses on the parts with strong characteristics to improve the performance of network model in denoising. Through testing and comparing with traditional methods, it is proved that ADNet has great advantages in noise reduction and efficiency improvement.

The random noise in seismic data seriously affects the accuracy of subsequent data processing and interpretation. Therefore, an intelligent seismic random noise suppression method based on convolution neural network is proposed. Firstly, a deep nonlinear noise suppression network is designed according to the principle of convolution neural network, and then the network is trained by using the constructed high-quality random noise training set, so as to realize the automatic learning of random noise characteristics in high-dimensional space, so as to fit the nonlinear mapping relationship between noisy seismic data records and random noise, Realize automatic suppression of random noise. This method is used for noise suppression of seismic data, and compared with the commonly used filtering algorithms ( mean filtering method and median filtering method ). The experimental results show that this method has higher signal-to-noise ratio and overcomes the problems of traditional methods. An example verifies the feasibility and effectiveness of this method.

Aimed at the shortcomings that the total energy efficiency and capacity of the system can not be
considered at the same time of the classical resource allocation scheme, subcarrier average power allocation
scheme and power allocation scheme based on maximum communication capacity and maximum system energy
efficiency in SCMA( Sparse Code Multiple Access) downlink system, a hybrid SCMA power allocation scheme
based on scale factor control in downlink is proposed. The capacity and energy consumption of the system are
optimized as the main direction, and the system could flexibly adjust the tendency by inserting the scale factor.
Since the problem is a nonconvex optimization problem, it is first transformed into a convex optimization problem
by GDA (Generalized Dinkelbach's Algorithm), and then the reasonable Lagrange value is selected to iteratively
solve the model by used KKT (Karush-Kuhn-Tucker) conditions. Finally, the energy efficiency and capacity of
the system are optimized by adjusted the scale factor. The theoretical analysis and simulation results show that
the proposed hybrid law allocation scheme can effectively improve the downlink system capacity and energy
efficiency compared with the traditional scheme.

 In order to achieve higher DAC(Digital-to-Analog Converter) resolution, a general structure of voltage
output DAC which is suitable for arbitrary radix is proposed. The general structure of the DAC mainly includes a
resistor network, a set of single-pole multi-throw analog switches, a set of reference voltage sources and a buffer
amplifier. On the basis of the proposed circuit structure, an 8-trit ternary DAC circuit module is designed using
discrete components and integrated circuit chips, and the test data is analyzed for errors. After theoretical
analysis, compared with the traditional DAC, the 8-trit ternary DAC has a resolution increase of 25. 6 times
without significantly increasing the complexity of the structure. And the DAC structure also has the characteristics
of simple structure and strong expansibility on radix and number of digits.

In order to improve the subdivision accuracy of the photoelectric encoder and realize the compensation
and correction of the Moire fringe signal, the source of the error that affects the signal quality is analyzed, and
the waveform equation of the Moire fringe signal is established. And a method using the cuckoo search algorithm
combined with the least square method is designed, and using the residual sum of squares as the fitness function
to realize the multi-parameter identification of the waveform equation. And the sinusoidal deviation, DC(Direct
Current) component and amplitude deviation in the signal are corrected by using the identified parameters and
the waveform equation. The identified phase parameters are used to construct a look-up table to compensate the
orthogonality deviation of the signal. The experimental results show that the method effectively reduces the
deviation in the output signal and improves the subdivision accuracy of the photoelectric encoder.

Aiming at many defects of the traditional speed sensorless vector control system of asynchronous motor
based on rotor flux linkage, such as serious overshoot, weak robustness and poor load capacity, our object is to
improve the traditional speed sensorless model of asynchronous motor. The speed PI ( Proportion Integration)
regulator in the traditional speed sensorless model of asynchronous motor is combined with the neural network
adaptive control algorithm, and a third-order low-pass filter is added to the current model of rotor flux linkage.
The simulation analysis is carried out for the start-up of asynchronous motor, the speed torque and the change of
estimated speed after sudden load increase. Experiments show that the improved speed sensorless vector control
system of asynchronous motor obviously eliminates the overshoot, effectively improves the robustness of the system
and enhances the load capacity of asynchronous motor.

Aiming at the disadvantages of lower speed accuracy, show dynamic response and complexity on
parameters selecting of vector control system of rail refrigerator used PMSM ( Permanent Magnet Synchronous
Motor), the traditional controller of speed loop and current loop is optimized by bacterial foraging algorithms.
The specific innovation is the combining of bacterial foraging algorithm with gravitational search mechanism to
form GS-BF(Gravitational Search-Bacterial Foraging) algorithm, so bacterial individual swimming action will be
endowed better target performance to speed up the efficiency of swimming. The bacterial foraging algorithm has
better performance on global optimal solution and convergence speed. Finally, the superiority of GS-BF algorithm
is verified by Matlab. The simulation results show that the optimized PID ( Proportional Integral Derivative)
controller has smaller ultra-tuning, faster response speed. The performances on static error and dynamic response
speed of the vector control system are effectively improved. This method has good universality and can be applied
to the speed control system for other motors.

In order to guarantee the welding quality of gas pipe joints and prevent serious accidents, a high performance welding control system for thermal melt pipe fittings is developed. The system uses bidirectional thyristor as the main control circuit, and uses timer interrupt technology to control the bidirectional thyristor through PWM( Pulse Width Modulation) to realize the adjustable voltage output of the thermal melt welding system. In order to improve the reliability of the system, the errors generated by the information collected by the sensor are analyzed. And the nonlinear and zero drift errors of the sensor output are reduced by the piecewise dynamic calibration of the sensor information using the least square method. The adaptive sliding filtering method is adopted to suppress the errors of noise and power frequency interference. For the control voltage, an improved PID( Proportion Integration Differentiation) fast voltage regulation algorithm is designed to improve the output response speed and output stability. The field test shows that the welding process can reach steady state quickly and effectively suppress the interference of environmental noise to the data acquisition, the voltage fluctuation is less than +-0. 1 V, lower than +-0. 5 V national standard, which meets the requirements of welding process.

In recent years, microblog has become an important platform for Internet public opinion disseminationand public opinion mining. In order to analyze the impact of epidemic events on Netizens' emotions, we should do a good job in prevention and control publicity and public opinion guidance scientifically and efficiently.Therefore, we integrate different deep learning methods to conduct emotional analysis of microblog comments on the COVID-19 outbreak at the end of 2020. A hybrid model based on RNN(Recursive Neural Network) and LSTM (Long Short-Term Memory) and using the FastText word vector representation in the embedding layer is proposed to reduce the noise data in the word vectors and thus obtain high-quality word vectors with semantically
rich and less noise. Training on Weibo corpora and compared with Bayesian and Support Vector Machine, RNN,LSTM multiple methods, the results show that the accuracy of the emotion analysis model proposed in this paper reaches 98. 71% , which proves that the model can effectively improve the accuracy of emotion analysis.

To solve the problem of incomplete information in air combat, research of offensive and defensive game strategy of UAV (Unmanned Aerial Vehicle) based on an improved sparrow algorithm is conducted. By analyzing the air combat information of the enemy and ourselves, the offensive and defensive game model is established in the case of determining the air combat situation advantage function, air combat performance advantage function and profit function of the enemy and ourselves. Then, combining the game payout function given by the game model and calculating the payout matrix of the enemy and ourselves, an improved sparrow algorithm based on reverse learning is proposed to solve the Nash equilibrium solution of the offense and defense game between the enemy and ourselves. Finally, the feasibility and the effectiveness of the proposed algorithm are verified through simulations. This scheme can preliminarily solve the problem of incomplete information encountered in the process of air combat confrontation.

Procedural knowledge plays an important role in the process of human learning and cognition. In the past, the organization of procedural knowledge was realized through a production system, which is a non-semantic method that can not reflect the relationship between knowledge and is difficult to maintain. Combining ACT-R theory, event cognition theory and event models in the literature, a seven-tuple semantic model of events that can represent procedural knowledge is proposed. Taking the procedural knowledge in the Java language as an example, a procedural knowledge event ontology model is constructed, and an ACT-R ( Adaptive Control of Thought-Rational) cognitive model based on the procedural knowledge event is designed. The procedural knowledge event ontology can realize the semantic complete description of the procedural knowledge and the relationship between them, providing new ideas for the organization of procedural knowledge teaching resources and the network teaching based on procedural knowledge.

With the continuous addition of distributed generation to the distribution network, the output of active distribution network will be unstable and the supply-demand side will be unbalanced, and its economic aspects need to be further studied. Based on the consideration of distributed energy, network loss cost and generation cost, an economic dispatching model of active distribution network is established. It adds soft open point to carry out flexible dispatching and reactive power optimization of active distribution network, and solves it by using the optimization beetle antennae search algorithm. The optimization method is to add genetic factor and tournament selection strategy to the algorithm. The algorithm results will not fall into local optimization and the minimum generation cost is obtained. Finally, the simulation results show that the optimized beetle antennae search algorithm is practical and efficient in the economic dispatching problem of active distribution network.

To address the adverse effects of missing data in the medical dataset on the performance of the classifier and on downstream tasks. We use the missing forest interpolation method to interpolate missing values in medical datasets. The method first trains a random forest model with observations of complete data in the dataset. Then the trained random forest model is used to predict the missing data. Finally, the above process is repeated iteratively to complete the missing data interpolation. On two medical datasets, according to NRMSE(Normalized Root Mean Squared Error) and PFC( the Proportion of Falsely Classified) evaluation metrics, the missing forest interpolation method has lower error and better interpolation than K-nearest neighbor interpolation,multiple interpolation and GAIN( Generative Adversarial Imputation Nets) interpolation. The stability of the missing forest interpolation method is demonstrated by analyzing the relationship between glutamate aminotransferase (ALT: ALanine aminoTransferase) and diabetes dose-response using the diabetes dataset.

In order to make the bearing fault diagnosis more accurate and intelligent, aiming at the problem of bearing fault feature extraction, a CNN(Convolutional Neural Network) based on residual structure and Inception structure is constructed, and a new bearing fault diagnosis method is proposed. First, the STFT ( Short-Time Fourier Transform) is used to transform the original one-dimensional signal of the rolling bearing into a two-dimensional time-frequency graph, which is divided into a training set, a validation set and a training set. Then the training set is used to iterate the built Inception-residual network model. The network parameters are constantly updated, and the verification set is used to check whether the model has over-fitting phenomenon. Finally, the trained model is applied to the test set, and the diagnosis result is output through the classifier of the output layer.The experiments proved the feasibility of the proposed method, and the average accuracy of bearing fault classification reached 99. 98% +-0. 02% , which has a higher accuracy and stability than other methods.

In view of the safety problems caused by low detection efficiency and illegal operation in traditional gas stations, a modified gas station oil unloading detection method based on YOLOv3 is presented. By introducing RFB (Receptive Field Block ) receptive field module after Darknet-53 backbone output, the model can select appropriate receptive fields to match different scale targets and improve detection accuracy. According to CSP (Cross Stage Partial) network, two RFB_CSP and RFBS_CSP structures are provided to realize cross-level splicing and channel integration and reduce calculation cost. Cluster 9 target reclustering in the field is realized by using K-means++ algorithm to determine appropriate network anchor parameters. The experimental results show that the optimized model contrasts the original YOLOv3 model. The average accuracy is increased by 2. 3% and 2. 9% , indicating that the optimized YOLOv3 model has high practical value in gas station scene detection.

Approximately 3% to 10% of hepatitis C cases can develop to hepatocellular carcinoma after viral hepatitis C virus infection. Worldwide, 27% of cirrhosis statistics are due to hepatitis C and 25% are due to hepatocellular carcinoma. Accurate prediction of hepatitis C infection is a matter of urgency. Machine learning is fast and accurate. Hepatitis research often used time series analysis or pathological analysis in the past and did not use machine learning algorithms as an auxiliary diagnosis method for hepatitis C. To select the optimal model for detecting hepatitis C, different machine learning models are compared and analyzed in UCI(University of California Irvine) hepatitis C data. The experimental results show that gradient boosting tree, random forest and light gradient boosting machine perform better, among which the gradient boosting tree is accurate in predicting hepatitis C up to 0. 935 1. The most accurate prediction of hepatitis C infection is performed using gradient boosting tree.

In order to solve the problem of low accuracy in user similarity calculation of traditional collaborative filtering algorithm, an item attribute and item tag information is introduced into the recommendation system, and proposes a hybrid recommendation algorithm is proposed based on tags and attributes. Firstly, the user's score on the item is transformed into the user's score on the item attribute value and label, and the user-attribute rating matrix and user-tag rating matrix are constructed as user description files. Then the similarity between users is calculated according to the user-attribute rating matrix and user-tag rating matrix, and the results are average weighted to obtain the nearest neighbor list of each user. Finally, the recommendation result is generated according to the neighbor's score on the item. Since the number of item attributes and major tags are much lower than the number of items, the algorithm can effectively solve the sparsity problem of collaborative filtering algorithm, and describe the user preference more intuitively. In the process of constructing the user description file, considering the law that the user preference changes with time, different weights are given to the user's scores at different time points, and the weight increases gradually with the passage of time. Experimental results show that the proposed algorithm can predict users' ratings of unrated items more accurately and improve the accuracy and recall of recommendations.

Aiming at the problem that the data extraction method can not carry out massive mining, the mining results are inaccurate and the mining time is long, a regional strong association rule data extraction method based on data mining algorithm is proposed. Combined with the data management system of strong regional association rules, user demand information is collected, feature relevance is retrieved, drama features are obtained. The data relevance is used to analyze the association between drama features in drama retrieval, calculate similar label information parameters, calculate the support and confidence, and mine association rules from the database of strong regional association rules. Kulczynski measure and imbalance rate is used to implement correlation
analysis and filtering, and finally the strong association rules are obtained with practical significance. The experimental results show that this method has high mining efficiency and wide application value.

In traditional statistical methods, the conjoint analysis method is can not estimate variables for a large number of parameters at the same time, therefore, a Bayesian 茁 regression model is proposed. In the newly established model, the Dirichlet distribution is used as the prior distribution of the model parameters, and the relevant MCMC(Markov Chain Monte Carlo) algorithm is designed to fit the model. By analyzing the results of applying the model to the evaluation of discrete index variables, it is shown that the model has a good fitting effect on the data and the algorithm has a fast convergence speed. It shows that the Bayesian hierarchical model makes up for the defects of the traditional conjoint analysis method, and optimizes and improves the conjoint analysis method.

In order to clearly understand the basic principle and mechanism of mbedos scheduling, the support characteristics of arm cortex-m4 kernel for real-time operating system is studied. The common scheduling strategies of real-time operating system is analyzed. Using the methods of flow chart analysis and
context core code analysis,this paper focuses on the implementation of three mbedos scheduling strategies, SVC( Supervisor Call) interrupt, pendsv interrupt and systick interrupt. Finally, taking stm32l431 chip of STMicroelectronics as an example, the thread scheduling analysis practice of mbedOS is given. By analyzing the thread scheduling of mbedOS, it is helpful for readers to deeply understand the scheduling mechanism of mbedOS, provide help for the design of multitasking system, and also provide reference for the analysis of other real-time operating systems.

In order to overcome the shortcomings of SIFT ( Scale-Invarivant Feature Transform) algorithm, a method which combines Harris algorithm with SIFT algorithm is put forward. SIFT feature points within the corner neighborhood range detected by Harris algorithm are used as feature points. Goss circular window is used to establish a 64-dimension description vector for feature points. The experiment result shows that this algorithm can improve the matching accuracy and the matching speed.

With the increasing popularity of complementary energy storage power generation system, its coverage is also expanding. But the capacity optimization technology is still in the single energy stage. Therefore, a capacity optimization algorithm is designed for the solar and wind energy complementary storage and power generation system. By analyzing the power output of the power generation system, the state of annual output and supply is obtained, and the capacity optimization objectives and constraints are established according to the power balance criterion. Then, the differential evolution algorithm is used for mutation, crossover, selection and other operations, and the particle swarm optimization algorithm is used to update the orientation and velocity. The current optimal solution of parent body and population is obtained by the combination of the two, and the capacity optimization of energy storage system is realized through iterative cycle. In the experimental, through the data of component configuration quantity and load power shortage rate it is found that this algorithm greatly improves the economic benefits, and ensures the reliability of power supply, which shows that the algorithm has significant optimization advantages and great application potential.

For image clustering, the existing methods are either difficult to choose the appropriate dimension transformation method in feature extraction, or the extracted features are weak and not rich enough for the expression of image features, which have a great impact on the clustering effect of images and lead to low clustering accuracy. Therefore, an image clustering method based on multi-scale residual convolutional autoencoder is proposed. By constructing several multi-scale convolutional modules with residual connections, the high-dimensional feature expression of the middle layer is obtained, and the image is clustered based on these features. The clustering accuracy on MNIST data set is 82. 2% , ARI (Adjusted Rand Index) value is 0. 781 0 and NMI ( Normalized Mutual Information) value is 0. 853 2, indicating that the model has achieved good clustering effect.

In order to improve the color matching effect of film and television animation scenes, a design method of film and television animation scene color matching system based on feature point extraction algorithm is proposed. A visual cognition model of color matching in film and television animation scenes is constructed. The feature space structure of the auxiliary visual elements of the color matching graphics of the film and television animation scenes is extracted through the auxiliary visual image sampling model. Combine the three-dimensional distribution characteristics of the color matching graphics of the film and television animation scenes, the RGB feature decomposition of the color matching visual elements of the film and television animation scenes is established by the feature point extraction algorithm model. Using the method of visual information parameter fusion and three-dimensional vision reconstruction, the software platform of the color matching system of film and television animation scenes is developed and the color matching system of film and television animation scenes is designed. The three-dimensional reconstruction software is used to realize the visual simulation of the color matching system of film and television animation scenes. The simulation results show that the designed film and television animation scene color matching system has high output stability and good reliability.

In order to solve the society's demand for long-term and continuous ward monitoring, a ZigBee-based ward monitoring system is proposed. The system implements the collection of the ward environment and the patient's physiological parameters and uses the ZigBee technology to integrate and transmit data of the sensors.The nurse station management platform is developed to monitor and manage the information of each ward and patients, and integrates the ward environment, patient information and management background. Experiment indicates that this system can achieve stable communication for more than 24 hours in a 30 m²room, and the error rate of data is less than 2. 5% , which meets the daily monitoring and management needs of ordinary wards.