This PDF file contains the front matter associated with SPIE Proceedings Volume 12779, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

In transmission line inspection, using UAV inspection technology will help improve the quality and efficiency of transmission line inspection and reduce the error and security risk of manual inspection. After the UAV collects a large amount of data during the inspection of power lines, when passing the ground station with stronger computing power, the data can be unloaded to the ground station to help it complete the calculation to make sure the task is finished within a specific limited time. The flight starts and finish points of the UAV patrol are fixed, and are constrained by factors such as flight, communication, and energy. In the paper, the flight trajectory planning scheme of UAV is proposed to minimize the total of its communication and flight energy or power consumption while completing the unloading of computing data within the specified time, and the experiment shows that the design scheme can greatly dwindle the UAV energy consumption in the mission execution process.

The demand for lean management of distribution network is growing. However, the current identification of distribution network data quality lacks a unified standard and cannot achieve automatic data verification and processing, which restricts the improvement of distribution network application level. In the actual distribution network application, more experts rely on their experience to solve the quality and availability problems of local distribution network topology data as needed, lacking a global data quality evaluation and solution. With the development of smart grid, the demand for lean management of distribution network is growing. However, the current identification of distribution network data quality lacks a unified standard and cannot achieve automatic data verification and processing, which restricts the improvement of distribution network application level. In actual distribution network applications, more experts rely on their experience to solve the quality and availability problems of local distribution network topology data as needed, Lack of overall data quality evaluation and solutions. This paper first combs the status quo and quality problems of distribution network topology data, collates and summarizes the current status quo of distribution network topology data, summarizes the problems existing in the application of distribution network topology data, and researches and classifies the problems, laying a foundation for the development of distribution network topology data quality evaluation system and distribution network topology data verification rules.

Due to the diversity of data types and data organization methods of electric multi-source heterogeneous data, the organization and storage requirements of heterogeneous data are also different, so the difference of heterogeneous data must be considered for data integration and fusion, the integration and fusion level of power distribution and distributed new energy data resources needs to be improved. The distribution network data resources including electrical equipment, spatial information, grid topology, power consumption information, operating conditions and other types of resources, taking into account the new energy, are obviously different in terms of quantity, scale, data model, data type, organization mode and other aspects. The data of distribution network and distributed energy comes from business systems in multiple professional fields, and there is a strong correlation between data resources at the business level. However, due to the certain independence between the data of various business systems for power distribution and the differences in field definitions and descriptions, traditional key field matching methods are difficult to achieve automatic data matching, and data fusion across business systems is faced with the problem of no uniform rules to follow.

Aiming at the typical characteristics of constant changes in the power grid topology caused by multiple factors such as large-scale access of distributed power supply and diversified loads, this paper analyzes the shortcomings of existing power grid topology data management methods in temporal management of topology data, designs a power grid topology data model integrating time dimension based on big data and graph computing technology, and establishes a multi-temporal power network topology data management method. This method realizes the importing, processing, storage and reading of multi-temporal power grid topology, which provides a new solution for the power grid topology data management with multi-temporal version attribute, and also provides a powerful data support for the subsequent analysis and application of multi-temporal power grid topology data.

In the context of the poor time performance of some distributed energy storage capacity optimization methods, a distributed energy storage capacity optimization method based on smart node terminal technology is designed. It Quantifies the distributed energy demand response characteristics, extracts the economic index of the energy system, calculates the natural gas consumption of the internal combustion engine and the recovered waste heat respectively, obtains the storage capacity equipment operation coefficient, sets the application scenario selection mode based on the intelligent node terminal technology, improves the load factor of the gas internal combustion engine, and designs the storage capacity optimization allocation method. Results show that the average time required for the distributed energy storage capacity optimization allocation method in the paper is 72.36s, indicating that the designed distributed energy storage capacity optimization allocation method is more suitable for application to practical tasks after the introduction of smart node terminal technology.

Under the development and reform of the new power grid, the technology of drawing single line diagrams also needs to be developed and reformed urgently. In the past, single line diagrams were drawn mostly by manpower. On the premise of given data and topology structure, the drafters first sorted out the topological relationship between the components, and then placed the components on the canvas with the principle of minimizing line crossings, and the canvas size needs to be reduced as much as possible and draw the graph W in a way that is easy to understand with the aid of CAD and other tools. This paper adopts a new decomposition model to solve the problem of power system single line mapping. First, the steps of electrical topology automatic mapping are decomposed into H steps. Using the knowledge of graph theory in mathematics instead of the form of polynomials, multiple processing objectives are processed separately, and H different methods are used to process them step by step. Then, the improved mapping algorithm is used to generate the plane embedding of electrical graph topology, which is used to achieve the goal of minimizing the intersection of lines. Then, the improved least cost flow graph method is used to process the graphic dexterity with the least cross. On the one hand, the graphic elements are evenly distributed, on the other hand, the horizontal and vertical orthogonal graphs are generated and the bending of the lines is minimized.

With the sustained growth of scale of electric power system, the safety of distribution grid and the ability to deal with failures are facing new challenges. Aiming at the problem that the failure of a middle and high voltage transformer in the electric distribution grid leads to a large area of power failure and it is impossible to quickly select a transfer path to restore power supply in a short time, a transfer path identification scheme is proposed, which assumes that when a main transformer fails, multiple path selections are made after the transfer scheme is obtained through integer programming, and the electricity fluctuation problem before and after the new extended energy is considered. First, the topology of the distribution network diagram is simplified through topology analysis. Then, integer programming can overcome dispatching of electric power trouble to get the maximum recovery amount and the transfer path. Then, the transfer path obtained is filtered for many times, and finally the transfer path satisfying the conditions is output. The results after application in provincial power companies show that the method has short recovery time, strong real-time, high execution efficiency and strong adaptability.

In recent years, with the high level of China's market economy, the use of electricity resources is increasing, and the scale of thermal power plants is expanding, resulting in an increase in fire hazards. Therefore, thermal power plants are required to improve the design level in the design of fire protection system to ensure the design standardization. This paper provides a detailed analysis of the design of fire protection systems in thermal power plants, hoping to provide some help for the improvement of fire protection systems in thermal power plants in China.

At present, the power Internet of Things (IoT) border protection equipment gradually exposes its functions, and the technical specifications are not unified to deal with the problems of insufficient connection ability of high concurrent data, imperfect automatic monitoring function, complex product types, and difficulty in equipment selection. Although the traditional research supports remote maintenance, the information interaction effect is poor, which can't meet the high security characteristics. This problem forces the need for a data security communication protocol with high security and suitable for application scenarios. In this paper, the encryption technology of IoT gateway based on Bayesian algorithm is proposed to ensure the safe operation of power system. The simulation results show that the IoT gateway encryption method in this paper is safe and has practical application value for the safe operation of power system. Facing the needs of different business development, the software and hardware architecture should be integrated to form typical application protection schemes in different business scenarios, so as to achieve the goal of unified application security and accurate protection of basic network security.

The implementation of digital fraction delay has found diverse applications in signal processing, leading to extensive research in this area. To improve the accuracy of channel estimation in practice, Fractional Delay Finite Impulse Response (FD FIR) filters have been proposed. This paper presents the implement principle of fractional delay FIR filters. Building upon existing solutions, it makes improvements to the methods of windowed sinc function, smooth transition function, complex general LS design, and maximally flat design: Lagrange interpolation from simplicity, efficiency, and filter performance, particularly in the reduction of Gibbs phenomenon caused by truncation in time domain. The effectiveness of the proposed method is validated through MATLAB simulation and comprehensive magnitude response analysis, which demonstrates that the complex LS design exhibits the best response with a wider passband and almost no ripple or Gibbs phenomenon.

In order to address the lack of integrated simulation verification of control object and control strategy in the early design of most electromechanical integrated drive control systems as a reference for practical design, this paper proposes a modeling and simulation method for electromechanical integrated drive control systems based on Matlab models. This method combines the detailed parameter design of the control object with the design of the control strategy, and establishes an integrated simulation model in Matlab. Through system simulation analysis, it guides the design of the control object parameters and control strategies, and aims to use the simulation model as a reference for actual electromechanical integrated drive control system design.

As an important part of the DMS, the operation state of distribution lines is related to the operation safety of the DMS. Lightning is a natural transient weather phenomenon with high voltage, high current and strong electromagnetic radiation in nature, and it is also one of the more common natural disasters, which has become the primary factor of distribution line tripping. In this article, an algorithm for lightning monitoring and fault identification of overhead distribution lines based on the Internet of Things (IOT) is proposed. According to the distribution characteristics of wavelet energy spectrum under different lightning strikes, the response characteristics and fault identification criteria are constructed. The simulation results verify the real-time and reliability of the online monitoring data transmission of the lightning monitoring system for overhead distribution lines. The method in this article can avoid the influence of lightning impact corona and reflected wave at the end of the line on the accuracy of the algorithm by reasonably selecting the judgment threshold and the monitoring point of current traveling wave, which has high practical value. Through the statistical analysis of various monitoring data of the system, the operation and management department of distribution line can grasp the changes of key operation state of the line in time and lay a solid foundation for the condition maintenance of distribution line.

In this paper, a kind of digital twin dynamic polymerization technology based on clustering and intelligent analysis is studied. In the first step, the goal and business requirements of device-level digital twin polymerization are studied, and the principle of hierarchical and distributed dynamic polymerization of device-level digital twin is proposed. The second step analyzes the dynamic characteristics of the county energy Internet source network charge and storage, and studies the spatio-temporal dynamic aggregation modeling technology of the device-level digital twin based on the dynamic characteristics of the source network charge and storage. The third step is to study the hierarchical polymerization framework and polymerization method of the virtual power plant at the source, load and storage side facing the distributed multi-level interaction framework of the county energy Internet, the hierarchical polymerization framework and polymerization method of the county energy Internet side, and the hierarchical polymerization framework and polymerization method of the virtual power plant at the source, load and storage side facing the county network side.

In order to solve the problem of consuming a lot of time in data analysis, condition formulation and summary report compilation of satellite mechanics test, a set of software for data analysis of satellite mechanics test was developed. Firstly, the system framework is constructed from the process of satellite vibration test. Then, the functions of the software are divided by sorting out each link of the test data analysis. In addition, according to the flow of data in the test, design the engineering data structure. Finally, combined with the characteristics of data interaction and convenient use of the software, a general module of the software is constructed.

With the vigorous development of urban rail transit industry in recent years, the quantity and quality requirements for subway vehicles are higher. The existing electric control system of subway vehicles is controlled by the contact scheme represented by relay, which has many shortcomings such as high cost, high failure rate, and no redundancy in logic cascade. If the single point and relay fault in the control circuit, the subway vehicles cannot operate, and the problems of vehicle cleaning and rescue will occur. This paper introduces a design scheme of the standby mode of subway vehicle control circuit. The programmable Logic Control Unit (LCU) is used to design bypass, which can affect the key control circuit of the train, so that the subway can continue to run to the station in the fault state, avoid the congestion of the line, and solve the problem of vehicle clearing and rescue caused by the fault.

In the production work of the State Grid, the power equipment is often in a high and low temperature environment. The timely detection of the fault of the power equipment is an important part of the stable development of the State Grid. This paper proposes a method to detect the quality of power equipment based on the measurement data of power equipment at room temperature. Based on the LSTM model, the working data of power equipment is predicted in the form of serialized time nodes during the experiment. Then, the feature data is extracted and dimensionality reduction according to the operating environment, and finally the anomaly detection is carried out according to the Local Outlier Factor (LOF) algorithm. The experimental results show that the prediction effect of the model prediction diagnosis results is ideal, and the LSTM-LOF combined with the model can be used to assist engineers in the quality diagnosis of power equipment.

In order to solve a series of problems existing in the existing smoking rooms in public places, such as "poor indoor environment of smoking, easy adhesion of second-hand smoke to walls, tables and chairs and other dead corners, low smoke collection and processing capacity, resulting in pollution of the smoking room and the surrounding environment", this device designs a smoke purification device in the smoking room based on cutting off the continuous vortex ring air curtain, The utility model can efficiently replace the indoor air by isolating the air curtain, so as to achieve the comprehensive purification of the air in the smoking room. MQ-2 smoke sensor is used, which is controlled by STM32 single chip microcomputer through AD conversion and serial port input, and ADC0832 is used to convert the collected data to obtain the actual smoke concentration, The smoke concentration is controlled by the P port of STM32, and the oledspi display module located inside the smoking room displays it in real time. The smoke concentration digital signal compares the parameters through the obtained smoke concentration and the set smoke concentration threshold. When the processing requirements are met, the circuit of the digital relay controls the motor drive of the forced draft fan, controls the isolated suction of the vortex ring air curtain, and pumps the indoor air into the subsequent pretreatment system efficiently and quickly, In the prediction system, nano mineral crystal flue gas toxic gas adsorption and activated carbon filter plate are used to filter the flue gas generated by smoking, and then discharged outdoors after passing the detection of tail gas detection sensor.

A large quantity of Electric Vehicle (EV) charging station loads connected to the power grid will aggravate the peak-valley difference and reduce the stability of the power system and the economic benefits of operation. The connection of EV charging station loads will also inject a large quantity of harmonics into the power grid, further aggravating the risk of power supply equipment failure. EV can also be regarded as distributed energy storage devices. Under special circumstances, EV can feed back electric energy to the power grid through power conversion devices and participate in frequency modulation of the power grid. In order to reduce the idle rate of charging stations and promote the distributed photovoltaic absorption capacity of the system, this paper proposes an electric load forecasting model based on Particle Swarm Optimization Neural Network (PSO-NN). The charging and discharging power of EV cluster and energy storage equipment are solved in turn by PSO algorithm, and when the regulation capacity of EV cluster is insufficient or limited, it is supplemented by energy storage equipment. The simulation results show that with the increase of the quantity of experiments, the accuracy of this algorithm is stable at about 95%, and the real-time wavelength tends to be stable. Therefore, the scheduling strategy can effectively improve the safety and economic performance of power grid and the capacity of transportation system, which is conducive to improving the operational performance of power grid and transportation network and providing algorithm and technical support for the construction of charging station number planning model.

With the increasingly serious energy crisis and environmental problems, EV (Electric Vehicle) has become the development trend of automotive energy and environmental protection in the future. As an important supporting system for the development of EV, the charging infrastructure will inevitably affect the power quality of the distribution network when large-scale EV is put into operation. Therefore, this paper proposes an EV charging facility demand forecasting analysis method based on the charging load forecasting of the Markov model and establishes a charging facility demand forecasting model based on data fusion. Finally, the research shows that the voltage optimization of the simulated Markov model is better, but the satisfaction of the operation results of the charging pile configuration scheme before the simulated Markov model is 0.78, The operation result of the charging pile comparison planning scheme smaller than the simulated Markov model is 1.58, which indicates that the charging pile utilization rate of the charging pile comparison planning scheme simulating the Markov model is higher and the customer satisfaction is better. In this paper, Markov model is used to describe the change of the state of charge of EV users' charging piles during a day's trip, and determine the real-time charging requirements of different types of EV users.

The main goal of Nondestructive Testing (NDT) is to regularly test the pressure vessels in order to reduce the probability of failure accidents. On the basis of NDT, it can effectively avoid container breakage and ensure the safe use of containers. In the actual use process, we will face complex conditions, so we must attach great importance to the safety of container use. At present, there are still many problems in the inspection of pressure vessels and pipelines. Only by objectively understanding these problems can we better carry out the inspection of pressure vessels and pipelines and ensure the application safety of pressure vessels. This article takes NDT technology as the main research content, analyzes the combination of NDT and pressure vessel and pressure vessel management, and puts forward a fault detection algorithm for pumps, pressure vessels and pressure pipelines based on big data analysis. The simulation results show that the pressure equipment fault detection algorithm based on this algorithm is better than Support Vector Machine (SVM) in time measurement accuracy and efficiency and can accurately detect the equipment state and fault. Non- destructive testing technology has incomparable advantages over other technologies, but it also needs to constantly improve the application level to improve the work efficiency.

With the digital transformation of distribution network, it is a trend to use digital twins to simulate the status and operation of distribution network system. However, the current distribution network digital twin system is inclined to model three-dimensional display and data superposition, lacks the characteristics of power business, and is difficult to grow synchronously with the dynamic evolution of the power system. At the same time, the twin construction for massive distribution network objects consumes a lot of labor and time costs. Therefore, this paper introduces the grid topology and electrical state calculation model based on graph calculation, integrates the electrical topology to realize the rapid construction and automatic layout of distribution network data twins, and carries out distribution network business applications such as distribution network panoramic state monitoring, power supply range analysis, etc. based on the distribution network twins scene.

Based on the current safety relationship of new energy vehicle batteries, this paper proposes a battery supervision system based on NB-IoT cloud. In this paper, the current state of battery charge in the pool is estimated by establishing a parametric model combined with the extended Kalman filter algorithm, so as to obtain various safety indicators and nonlinear parameters of the battery of electric vehicles, and real-time monitoring and timely adjustment of the use state and running state of the new energy battery, so as to improve the efficiency of the battery and extend the life of the battery. Aiming at the problem of fault diagnosis and safety regulation during battery use, this paper adopts the combination of multiple evaluation indicators, and realizes the termination and fault adjustment of the operation process by setting parameter threshold and interrupt procedure, so as to ensure the safety and stability of the operation process. In this paper, the upper computer system is designed to complete high-speed interaction and data storage of various parameters of external modules and battery fault diagnosis through NB-IoT cloud, so as to realize accurate supervision of batteries in the new-energy vehicle group.

Aiming at the problem that the life of electrolytic capacitor is short, which restricts the life of LED drive power supply, this paper studies the LED drive power supply without electrolytic capacitor. Based on the analysis and research of the existing LED driving power supply, the method of organically combining the voltage stabilizing circuit of the non-polar capacitor with the linear control driving circuit of the partial voltage rectification is adopted. The non-polar capacitor is used to replace the electrolytic capacitor, which reduces the magnetic and thermal losses of the power supply, extends the switching frequency life of the LED lamp, and provides a new design method for the popularization and wide application of the LED driving power supply.

10kV cable joints often cause cable failures during operation due to cable breakage, limitations of traditional connection methods, quality problems at construction site, etc.Which often lead to potential safety hazards, power outages for users, threats to the stable operation of the power grid, serious economic losses and other negative effects. This paper presents a design scheme of plug-in intermediate joint. The prefabricated intermediate joint designed in this scheme is simple in structure, convenient to carry, easy to operate and relatively easy to install, which reduces the operating difficulty of the construction personnel, and can improve the reliability of the intermediate joint to a certain extent. In this paper, a new type of 10kV prefabricated intermediate joint is designed. Two cables are inserted into the intermediate joint part and the plug-in part of the device respectively. The two cables are connected and connected through the copper tube guide inside the intermediate joint part, and then the intermediate joint part and the plug-in part are fixed. This scheme accelerates the recovery speed, saves manpower, shortens the operation time, greatly improves the work efficiency, greatly reduces the failure rate, and creates good conditions for the safe and stable operation of the medium and low voltage power grid.

In view of the stability and reliability of power system operation, it is necessary to adopt accurate short-time load forecasting. In order to improve the forecasting accuracy, a very short time power load forecasting method based on EEMD-LSSVM- integrated BPNN is proposed. Firstly, EEMD (Ensemble Empirical Mode Decomposition) is used to decompose the initial power sequence into a plurality of components with different frequencies, calculate the SE (Sample Entropy) of each component, and recombine the similar components with entropy values. LSSVM (Least Squares Support Vector Machine) is used to predict the low-frequency trend components, and integrated BPNN (Back Propagation Neural Network) based on Bagging integrated learning algorithm is used to predict the remaining components. Finally, the distributed power series are comprehensively predicted. The example shows that this method is a feasible short-term power load forecasting method and can effectively improve the forecasting accuracy and stability.

The Integrated Energy System (IES) contributes to the target of "double carbon". In the wind-electricity-gas energy optimization, a scheduling model is proposed in this paper considering the minimum sum of the three costs of energy purchased, carbon transaction and abandoned wind as the objective function with the ladder carbon trading. Then compare the results of scheduling in scenarios which set up differently and analyze the data of carbon emissions and different type cost in scenarios which is set up four. Finally, the environmental protection coefficient ‘a’ and the economic coefficient ‘b’ of the four scenarios are compared. It is concluded that the IES with P2G equipment under the stepped carbon trading in scenario four achieve the optimal balance on ‘a’ and ‘b’, which verifies the superiority of the model.

With the development of power system automation, the way of replacing manual line inspection with unmanned equipment is rising day by day. Inductive energy extraction technology is installed on the low-voltage side of the tower, and energy is extracted by the changing electric field near the inductive transmission line. When the line works, the electric field is stable and the power supply is relatively stable, so it is a good choice for online monitoring system with low power consumption and periodic operation. The main formation mechanism of induced overvoltage in distribution lines is that discharge channels will be formed in the process of discharge, and electromagnetic fields will be formed in the surrounding space, which will lead to induced overvoltage in distribution lines with surrounding power grids. Therefore, based on pulse power technology, this paper further studies the electric field induced energy extraction of distribution lines, and the simulation results show that the simulation results of finite element software are basically consistent with the experimental values, no matter whether the plate size or voltage changes, so we can know that the simulation results of the method proposed in this paper are desirable. However, the initial permeability is low, which is suitable for energy extraction at high current. Permalloy has a high initial permeability, but it is easy to saturate when the current is slightly larger.

The function of the connector on the mountain bike is to connect the two sides to ensure the stability in the driving process but wear often occurs in daily use. First of all, this paper carries on the lightweight treatment to the bicycle connector, after the topology optimization, has realized the weight reduction 54%. Then carbon fiber modified polyester composites were used to improve the materials. The optimum aspect ratio of carbon fiber and the proportion of carbon fiber were determined through a series of experiments, and then its mechanical properties, thermophysical properties and friction properties were analyzed. It is proved that the material has the advantages of light weight, good thermophysical properties and strong mechanical properties compared with steel structure. The connectors of different materials are explored by finite element analysis. Through the finite element analysis and comparison of the stress of different materials, the strength of carbon fiber modified polyester composite is obviously higher than that of medium carbon steel, the deformation of carbon fiber modified polyester composite is less than that of medium carbon steel, the service life of the connector is increased, and the safety hidden danger in the process of riding is reduced.

PWM frequency converter is one of the important tools in modern industry, which plays an important role in solving the production and application problems. However, the high voltage common mode voltage generated by it has certain restrictions on the use of PWM frequency converter. Based on this, this paper proposes the research on the output common mode voltage suppression technology of PWM frequency converter. First, the SPWM modulation uses the equivalent signal to replace the sine wave signal, and then the common mode voltage generated by SPWM is analyzed for data, explain the causes of the common mode voltage from the theoretical data, clarify the characteristics of the PWM inverter, analyze the influencing factors of the common mode voltage, and design the suppression method according to the influencing factors. Finally, the feasibility and correctness of the suppression method are verified through simulation experiments. In order to provide a new idea and method for the related research of the output common mode voltage suppression technology of PWM inverter.

With the advancement of the global dual carbon economy, the construction of a new type of power system is accelerating the digital and intelligent development of power grids. Non-Technical Losses (NTLs) in smart power systems are usually related to the usage behavior of the customer. Amongst other things, electricity theft is a serious threat to electricity suppliers and can lead to serious financial losses and wastage of electricity resources. While electricity consumption information on the electricity customer side allows for analysis of the electricity consumption situation, the issue of customer electricity consumption anomalies is a challenge. Traditional abnormality detection methods are often difficult to apply to NTL detection and have low detection rates. To address this issue, we suggest an innovative model of multiple self-attention. It connects the self-attentive to the dilated convolution and is unified by a convolution kernel of size one. Designed to solve the problem of detecting electricity theft on unbalanced data sets. This method is able to accurately diagnose NTL anomalies and provides more comprehensive and accurate data support for big data analysis in the power industry, helping to improve the efficiency and quality of service in the power industry.

Power equipment knowledge base is an important part of the digital transformation of power grid equipment management, but the structure of power equipment knowledge is complex, and its operation and maintenance involve equipment technical standards, general systems, product information, equipment encyclopedia and so on. In order to improve the utilization rate of data and mine the related information between devices, it needs to be processed and stored uniformly. In view of the above content, this paper introduces a construction and query method of power equipment knowledge map based on graph database. With the help of knowledge map technology, the text data of power equipment is processed, and the query method of power equipment knowledge is specified to complete the construction and query of power device knowledge map based on graph database. It realizes the standardized storage of power equipment knowledge, greatly improves the operation performance of complex associated data, and realizes the rapid query of power equipment technical standards, general systems, product information, equipment encyclopedia and other information. Experiments show that the model has a better recognition effect for entities with more times and stronger regularity, and achieves a more efficient and convenient effect in practical application.

Analyzed the main process, influencing factors, and key links of battlefield material supply support. Established an optimization model for material supply support task planning. Aiming at the characteristics of the model with multiple algorithm constraints, large decision-making planning scale, and multiple variables, proposed a strategy of step-by-step solution and gradual optimization of calculation, solved the problem of easy establishment of battlefield material supply support model but difficult solution. Provide theoretical and model algorithm basis for effectively improving the efficiency of wartime material supply and support and military economic benefits. Provide task planning model algorithm support for the construction of wartime logistics support information intelligent system.

According to satellite structure with slender rods or thin-walled, drastic temperature changes on orbit cause structure vibration, and further thermal vibration and even thermal induced bending vibration. The paper describes a thermal machine two-way coupling analysis developed based on existing commercial software. A flexible structure was analysis using the method. The analysis results indicate the amount of displacement of the flexible structure on orbit changes over time providing a reference for the follow-up work. The thermal machine two-way coupling analysis proposed in the paper provides a new idea for thermal induced bending vibration analysis of the flexible structure of satellites.

For the seven hydropower stations with different installed capacities and substations with different construction scales built in different periods under the power system of Yunnan Agricultural Reclamation Power Co., Ltd., the technical standards of different equipment are not fully compatible with the centralized control system, forming several information silos, which has failed to achieve the power generation and supply balance monitoring of the power system of Yunnan Agricultural Reclamation Power Co., Ltd. This article uses cloud computing and 4G communication technology, combined with deep learning algorithms, to construct an online monitoring system for electric energy balance that is suitable for the production and operation of the current power system. Practice has shown that this system can improve the overall monitoring level of the power system, effectively reduce power dispatch and technical management costs, and provide reference for the construction of a new regional power grid in the next stage.

According to incomplete statistics, from 2001 to 2018, the number of Chinese civil aviation passengers has maintained an annual growth rate of 15% to 17.4%, and the national flight volume has increased by about 8% year-on-year. The rapid growth of air travel has accelerated the development of the civil aviation industry. Aviation safety has always been a focus of attention in the industry. With the continuous development of China's aviation industry, the aviation manufacturing industry in China is also growing, and the business volume of aircraft manufacturing and aircraft test flights is also rapidly increasing. How to ensure the safety of civil aircraft test flights is becoming a focus of active research in the industry. Safety is the bottom line and cornerstone of flight testing. When safety issues arise, they often lead to unimaginable disasters, and even lead to the failure of the entire model development. The test flight safety management system is an important means to ensure test flight safety. By systematically analyzing the various elements of the test flight safety management system, a test flight safety management capability evaluation model is established based on the matter element extension theory. The test flight safety management capability of the test flight unit is evaluated, and improvement strategies are sought, providing reference for airlines and test flight units to continuously improve their safety management level.

The current design concepts and systems of aircraft have developed exceptionally complex and precise, various newly developed or improved aircraft models need to undergo several years of rigorous flight test validation before being put into practical application. Therefore, a large number of flight test data are generated during the flight test. Effectively managing and analyzing the data obtained during flight tests to ensure safety is the key to the entire flight test. Based on the C/S architecture of web technology, an attempt is made to establish a flight test database system, which is divided into client software and service support subsystems. Using C/S structure can not only enhance the security of the system, but also provide users with a powerful system management interface and rich database management means, which makes the maintenance of the system simpler and more reliable. To improve the query efficiency of the database, Spark tool is used to establish query connections for multi-channel spatial datasets using iterative methods. The system test results show that the actual rate of file upload and download can reach 60 -70 Mb/s per second, which can meet the performance requirements of flight test organization for file upload and download. This database can provide reference for flight test units and civil aviation enterprises, providing strong technical support for flight test safety and civil aviation safety.

In recent years, with the development of Deep Learning (DL) technology and the continuous improvement of algorithms, DL-assisted diagnosis systems based on medical images have rapidly developed. Compared with traditional image processing methods, DL trains models by combining a large amount of relevant data (e.g. clinical data, imaging data, etc.) and then uses the models to predict disease-related information. Compared with traditional medical image processing algorithms, DL has better performance in recognition, segmentation and classification of medical images. In this paper, an AI-aided diagnosis algorithm is developed for lung cancer, a malignant tumour disease, based on clinical chest CT data and imaging data. The algorithm uses chest CT images as the object of study, and classifies and evaluates patients based on various factors such as their age, gender, tumour volume and location, as well as their knowledge of the disease. A pre-trained model was first used to establish an algorithm for lung cancer tumour segmentation and recognition. Convolutional NNs are then applied to learn solutions to the feature extraction and classification problems. Finally, the results obtained are used as model output and the system performance is evaluated to complete the diagnostic process for lung cancer as a class of malignant tumour disease.

The bullet point acoustic detection and location system determines the location of the attack target by receiving and measuring the muzzle shock wave of the attack target and the shock wave generated by the projectile flight. This paper designs an attack target location method based on the quaternion detection array through the analysis of the ballistic wave signal, that is, the location of the attack target is preliminarily determined by the time difference of the muzzle shock wave reaching each sensor, Then, the ballistic trajectory is determined by the time difference between the ballistic shock wave generated during the flight of the projectile and each sensor, and the positioning accuracy is improved through the combination of the two.

With the continuous development and progress of automobile design technology, most of the main engine manufacturers no longer develop new models through large-scale reverse mature models. However, the independent research and development and production capacity of automobile manufacturers are still obviously insufficient. Automobiles designed and produced independently are far behind foreign automobiles in terms of first failure mileage, average failure interval mileage, average life, reliable life and rated life. Data mining draws lessons from the knowledge of statistics, machine learning, database and other fields in its methods. Data mining was first carried out by experts in the field of database, and getting up early in the database is a very difficult professional technology to master. Based on the big data theory, this paper establishes a performance target evaluation model, and uses this model to set the performance target of mass production vehicles to improve their reliability. In this way, in the early development stage of the mass production vehicle, the performance target of the mass production vehicle can be formulated under the consideration of the market, which can not only reflect the market competitiveness of the quality, but also improve the reliability of the performance target. The results show that RF is more suitable for this experiment than DT and KNN models and has better classification performance on data sets.

Blockchain technology is a paradigm of distributed systems, and its consistency problem is solved by consensus algorithms. This paper summarizes the existing consensus algorithms and divides them into three types: proof, voting, and Paxos-like. The implementation methods and advantages and disadvantages of some representative algorithms among the three types of consensus algorithms are introduced. Finally, the research hotspots and development directions of consensus algorithms in terms of performance and scalability are pointed out.

In the face of massive book resources, users need to spend a lot of time and energy to obtain the required book resources, while the traditional search query can no longer meet the needs of users. In the process of informatization, the library management system has accumulated a large amount of data. There is a lot of valuable information among these data, which can provide decision support for decision makers of enterprises and departments. Users' demand for resources is becoming more and more detailed and precise, and it is increasingly difficult for users to quickly and accurately obtain the resources they need. Based on this, this article puts forward a personalized recommendation model of library based on Collaborative Filtering (CF) algorithm, which can mine the hidden information behind the data, predict the information needs of readers, organize more and better targeted high-quality information for readers, and provide more personalized services. The experimental results show that the personalized recommendation system can improve the recommendation accuracy and realize dynamic personalized recommendation, which provides a theoretical basis for the construction of university libraries.

In the Internet era, Internet technology is widely involved in teaching life, and the use of mobile phone software to learn English has become an important way for many non-native learners to learn English. There is still a great demand for scholars at different stages to learn spoken English. There is a lack of methods and ways to practice oral English for university students, while offline oral English teaching tests still have many limitations. Using AI (Artificial Intelligence) to assist oral English testing, especially speech recognition technology is a new direction of oral English learning. EAP TALK is an AI oral English evaluation system based on statistical calculation, big data, language cognition and deep learning, and automatic real-time scoring. This paper takes "questionnaire Star" as the platform to design an electronic questionnaire, uses SPSS (V26.0) for statistical analysis, and conducts semi-structured interviews with students. The results show that participants have confidence in using EAP TALK for oral practice, and EAP TALK has some advantages over classroom teaching.

Aiming at the problems existing in the operation and maintenance of distribution network at present, this paper studies the design of remote monitoring system of environmental variables in distribution room based on FNN (Fuzzy Neural Network). A remote monitoring system of environmental variables in intelligent power distribution room is established. Aiming at the fault problem of distribution system, a fault diagnosis method of low-voltage distribution system based on FNN is proposed. The combination of fuzzy reasoning and BPNN (BP Neural Network) is adopted, that is, the output of fuzzy system is used as the input of neural network. First, the collected fault signal is processed by fuzzy set theory, then the fault signal is diagnosed by three-layer BPNN, and the adaptive learning rate is used for learning. The research results show that the diagnosis accuracy of the method based on fuzzy reasoning and BPNN is obviously improved compared with the method based on neural network alone. The diagnosis error is greatly reduced, the situation of misjudgment and missed judgment is basically eliminated, and the overall performance of the system is improved. The research results show that the system is reliable and can meet a wider range of needs.

With the development of society and the improvement of people's living standards, tourism is becoming a new way of life. Anyone can take a trip as soon as they say, and the era of national tourism and mass tourism has arrived. Moreover, with the rapid popularization of information technology, when making travel plans, users will generally give priority to searching relevant travel information through search platforms. Personalized travel intelligent push service is mainly based on the comprehensive analysis of user basic information, behavior preference information and scene feature elements to build an interest model for each user, and then push the corresponding personalized service. This paper proposes a smart tourism information push system based on the cloud platform, and studies and solves some key issues involved in the system. In view of the massive data processing problem in the tourism system, the massive data of tourism information is processed through the powerful computing power of cloud computing.

Industry 4.0 has a huge impact on discrete manufacturing industry. The fierce market competition has expanded to the global scope, which also brings a series of unpredictable "turbulence" factors such as order diversity and market demand changes. Manufacturing enterprises can not only rely on product cost and quality to win, but also need to have the ability to meet the needs of users with the fastest response speed under the unpredictable environment, that is, the flexibility of the system. Based on the development trend of manufacturing system, this paper tries to establish a new method for quantitative evaluation of manufacturing system flexibility.

Escalators are becoming more and more popular in daily life, bringing a lot of convenience to people. However, in recent years, the frequent safety accidents of escalators have caused people's concern. How to ensure the safe operation of escalators, timely diagnosis and early warning of faults has become a focus of research for many elevator manufacturers and operation and maintenance personnel. The traditional regular inspection of escalators causes waste of human and material resources. The rapid development of Internet of Things technology and sensor technology has brought new development direction for the operation condition monitoring of escalator parts. Under this background, the escalator condition monitoring and early warning system is designed and constructed, and the results are displayed visually through the user interaction page.

The subsystems in a complex system are not independent of each other, but interrelated, and because they are in different micro-level and macro-level, this makes the complex system hierarchical. In the process of rapid economic development, the integration of economics and management is deepening. Applying the statistical analysis method of big data to regional economic management can find out the rules of management and operation, improve the efficiency of management and promote the healthy development of regional economy. This article discusses the application of big data analysis method in the field of regional economic management and puts forward an intelligent decision management model of regional economy based on AFSA (Artificial Fish Swarm Algorithm) driven by big data. The simulation results show that the stability of the model system is high, and it can still keep about 90% stability when there are many transaction sets. The results show that the method is feasible and available, and it has good performance. There are a large number of levels in the regional economic system, and their hierarchical structures are different. In order to realize the overall optimization of the system, we must actively coordinate the mutual interest relationship between elements and levels.

The power intelligent IOT system is an important support to guarantee data analysis and intelligent decision making of power system, which realizes power information communication and resource business scheduling through edge-side and cloud master station architecture. With the deployment of massive smart terminal devices leading to problems such as insufficient channel communication resources and delayed middle station service scheduling, this brings obstacles to resource scheduling for the intelligent operation and maintenance of power grid. To address the above problems, this paper proposes a quality management optimization method for the resource scheduling strategy of smart IOT system. The method uses cloud-edge collaboration technology to analyze the existing smart IOT system, constructs a cloud-edge collaborative resource quality management architecture, and studies the collaborative resource interaction method between the cloud and the edge. A multi-dimensional quality optimization model is established based on the edge computing terminal container to realize the optimal management of resource quality. The experimental simulation results prove that the method can efficiently solve complex computing tasks and reduce the data communication frequency delay between the edge side and the cloud center by up to 15%. Its overall improvement of the quality of resource scheduling in the smart IOT system provides a technical method for intelligent assessment of grid quality.

The transmission process of automobile monitoring video network will be disturbed by complex network environment, such as network delay, jitter and other factors, which will lead to the phenomenon of multimedia asynchronization, and audio-video synchronization, as one of the key technologies, has attracted more and more attention. In this paper, an audio-video synchronization algorithm of automobile monitoring video based on neural network model is proposed. By improving the variable-length coding algorithm and its mapping rules, the audio information is dynamically grouped, and the dynamic double mapping relationship between prediction mode group and variable-length code group is established. The simulation results show that the algorithm can ensure the correctness and integrity of audio data, and realize the synchronous coding transmission of high-definition and ultra-high-definition video. The algorithm is applied to vehicle monitoring video, which ensures the time constraint relationship between audio data and video data, that is, realizes the synchronous playback of audio and video, thus effectively ensuring the driver's driving safety, avoiding traffic accidents and making it easier for the driver to drive the vehicle.

The theory and design method of campus traffic planning is an important part of the overall campus planning. At present, the planning and design of university campus roads refer to the planning and design of urban roads to a great extent, but there are some differences between the urban traffic organization and that of university campus. The intelligent transportation system needs the traffic state prediction of the road network, but most of the current traffic state prediction methods are single-section prediction methods, which seldom consider the relationship between adjacent sections. Only the simple sum of traffic conditions of each section cannot reflect the spatial distribution of road network sections. In this paper, through the analysis and research of road planning and design and road traffic organization of university campus, an intelligent planning model of university campus road based on Support Vector Machine (SVM) algorithm is proposed. The experimental results show that the errors of this algorithm are small, and the accuracy rate is high, reaching 92.84%, which is about 18% higher than the traditional algorithm. Therefore, this method can build a good campus humanistic environment and natural environment, and effectively reduce the interference of motor vehicles on students.

Moving target detection and tracking is the core of computer vision system research, and it has always been a hot topic for scientists and engineers. Many remarkable achievements have been made in this field, and many preliminary applications have been made. With the introduction and development of the concept of open architecture CNC system, when designing a new generation CNC system, we should not only consider the standard questions of system real-time, machining accuracy, simplicity of user operation, fault diagnosis and system security protection, but also consider the new problems brought about by the new requirements of system software upgrade and maintenance, cross-platform transplantation and user's secondary application development on the system platform. Based on the self-developed DSP type motion controller, this paper designs and develops a special motion control software for single-needle computerized quilting. The system creates an application environment with Genetic Algorithm (GA) to deal with complex problems on the TMS320C6201 hardware platform and performs functional verification on image matching problems with large data volume and complex fitness function calculation in the system environment. The experimental results show that the overall scheme of the system is feasible.

In the field of sports analysis, combining sports action recognition with sports events and analyzing the athletes' sports situation in the video can provide athletes with auxiliary training, find their strengths and weaknesses, optimize the training system and promote their growth. Aiming at the low accuracy of sports video action recognition method, this paper applies the Deep Learning (DL) technology in Artificial Intelligence (AI) to sports action recognition, constructs the functional architecture of sports application system, and conducts comparative experiments on standard sports video action data sets to verify the feasibility and effectiveness of the proposed algorithm from the recognition accuracy and response time respectively. The test results show that the proposed method is superior to the traditional motion recognition model in accuracy and efficiency.

A new computing paradigm that extends the capabilities of the cloud to the edge of the network, Vehicular Edge Computing (VEC), provides vehicles with a wealth of resources that can accommodate the rising demand for in-vehicle application computing and storage resources. This is accomplished by seamlessly fusing Mobile Edge Computing with in- vehicle networks (MEC). It provides low-latency, high-bandwidth, and extremely dependable computing services to vehicles and other users with limited computing resources. This paper reviews existing game theory-based task offloading mechanisms in the vehicle edge environment. The first section introduces the current research status of vehicle edge computing in terms of its features, architecture, and applications. It covers the background and significance of task offloading in vehicle edge computing, the VEC architecture, the advantages of VEC, and several alluring application scenarios. A thorough analysis of the body of prior research is then presented, followed by a categorical discussion of game theory-based computational offloading mechanisms in the context of vehicle edges. Finally, predictions for the future of technology are made.

More digital noise monitoring systems are put into use with the attention to environmental noise monitoring. In order to solve the problem of accuracy and traceability of acoustic performance of environmental noise monitoring system, it is necessary to design an intelligent online noise detection device based on Internet technology and develop technical research on acoustic performance of online noise monitoring system. The device provides a method for realizing the online automatic detection for acoustic performance of digital noise monitoring system and improving the automatic detection ability of environmental noise monitoring system. The experimental results show that the device meets the requirement of online detection of environmental noise monitoring system, makes up for the shortage of online detection ability of environmental noise monitoring system with better applications.

Motion control of mobile robot is the premise of completing all tasks, and trajectory tracking control, as an extremely important part of motion control, is the basis for robot to complete all tasks. For different fields, the problem background and application scenarios faced by robots are different, but in any application scenario, it is necessary to ensure that the mobile robot can track a given trajectory or target quickly and stably in the process of motion control, so as to realize real-time trajectory tracking control. The path-finding method introduced in this paper is to optimize the traditional Back Propagation Neural Network (BPNN) from the perspective of structure and weight setting under the environment model established by grid method, and then, according to the characteristics of high-speed parallel operation of neural network, an automatic path-finding algorithm based on logical judgment and memory is proposed, which can achieve ideal results by combining it with the improved BPNN. The simulation results show that the algorithm can effectively overcome the influence of uncertain factors such as model error and external disturbance, and its trajectory tracking effect is better than the traditional trajectory tracking control algorithm.

In order to solve the problems of large amount of calculation and a large number of false points in the case of target source in the multi sensor direction finding cross positioning system, an AOA positioning algorithm based on three-dimensional space is proposed, which introduces the sensor height information that is usually ignored in the traditional two-dimensional positioning algorithm into the algorithm. The method only needs two measurement platforms. Computer simulation results show that the algorithm has high positioning accuracy and less computation.

With the rapid growth of social productive forces, changes in market supply and demand structure and increasing market competition, the standard for judging products has gradually changed from the previous product performance to the customer's satisfaction with products. According to the investigation and analysis of customer satisfaction, it can help automobile enterprises to allocate resources reasonably, grasp consumers' purchase intention, tap potential consumers and improve sales quantity and quality. In this paper, an algorithm for evaluating the customer satisfaction of commercial vehicles based on fuzzy comprehensive assessment and Artificial Neural Network (ANN) is proposed. Firstly, the assessment system of customer satisfaction is qualitatively described by linguistic variables, and experts score according to the importance of each assessment factor, and then the quantitative assessment results are counted, and then the membership degree theory of fuzzy mathematics is applied to evaluate and identify the customer satisfaction of commercial vehicles. The results show that the error of the proposed commercial vehicle customer satisfaction assessment algorithm is small, and the accuracy is excellent, which can reach more than 96%. This proves that ANN and fuzzy comprehensive assessment have a good effect on evaluating customer satisfaction of automobile enterprises.

How to build a collaborative mechanism of scientific and technological innovation and development in Guangdong-Hong Kong-Macao Greater Bay Area (abbreviation: the Greater Bay Area (GBA)) is undoubtedly an important link in building a world-class scientific and technological innovation center. Based on the panel data of 11 cities in GBA from 2012 to 2020, this paper empirically tests the influence of GBA financial agglomeration on the technological innovation efficiency of high-tech enterprises by using SDM (Spatial Dubin Model) model, starting from the innovative effect of financial agglomeration and considering two weight matrices of geographical distance and economic distance. The research results show that the regression results of GBA's SDM model show that the autocorrelation coefficients under zero to one adjacency weight matrix and geo-economic distance weight matrix both exceed the significance levels of 1% and 10%, which shows that there is obvious correlation between the technical development efficiency of GBA's high-tech industries, but they are not independent of each other. The level of economic development in different regions and stages is quite different. The results show that, at the significant level of 1%, financial agglomeration has a greater role in promoting the efficiency of technological innovation.

With the rapid development of the Internet, the number of digital music has increased dramatically. It is very difficult for users to dig out the songs they are interested in from the massive digital music information, and the problem of "information overload" has emerged. A search engine is a tool for retrieving information. When a user submits a query to the search engine, the relevant content is returned to the user after retrieval. The search results of different search engines are different. Through the analysis of users' behaviors, we can understand users' search habits and realize a search engine that is more in line with users' needs. In order to accurately recommend interesting music to users and improve the market share of application objects, this paper designs a music intelligent recommendation system based on user interest preference, which is composed of data acquisition module, offline data processing module and online recommendation module. The data acquisition module collects user registration information, online scoring data of target music, and conducts online survey to collect relevant data. Data analysis and experiments show that the model is superior to the comparison model in terms of training efficiency and accuracy of recommendation, thus verifying the effectiveness of integrating user's long-term and short-term preferences and user influence for recommendation.

By collecting the listener's response to the generated music word and song matching degree, such as heartbeat, pulse and skin conductance, such as evaluation score, preference or physiological signal. Then use the collected information to evaluate the generated music works. How to evaluate the performance of many algorithms is a difficult problem, not only because of the complexity of music data itself, but also because the query requirements of users in different application environments are very different. For a piece of music, rhythm is very important. In this regard, this paper attempts to design an intelligent evaluation algorithm for music word and song matching based on NN (Neural Network), and generate TopN recommendations for target users by calculating the similarity between user preference features and music potential features. At the end of the simulation experiment, it is further proved that the algorithm proposed in this paper is convergent in the iterative process, and the accuracy in the training set has reached 93.4%. When the lyrics are looked at alone, the lyrics reflect the sad emotion. But the lyrics combined with the melody can be found that the rhythm of the song is very strong. Through the experiment, the accuracy in the test set has reached 86.2%, which shows that the accuracy of the algorithm proposed in this paper is considerable.

Improving the effect of music annotation task through various advanced technologies has become a hot direction in the field of music information retrieval. The research of music automatic annotation has high value in both theoretical research and practical application. The deep CNN (Convolutional Neural Network) model in the field of deep learning has achieved good results in the fields of image and voice. In this paper, the construction of multimodal music automatic labeling model based on neural network algorithm is launched. In this paper, CNN combined with SAM (Self-Attention Mechanism) is used to learn the appropriate feature representation from the low-level Mel spectrum description of music and the original audio waveform data. Two-dimensional convolution is applied to the Mel spectrum input of music, and one-dimensional convolution is applied to the original audio waveform input, so as to better capture various structural features of music and annotate it. The results show that the accuracy of CNN combined with SAM method is 3.9% higher than that of linear weighted fusion method, and the AUC value is 8.5% higher than that of linear weighted fusion method. The comparison results show that the multi-modal music automatic annotation model framework proposed by CNN and SAM in this paper is effective for the automatic annotation task of music.

With the development of IT (Information Technology), the frequency of digital management is increasing. For music audio-visual archives, the traditional management mode is no longer suitable for the development of the times. Realizing digital management can better improve the management efficiency of music audio-visual materials and improve the storage safety and utilization efficiency of music audio-visual materials. Therefore, on the basis of analyzing the structure and characteristics of the digital management system of audio-visual archives, this paper constructs a music audio-visual archives resource management system based on RF (Random Forest) algorithm. RF algorithm is a statistical learning theory that includes multiple decision tree classifiers. It uses feature subspace to construct the model, which can better deal with noise and avoid over-fitting. Based on the construction process of decision forest, this paper puts forward an improved RF method. The results show that the error of the algorithm is low, and the accuracy is high, which can reach 95.47%. At the same time, the proposed method can intelligently recommend relevant music audio-visual archives resources to users according to their needs, which further improves the management efficiency and utilization efficiency of music audio-visual materials.

Traditional security screening methods mainly use manual identification of security images, there is a low identification of inefficiency, high error of judgement rate, which has become a bottleneck limiting public safety and security. Therefore, to deal with this problem, this paper proposes the security image contraband detection model PP-YOLOE+_DCS, which makes three main improvements on the basis of the PP-YOLOE+ model. To begin with, we introduced deformable convolution within the backbone network that strengthen the models' feature extraction capability. Secondly, we introduced a coordinated attention mechanism among the backbone network and the detection neck for better focusing the model on the object region. Finally, we replaced the original GIOU loss function with the SIOU loss function to improve the detection accuracy and training speed. The improved PP_YOLOE+_DCS model obtained achieves 91.4% detection accuracy, 2.8% improvement compared with the baseline model mAP, only 0.24 M additional parameters, and 420.2 ms inference delay on embedded devices, which provides a new solution for the intelligence of contraband detection.

Sleeping is a vital biological state which help maintaining the homeostasis of organisms of all biological lives. A full sleep can be divided into different repeating stages, Rapid Eye Movement sleep (REM) stage, Non-Rapid Eye Movement Sleep (NREM) stage one to four. An effective sleep staging system can help patients improving their sleep quality. In the past, patients are required to wear Polysomnogram (PSG) for the whole night to collect signals like Electroencephalogram (EEG), Electrooculogram (EOG), Electromyogram (EMG) for diagnosis. And the traditional sleep staging system use one or more signals above to predict a sleep stage. In this paper, we introduce a new sleep staging algorithm based on machine learning. Our model has two main inputs: patients’ respiratory signal and their physical data, like age, gender, and weight. The strategy is to use two CNNs to extract features from raw respiratory signal in time domain and frequency domain, several Word2vec layers are built to extract features from patients’ meta data and a transformer encoder to collect all the features. Using the MIT-BIH Polysomnographic database, our model achieves a result of 81.96% accuracy. This shows that it is completely feasible to classify patients’ sleep stage with their respiratory signal and meta information.

With the popularity of the Internet and mobile terminals, the amount of movie entertainment information on the Internet has increased dramatically, and users' demand for personalized movie services is growing. Personalized recommendation service can be used to mine the hidden correlation between user's historical information, movie project information, user's historical operation log and data, and recommend the obtained video resources that users may be interested in to users, so as to better serve users. Among them, video recommendation is an important field of recommendation system technology research. The existing film and television recommendations are mainly popular recommendations and related recommendations. Popular recommendations are easy to lead to Matthew effect, while related recommendations are in line with users' preferences to a certain extent, but the degree of personalization is low. Different users often see the same recommendation list on the same play page. This article first analyzes user behavior data in the user behavior data modeling phase, and establishes an initial user behavior data model by combining user scoring behavior and labeling behavior. Finally, based on the improved recommendation algorithm proposed in the previous article, a video recommendation system is designed and implemented. First, the system requirements are analyzed, and then relevant designs are carried out according to the requirements. The system is implemented using the SS2H framework, and the main data table display and functional interface display of the system are given.

In the 21st century, with the rapid development of Artificial Intelligence (AI) technology, rapid and automatic recognition of ground objects in remote sensing images has gradually become the current research focus. This technology solves the problem that a single remote sensing image cannot cover the target area and cannot meet the actual operation requirements. Because of the advantages of high efficiency, low cost and reliable mosaic image quality, it is widely used in engineering surveying and mapping, traffic monitoring, military investigation, disaster investigation and other fields. With the rapid development of aerial remote sensing technology, the number of optical remote sensing images has increased explosively, which provides necessary data support for using deep learning method to detect targets in remote sensing images. Based on the research background of automatic detection and recognition of bridge targets in visible light remote sensing images, this paper studies the problems of feature extraction, river detection, river region correction and bridge target recognition, and constructs a complete automatic detection and location system of bridge targets. After that, the bridge is accurately detected using binary morphology combined with the feature of bridge crossing both banks.

At present, cable-stayed bridges and suspension bridges are widely used in highway traffic construction in China because of their beautiful lines and strong spanning ability. Cable is an important part of cable-stayed bridges and suspension bridges, and it is the key stress component of this kind of bridges. The traditional method of bridge crack detection mainly relies on manual field investigation, which has the disadvantages of high risk, low accuracy and low efficiency. At present, many crack measurement systems based on digital image processing technology are mostly semi-automatic or highly dependent on manual work, with low efficiency. In this paper, the crack image obtained in the field in the engineering test is used to enhance and de-noise the crack from the image preprocessing, including color image to gray image, image enhancement and image wavelet thresholding de-noising. At the same time, the corresponding contrast image of the crack is given. Then, an adaptive multi-scale watershed segmentation algorithm is proposed. First, the filtered image is processed by multi-scale gradient transformation, Then the gradient image is transformed by adaptive marked watershed transform, and finally the obtained image is combined with improved regions to effectively suppress the over-segmentation phenomenon.

The corresponding relationship between magnetic induction intensity and GMM strain was analyzed theoretically. The finite element software COMSOL Multiphysics was used to select and compare the transducer, and the static simulation analysis of the bias magnetic field was carried out. Taking the axial magnetic induction intensity and magnetic field heterogeneity of GMM rod as optimization evaluation indexes, the optimal magnetic circuit design scheme was determined by optimizing the transducer shell material, the relative permeability of magnetic block and the structural parameters of permanent magnet. The two indexes were significantly improved, and the magnetic leakage was reduced, and clear test results were obtained.

Nowadays, with the continuous development of technology, industrial robots are widely used in the manufacturing industry, where tandem robots are working with increasing capabilities. Robot kinematics aims to study in depth the laws of robot motion in a specific frame of reference and as an important function of time, as well as the correlation between the spatial structure of robot joint variables, robot end-effector positioning and pose, and their internal interactions. It takes into account not only the forces and moments of robot motion, but also the temporal characteristics of robot motion. The purpose of this paper is to study the kinematic characteristics of the robot based on the desktop-level six-axis tandem robot GLUON-6L3. First, the DH joint coordinate system of this six-degree-of-freedom robot arm is constructed by the DH parameter method, and the related basic parameters of the linkage are defined. Then, the expression of the chi-square transformation matrix is given, and finally, the kinematic solution results of the robot are verified by using Matlab software to ensure its accuracy.

Intelligent network connected vehicle is an important development direction of cutting-edge science and technology, which has a significant impact on intelligent transportation, modern logistics and other aspects. The environment awareness system is the foundation and guarantee of the safety and intelligence of the intelligent network connected vehicle. The application of intelligent and automatic driving technology to automobiles has become an inevitable trend in the future development of automobile industry. The complexity of the product structure and application technology of Intelligent and Connected Vehicle (ICV) has brought great changes to the industrial ecological chain, which also determines the potential high risk of its Supply Chain (SC). Taking the automobile industry as the research object, this paper identifies the SC risks in the development of new products in the automobile industry. After screening and combining with the characteristics of the automobile industry SC, the SC risks in the development of new products are classified into six categories: supply risk, manufacturing process risk, demand risk, cooperative relationship risk, logistics and distribution risk and external environmental risk. This paper aims to provide theoretical methods and strategic support for the risk management of automobile green SC and promote the sustainable development of automobile green SC in China.

How to accurately determine the location of the impact point is an urgent problem in many application fields. By analyzing the influence of oblique incidence of the projectile on the positioning accuracy, a planar double disk positioning array is proposed. The disk array is used to correct the error caused by oblique incidence of the projectile. The mathematical derivation and error analysis of the model are carried out. The correctness of the model algorithm and its positioning performance are verified through simulation experiments.

The main function of active equipment is to bear weight and provide accurate guidance for the transmission of the whole device. Once the failure of active equipment, the safety and operation stability of the mechanical equipment will be affected. In this paper, a wear detection system based on image recognition technology is developed to solve the wear problem of key parts of dynamic continuously running equipment. Through online video image acquisition, preprocessing, analysis and recognition of wear particles, the wear type and wear severity of key friction pairs of mechanical equipment can be matched and identified, and the equipment fault information and fault development trend can be obtained So as to realize the early warning, diagnosis and maintenance of the key friction pairs of mechanical equipment, so as to prevent and eliminate the failure of active parts affecting the stability and safety of the whole device.

With the growth of the Internet, network attack means emerge in endlessly, and internet safety issues become more and more important. Anti-virus software and firewalls generate a large amount of alarm information. Faced with such a huge amount of data, network managers can hardly grasp the security status of the network system and cannot take corresponding preventive measures in time. In order to improve the data security transmission performance of the network, it is needy to evaluate the anti-risk index of intelligent networking. At present, most of the internet safety assessment methods are from the perspective of security vulnerabilities. By scanning whether there are some known vulnerabilities in the network, the corresponding assessment results and solutions are given. In this paper, an intelligent network anti-risk index assessment system based on improved machine learning algorithm is proposed. Through this system, the internet safety situation is predicted, and then the algorithm model of internet safety assessment decision system is constructed by using assessment rules. The simulation results show that this method has high accuracy in evaluating the anti-risk index of intelligent networking, strong ability in detecting intrusion information, and improve internet safety.

Since the implementation of China's financial budget reform, after more than ten years of continuous exploration, practice and improvement, it has played a positive role in deepening the concept of budget management, refining departmental budget preparation, strengthening the rigidity of budget constraints, and ensuring the performance of the organization's mission. The reform has achieved remarkable results. However, with the continuous development of the economy and society and the in-depth promotion of the construction of the public finance system, especially in the context of the current requirement to open the budget, final accounts and "three public" funds, the current financial budget management in China is still uneven in the implementation of the budget, project funds and public funds, and a large number of year-end carry forward balances due to the insufficient accuracy of the prediction at the beginning of the year and the inability to accurately integrate various data resources Some units spend too much money, and performance evaluation is lack of grasp. These problems, to some extent, hinder the scientific rationality of funding arrangements and the further improvement of data resource allocation efficiency. This paper proposes a method to build a basic database and a cross year rolling project database based on data resource fusion technology to improve the scientific rationality of the budget allocation at the beginning of the fiscal year.

This article proposes an image processing algorithm based on multi-scale clustering, which can improve the probability of a classifier correctly classifying an image even after it has been subjected to noise attacks. The algorithm mainly involves performing multi-scale clustering on the pixels of the image based on features such as color space, distance, texture, etc., and then incorporating the clustering results into the training of the classifier. By directly extracting the main feature information from the image for training, the algorithm prevents large classification errors due to changes in certain pixel values, thus reducing the impact of attacks on other pixels in the image and greatly improving the robustness of the image. Therefore, this algorithm is a robust and efficient denoising method.

Almost 80% of the vehicles required for Bangladesh's road transportation industry are supplied by reconditioned cars. Using Machine Learning (ML) to predict car prices refers to using ML algorithms and techniques to make assumption about future car prices. This can be useful for a variety of purposes, such as helping car buyers and sellers make informed decisions, assisting car dealerships with inventory management, or providing insights for car manufacturers and other industry stakeholders. To predict car prices using ML, data is collected on a variety of factors that can affect the ongoing cost of a car, such as its make and model, age, mileage, condition, and location. This data is then fed into the Random Forest ML model, which uses statistical techniques to analyze the data and identify patterns and trends. The model performs 99.59% accurately in the tested portion of the data set and ensures that the model can then be used to make predictions on the future cost of an automobile based on these patterns and trends.

At present, scientific and technological personnel in electric power enterprises mainly rely on patent search websites to obtain cutting-edge electric patent information. But these websites are mainly based on string matching, which fails to capture the connection between patents, making the recommend result unsatisfying. In view of the above problems, we first constructed a grid patent knowledge map, where knowledge extraction was carried out for entities such as title, abstract and applicant in the patent text, and the entity and defined relational data were stored in the Neo4j graph database. Secondly, the Transe-SNS algorithm with optimized negative sampling was used for vectorization of the graph entity relationship. Experiments showed that Mean Rank and hit@10 improved by 27.5 and 2.2% respectively compared with the traditional TransE algorithm. Finally, the similarity between patents was calculated by combining the results of knowledge graph vector embedding of patent entities with the word vector embedding of patent title abstraction, and top- k patents in similar fields were recommended for users. The experiment proved that the proposed method is superior to the traditional text embedding method in recommending similar patent technology fields.

After the reform and opening up, the prosperity of market economy has promoted the transformation of China's social structure, and thus promoted the transformation of social human resources management mode. Teachers in universities, like members of other social trade organizations, have more opportunities for communication and re-selection of occupations and posts in the market economy. It is bound to become a trend to use a more effective decision support system for teacher management to provide decision support basis for managers. This article puts forward a micro-decision-making mechanism of teachers' mobility based on Data Mining (DM). By using DM method to improve the teachers' management system, the decision-making ability of teachers' management is improved. Through the practice of DM technology in teachers' management, the DM is further explored. The results show that the algorithm in this article has obvious advantages in the middle and late period of operation, and the error is reduced by 30.57%. The algorithm has the advantages of simple description, fast classification speed, suitability for large-scale data processing, etc. It can be applied to classification, prediction models, and purposefully classifying a large number of data.

The environment of the transmission line is complex and easy to attach foreign matter, which has always been one of the reasons for the safety hazards to the operation of the transmission line. In view of the current problem that the detection accuracy of the foreign matter on the transmission line needs to be improved, an improved transmission line foreign matter detection method based on the Yolov5 algorithm is proposed. Detection model. First, the RepVGG module is introduced into the feature extraction network to enhance the network feature extraction ability and improve the model reasoning speed; secondly, the ability of the network to identify important feature information is strengthened by integrating the attention mechanism module; finally, by adding the prediction layer and Soft-nms the algorithm processes the target prediction frame to improve the detection accuracy of the model. The experimental results show that the improved Yolov5 transmission line foreign object detection algorithm proposed in this paper has a mAP value of 4.1% higher than the conventional Yolov5, and it also has certain advantages in performance compared with the conventional target detection algorithm.

Vehicle user information and big data are already inseparable, so rapid data processing and analysis has become a major challenge for the current Internet of Vehicles. It is also imperative to study the collection, utilization, protection and compliance governance of personal data in the vehicle during the development of the auto-drive system. This paper analyzes the collection and utilization, protection and compliance governance of personal data in the vehicle during the development of auto-drive system. This paper makes a comparative analysis of the speed of personal data collection and the security of data. Through Hive data collection survey, this paper analyzes the personal data of 2000 different types of car owners in City A and constructs an experiment from the speed of personal data collection in the car and the degree of personal data protection. The results show that Hive's data collection speed is far faster than that of traditional methods in any aspect. The use of new data protection system can not only reduce the number of external attacks by about 33% through its powerful functionality, but also effectively defend attacks, and the defense success rate is increased by 33%. This is of great significance for the safety protection of personal data in the car during the development of the auto-drive system and will escort the research of personal data in the car during the subsequent development of the auto-drive system.

A two-layer labeling-based dependency analysis model is proposed to introduce the location features of words for the coreference disambiguation of entities. Firstly, two layers of labels are used for labeling, the second layer labels the location information of words, the features of sentences are learned using a bidirectional long short-term memory network, the dependency syntax analysis based on deep graph decoding obtains the dependency tree of sentences, and the two layers of labels are fused to improve the performance and accuracy of coreference disambiguation. Experiments are conducted on the text dataset of power security entity relationship extraction, and the results show that the two-layer labeled dependency analysis has an improved effect on co-finger disambiguation, which verifies the effectiveness of the model for co-finger disambiguation on the experimental dataset.

The rail transportation industry is moving towards heavy loads, high speeds as well as high density operation, which places higher demands on the quality of heavy rail. The quality inspection of rail defects using machine vision first requires obtaining clear surface/planar images of multiple surfaces of the hot heavy rail. Therefore, a vision inspection solution is proposed, which uses a 6-lane linear CCD to photograph the tread surface, bottom surface, and upper and lower waist surfaces of the heavy rail. A comprehensive comparison of various types of light sources was conducted through experiments on chromaticity, brightness, and spectral power. To solve the problem of image overexposure due to infrared radiation of the hot heavy rail, the spectral radiation characteristics of the hot heavy rail were analyzed, and the imaging effects of adding different filters were compared and analyzed. Finally, to solve the problem of out-of-focus due to oscillation of the heavy rail during rolling, the image is analyzed by an automatic focus search strategy, and the results are fed back to the camera to achieve automatic focus, thus adaptively obtaining good image quality and laying a good foundation for further defects detection.

During the growth of universities, carrying out Physical Education (PE) courses can not only promote students' physical and psychological health, but also strengthen students' physique, which plays an important role in students' future development. To realize the effectiveness of PE classroom teaching, we must correctly understand the effectiveness of teaching, design teaching scientifically and reasonably, and use instructional assessment reasonably and objectively. Many factors of educational phenomenon can't be classified into absolutely clear categories, and any specific identity is relative, which contains differences and changes, thus showing certain uncertainty. According to the viewpoint of fuzzy mathematics, many things that people regard as the same are a fuzzy set. Combining with Data Mining (DM), this article constructs an assessment model of PE instructional effectiveness based on constrained clustering algorithm. For PE teaching, the construction of the assessment model of PE instructional effectiveness based on constraint clustering algorithm can further improve the monitoring quality of PE teaching, strengthen the information monitoring level and information summarizing ability of PE teaching, and ensure that the relevant information of PE instructional level can be applied to practice more accurately.

With the development and application of computer technology, the design and production mode of manufacturing industry is gradually moving towards digitization and intelligence. As a key component in space exploration, navigation, and national defense, the assembly quality of array antennas is crucial. In response to the problems of invisibility and difficulty in assembly of array antennas, this paper analyzes the current research status of virtual assembly from four aspects: virtual assembly system, assembly sequence planning, collision detection, and virtual assembly interaction, by introducing virtual reality technology. Finally, future work is pointed out for Industry 4.0. This research can provide a theoretical basis for the virtual assembly of array antennas.

In order to meet the high precision and high efficiency processing requirements of multi-variety small batch aluminum magnesium casing, the construction of flexible manufacturing cell of aluminum magnesium casing was studied. By analyzing the processing characteristics of aluminum magnesium casing and the current situation of manufacturing workshop, the overall design of flexible manufacturing cell was completed, and the model facility layout within the cell was constructed. The Tabu search algorithm was used to solve the model, to further complete the study of the layout of flexible manufacturing cell of aluminum magnesium casing was. The construction of flexible manufacturing cell for aluminum magnesium casing is of great significance to solve the existing problems in the manufacturing process of aluminum magnesium casing and summarize the new processing mode of key parts of aero engine, which lays a solid foundation for the construction of intelligent workshop.

A lot of rich auxiliary information can be integrated into the scoring and show good performance, which plays a good role in many recommended methods. In this paper, we use the rich semantic of knowledge graph to explore user preferences and improve the algorithm performance by using item attributes. A large number of experiments have been carried out on two data sets to prove the validity of the graph convolution recommendation model.

Joint replacement is a frequently performed surgical intervention to address end-stage joint ailments. However, successful surgery relies on early diagnosis and appropriate treatment of joint diseases. With the advancement of machine learning technology and its integration with the medical field, the use of machine learning technology in joint replacement has made substantial advancements. This article will explore the implementation and current standing of machine learning in joint replacement from four aspects: preoperative auxiliary diagnosis of arthritis, preoperative auxiliary decision-making, postoperative complication diagnosis and postoperative prediction.

Traditional over-sampling and under-sampling algorithms suffer from overfitting and high noise when unbalanced data classes are in the sample set. To improve the performance of the data classifier, this study proposes a SMOTECU algorithm combining SMOTE and ClusterCentroids under-sampling. It absorbs the advantages of both algorithms and avoids generating or rejecting excessive samples in the dataset, effectively reducing the harmful effects of overfitting and noise. We experiment with 16 unbalanced standard datasets combining three classifiers: RF, RBFNN, and RBFSVM. By comparing three evaluation metrics: F1-score, AUC, and running time, the results demonstrate that the performance of the SMOTECU-based random forest model is better, and compared with SMOTE and ClusterCentroids, SMOTECU can effectively avoid overfitting and save running time.

In the current market competition, if enterprises want to occupy a favorable position, they must use information technology tools, give full play to their own production and management advantages, improve market competitiveness and resist risks. The emergence of cloud computing brings hope to the current financial management information construction model. In this paper, an accounting big data analysis platform driven by cloud computing technology is established, and an accounting big data cloud computing task scheduling algorithm based on improved FA (Firefly Algorithm) is proposed. Aiming at the problem that traditional FA is prone to premature and fall into local optimum, chaotic disturbance strategy is introduced to disturb the part with low target value, and fireflies flying out of the search space are bounced back according to the rebound theory based on real physics in firefly position update, so as to maintain the optimization ability of the population. The research results show that this algorithm can shorten the total execution time of cloud computing tasks by about 21.06%. It is proved that the improved FA has better optimization ability and real-time performance.

Image-based insect species identification is a comprehensive application of computer vision technology, image processing technology and pattern recognition technology to realize insect species identification. It is of great significance to find and identify crop insects in time and quickly, so as to take corresponding measures to prevent and control pests in time and reduce the losses caused by pests. In this paper, intelligent insect species identification based on computer vision is studied. For insect images, using feature fusion technology will obtain more comprehensive insect image features, which will greatly improve the recognition accuracy and improve the recognition efficiency of the system. In the learning process, firstly, a certain number of insect images are selected as training samples, and the feature vectors of insects are extracted as the input vectors of SVM (Support Vector Machine) training, and the input vectors are learned to get the best classification model. The experimental results show that the average recognition rate of insects by feature fusion is 94.747%. The experimental results show that the research results of this paper provide theoretical and practical basis for intelligent insect species recognition technology based on computer vision.

The experiment is based on plug-in construction design and pattern adaptation design, explore the design methods and processes of cultural and creative patterns with preliminary intelligence. The experimental team designed a plug-in based on Grasshopper by programming, using Grasshopper as the running platform and the Bird-and-Insect as the theme, controlling and automatically generating basic patterns with parameter values. The purpose of the experiment is to reduce the repetitive and heavy workload of pattern designers and improve the cultural attributes of patterns through digitization and personalization. This experiment has proved the feasibility of using parametric technology to pattern the cultural creation of Bird-and-Insect and has been verified by the design and application of related cultural creation products.

To address the negative quality of RT-PCR and the time taken to obtain results, this paper reports on the classification and discrimination of COVID pneumonia images by developing a neural network adaptively. Given the limited number of published COVID-19 CT data currently available, we suggest using the proposed negative process to refine the data to obtain more CT data to reduce matching risk. Experimental results show that compared to network models such as AlexNet and GoogleNet, the proposed BUF-Net network model has the best performance with an accuracy of 93%. Visualization of systemic findings using Grad-CAM techniques may further elucidate the critical role of CT imaging in the diagnosis of COVID-19. The use of deep learning tools in clinical practice can help radiologists make better diagnoses.

The mode of film and television information dissemination has undergone great changes under the promotion of social network information technology. The disseminators of film and television information have changed from single production information in the past to multi-channel release information now, from simple text information to various forms of multimedia network information interaction. This paper attempts to build a film and television information dissemination model based on grey clustering algorithm and sorts out the application in film and television information dissemination according to the time development order and summarizes the development process and mechanism of film and television information dissemination model. The simulation experiment is used to simulate the film and television information dissemination process of information in this network, which proves the practicability of this model. The research results show the changing trend of the number of final nodes in the unobserved state S state with P value under different μ values. When μ=5.5, with the increase of P, that is, the proportion of group effect of group heat information in film and television information network in decision-making gradually increases, the final number of people who have not watched it gradually decreases, and the final spread range of film and television information gradually expands with the increase of the proportion of P value.

Cultural trade is one of the main activities of cross-cultural communication. With globalization, the proportion of cultural trade in international trade has gradually increased, which plays an important role in promoting national economic growth and enhancing cultural influence. Among all kinds of cultural media, film and television media have great international communication ability, and the visual images spread through film and television media are easy to be understood and accepted by people of different countries and cultures, with the widest audience and coverage. Based on SWOT analysis, this paper constructs an information processing model of cross-cultural communication of film and television culture in the context of new media and puts forward a prediction algorithm of social network information communication based on dynamic adaptive network. The results show that the prediction algorithm proposed in this paper performs well on both data sets, and the accuracy can reach 95.88%. Compared with other methods, its prediction accuracy is obviously improved. The proposed social network information dissemination prediction algorithm can more truly reflect the information dissemination law of social networks in reality, improve the efficiency of film and television cultural information processing, and can provide support for the study of cross-cultural communication strategies of film and television.

University students are especially susceptible to the negative effects of mental stress in today's environment, which is a serious issue overall. A great deal of pressure is now being placed on a period of life that was traditionally considered to be the most carefree. People in today's culture are exposed to increasingly high levels of mental stress, which has been related to a broad variety of health problems, such as depression, suicide, heart attacks, and strokes. Because of this, in order to primarily extract, for the purposes of this research, the mental stress ratings of university students, we applied a total of six distinct machine learning methods. The Decision Tree Classifier, the Random Forest Classifier, the SVC, the KNN Classifier, the Multinomial NB, and the K-Nearest Neighbors Regressor are only some of the machine learning algorithms that are available. This investigation's principal objective is to determine the percentage of students who are struggling to deal with emotional pressure in their lives. The dataset was put together by hand with paper and manual information obtained from a survey. Out of the six distinct classification strategies, the Decision Tree Classifier and the Random Forest Classifier both achieved a test result of 0.99, which is the maximum score that can be achieved.

Aiming at the advantages of processing time series data, an excellent variant network seq2seq of recurrent neural network is constructed, and the basic unit of the network adopts LSTM, and a wind turbine fault prediction method based on SCADA data is proposed. The method first reduces the dimensionality of a certain sequence length through the encoder and decoder of SEQ2SEQ, and then predicts the magnitude of the active power through the fully connected layer, and then calculates the residual size between the predicted value of the active power and the actual value to analyze the operating state of the wind turbine, and finally verifies the method with obvious fault data. The results show that this method detects the abnormal occurrence of the wind turbine 6 days earlier than the alarm time of the SCADA system, which provides a technical guarantee to avoid the further deterioration of the wind turbine failure.

It is popular approach to employ Zeroing Neural Networks (ZNN) to handle time-varying problems. However, most of the existing methods have limited flexibility in the discretization process, which seriously affects the final effect of the model. To solve this problem, in this paper, we present a new discrete-time ZNN (DTZNN) for handling discrete time-varying minimization problem. In detail, the new DTZNN is proposed with four-instant general discretization formula, which has a adjustable parameter d. Generally speaking, in the process of discretization, different values of d and time interval can influence the residual error observably. In order to verify the effectiveness of our method, we conduct multiple numerical experiments. The numerical experiment results show that the new DTZNN has effectiveness and superiority for handling discrete time-varying minimization problem.

In this paper, a technical solution for the integration of vehicle-road cooperation and advanced emergency braking system is proposed. Firstly, the problems existing in the practical application of the system are analyzed, and the application scenarios are summarized by multi-level screening method. According to the different functional requirements of application scenarios, the overall system architecture and control strategy algorithm are put

The existing space-based remote sensing has problems such as weak collaboration, slow response, and long links, which cannot meet the application requirements of real-time anomaly detection, recognition, and transmission. This article studies the characteristics of existing surface anomaly classification, establishes a demand matrix for anomaly remote sensing, establishes a new surface anomaly real-time detection system, and proposes a working mode for anomaly real-time recognition, using the constellation system task and information flow design combined with the on-board intelligent processing unit, improve the anomaly recognition and service capability of the space-based system, design the mission flow and information flow of the constellation system, and finally analyze the communication link and timeliness of the system. The simulation analysis results show that the system can achieve minute level efficient anomaly recognition and early warning, effectively improving the service capability to users, this provides an overall idea and architectural reference for the construction of future space-based surface anomaly real-time detection systems.

Air brake system is a kind of brake system commonly used by large trucks, buses and other large vehicles. Air brake system structure is more precise, high frequency of use, so it is often prone to failure of a part. Once the air brake system fails, it is not only easy to affect the normal use of the vehicle, but also bring damage to the safety of the vehicle. Therefore, improving the reliability of the air brake system, reducing its failure rate and improving its troubleshooting level is an important principle to ensure its safe use. This paper mainly analyzes and discusses the common faults and solutions of air brake system and sums up some common points effectively.

Under the background of the development of new energy vehicles, the traditional three major parts of vehicles change with the change of energy supply. The traditional hydraulic brake can no longer adapt to today's electro-mechanical system, and the EMB system has become the inevitable direction of the development of the brake system. The current closed-loop control system of the EMB design is not fully developed, and there is a lack of consideration for the impact of brake clearance. In this paper, a lightweight EMB braking system is designed, and key mathematical models related to it are established. The control output of braking force is optimized by the PID algorithm. At the same time, a brake clearance compensation strategy is proposed. After simulation test, the system has good braking force control performance, and reduces the incidence of motor underspeed, and the overall system has good feasibility and stability.

In order to improve the environmental monitoring and operation management level of the power distribution room, this paper launches the construction of an integrated sensing and control system for intelligent power distribution room based on multi-sensors. In this paper, a scheme of integrated sensing and control system for intelligent power distribution room is proposed, and a multi-sensor structure is designed. Each sensor module converts the information related to gas, temperature and humidity into electrical signals or digital signals, and then the microprocessor processes and stores the data at the same time. When the detected parameters have exceeded the threshold set by the system, the upper computer displays the abnormal operation of the power distribution room. Aiming at three data fusion structures of multi-sensor parallel distributed detection fusion system, a soft threshold decision wavelet domain filtering algorithm is adopted, and this algorithm is effectively introduced into the multi-sensor parallel distributed detection data fusion structure. The results show that in the process of temperature and humidity measurement, the maximum relative error between the temperature measured value and the reference value is 2.6%, and the maximum relative error between the humidity measured value and the reference value is 0.3%, both of which are much larger than the relative error of the gas module measurement, but all meet the requirements. It is proved that the sensor meets the field application requirements of sensor equipment.

The system response caused by large power disturbance includes large deviation of generator rotor angle. Its stability depends on the initial operating conditions and the severity of disturbance. Fuzzy control is an attractive new control method and technology in the control field. This paper mainly takes PSS (Power System Stabilizer) as the research object to study the method of parameter optimization of damping controller. Fuzzy control algorithm is used to optimize and improve the control parameters of PID control algorithm, which makes the optimized PID algorithm more suitable for vibration control of PSS. Its core idea is to use fuzzy control algorithm to optimize and update the control parameters of PID algorithm in real time, overcome the limitations of PID control algorithm, and make it suitable for PSS vibration control. The simulation results show that the power system's ability to suppress low-frequency oscillation is significantly enhanced by adopting PID controller optimized by fuzzy control algorithm.

With the in-depth study of artificial intelligence, robotics and other related fields, many research directions have emerged, and intelligent robots, as one of the research topics, have attracted more and more attention. Intelligent robots have been widely used in various fields. Intelligent robot provides a standard, and quickly becomes an important part of international robot enthusiasts' research on artificial intelligence and robotics, which promotes the research on artificial intelligence theory, control algorithm, robot structure and specific applications. In this paper, a relatively complete intelligent robot motion control system is built through PID algorithm and artificial intelligence, and the preliminary design requirements of the motion control system are realized. The next step is to improve the system performance and add more design schemes to make the mobile robot meet more functional requirements. Therefore, the quality of the robot control system based on PID algorithm will directly or indirectly affect the working ability of the robot. For mobile robots, the control of mobile ontology is more important, because other actions of mobile robots are generally based on the control of ontology motion.

The popularization of automobiles has brought convenience to people's travel. The following problem is that non-renewable energy is facing depletion, and automobile exhaust has seriously polluted the environment, which is one of the main pollution sources causing "haze". As people increasingly face up to environmental protection and health, promote green travel, and strengthen the belief in the coordinated development of human and nature. In order to improve the regenerative Braking Energy (BE) recovery rate of pure Electric Vehicles (EV), Mamdani type fuzzy controller with total braking force demand, vehicle speed and battery SOC as input and motor braking force coefficient as output is adopted to determine the proportional distribution between motor braking force and mechanical braking force; Under the premise of ensuring the good braking performance of the vehicle, the power consumption of the battery under a CYC_UDDS cycle can be reduced, the energy utilization rate can be improved, and the driving range of the electric vehicle can be effectively extended on one charge. This study provides a reference for the formulation of regenerative braking control strategy for pure electric vehicles.

Path planning is a core algorithm component of AGV equipment, and its quality directly affects the operating efficiency and accuracy of the equipment. The current mainstream path planning algorithm A* and the traditional Dijkstra algorithm have the disadvantages of high computational complexity and limited accuracy. Aiming at the limitations of existing algorithms, this paper improves the Dijkstra-like algorithm by using the memorable precursor point list and introduces the kinematics model of the AGV car to optimize the PID control details. The comparison test proves that the improved algorithm can effectively improve the planning efficiency and stability of equipment operation.