This PDF file contains the front matter associated with SPIE Proceedings Volume 13071, including the Title Page, Copyright information, Table of Contents, and Conference Committee information.

This paper takes the research background of "the reverse load simulation test method for automobile steering gear" as the starting point. A test platform for an all-electric test system is set up to achieve the purpose of the research which is to realize the forward and backward drive follow-up load method of all-electric tests. According to the principle of modular design of the experimental table, the layout scheme and the overall frame of the test bed were conceived, and the overall design of the test bed was completed. Based on this, It designed the input and output shaft load unit, and the reverse brake control method was studied emphatically. The problem of the reverse load energy of the output shaft is solved, and the shortcomings of the original design are improved.

Selective laser melting (SLM) technology is widely used in additive manufacturing of three-dimensional complex parts, but its high temperature gradient produces residual stresses on the components, which significantly affects the quality of the components. A 3D finite element model (FEM) of AlSi10Mg is built to simulate the SLM forming process, and evolution of residual stresses are analysed, as well as the effects of different exposure times and laser powers on residual stresses. The outcomes demonstrate that the distribution of residual stresses is characterised by periodicity, and the stresses in the scanning channel and between adjacent scanning channels, and at the lap of the melt pool are obviously larger than those in other regions. The residual tensile stress enhances with the magnify of laser power, however when the laser power maximise to a certain value, the remelting effect is significantly enhanced, and the residual stress is better released. The longer the exposure time is, the more uniform the heating of the melted area, the slower the cooling rate, and the gradual reduction of residual stress to ensure the processing quality.

This paper utilizes an enhanced YOLOv7 network model, incorporating the Swin Transformer as the backbone network, to enable automated identification of internal defects within 3D printed lattice structures. By harnessing the robust adaptability and contextual capturing capabilities of the Swin Transformer, it effectively mitigates the limitations of YOLOv7 in handling diverse image sizes and detecting small objects. Through validation using CT slice images of the 3D printed lattice structure, the results indicate the recognition accuracy of 96.2%, surpassing the conventional YOLOv7 approach by 1.7%. The effectiveness and superiority of the methods suggested in this study are supported by these findings.

Additive Manufacturing (3D printing) technology has become a necessary supplement to traditional additive manufacturing because of its unlimited structure, short production cycle, high density and high strength. As the core component of the through the drill pipe tool neutron compensation instrument, the pressure supply silo must integrate the neutron probe and other components in the limited space and must bear the complex loads such as high temperature (175°C), high pressure (140 MPa), tension and so on. The structure design and manufacturing process of the instrument are put forward high requirements. The Integrated Design, mechanical simulation, printing support optimization and high temperature and high pressure limit mechanical test of the pressure supply silo are studied. The integrated structure design and machining problem of the welding assembly (15 parts welding) of the pressure source cabin is solved, and the structure optimization and the mechanical properties of the material are realized. The near-shape part of the material adding technology simplifies the processing procedure, greatly reduces the cutting amount, and improves the material utilization ratio. The one-time integral forming of the curve hole simplifies the processing and assembly process. 12 samples were tested at 175°C, 140 MPa, 30 tons of maximum safety mechanics and limit mechanics simulation. The results show that the integrated design and processing technology of welding components based on additive manufacturing technology can meet the technical requirements of pressure source cabin, the problems of unstable structure and process reliability and complex machining are solved effectively, and a new feasible technical route for the integrated design and manufacture of similar welding components is opened up.

The resonator's manufacturing process will result in an uneven mass distribution, and when it operates, the vibration of the resonator will generate an inertia force on the surface of the hemispherical resonator sphere, which will impact the resonator's functioning state. In this study, the dynamic equation of the resonator is established under the excitation condition, the inertia force is built into the equation, and the dynamic equation of the resonator under vibration is obtained. To investigate the impact of the relative amplitude and phase of the vibration amplitude and mass deviation on the deviation angle, the dynamic equations are simulated using Matlab. It is proven that the error is caused when the vibration frequency is equal to the intrinsic frequency of the resonator.

In the production process of OLED, the test equipment of each process section is an important guarantee to improve the output. In this paper, a defect detection device for OLED panel is developed, including an electrical control unit, an image acquisition mechanism, a light source assembly and an industrial computer. Output the detection results by analyzing and processing the acquired images. The system has good real-time performance, high efficiency and strong detection ability in industrial field application. It solves the technical problems such as long detection time, low efficiency and strong subjectivity in manual detection of OLED panel defects.

The issue of AQI is a hot topic of public concern today and monitoring the main pollutants in the air is a prerequisite for addressing AQI issues. Research has shown that CO plays a major role in AQI and accurate monitoring of CO concentration in the air is particularly important. CO monitoring generally uses MQ-7 sensors, which needs to solve the problems of sensitive body resistance heating and temperature and humidity compensation during use. In response to these problems, a hardware compensation circuit with dual power supply cycle rotation heating is designed in this paper. At the same time, temperature and humidity data in the tested environment is obtained through the sensor. Based on the temperature and humidity characteristics of MQ-7, software compensation is used to improve its monitoring accuracy. Finally the article verifies through experiments that this method can make MQ-7 not affected by temperature and humidity, and accurately monitor the CO content in the air.

In order to protect the safe and stable operation of relay protection devices and make them retire in the best years, a service life prediction method of relay protection devices considering acceleration state and operation characteristics is proposed. The service life prediction structure of relay protection device is established, and its service life is evaluated from the characteristics of the device itself. By running the degradation track of service life, the service life of relay protection device in the current state is grasped, and the service life prediction model considering acceleration state and operation characteristics is constructed, and the model is solved to realize the service life prediction. The experimental results show that the prediction results of this method have a high degree of fitting with the actual situation, and the service life of relay protection devices can be accurately predicted more reliably, which is reliable and feasible.

In modern manufacturing technology, ultra precision machining technology has become an important and advanced machining process. In recent years, many new principles and methods of ultra precision grinding have been developed both domestically and internationally, and have achieved various applications. This time, the structural design of the precision grinding and polishing equipment has been redesigned. The mechanical structure adopts a rotating workbench, which is driven by a motor through gear transmission. Then, the workbench rotates freely through worm gear transmission. A cylinder loading device is installed on it, and a linear servo moving device and a rotary polishing device are installed on the left side of the rotating workbench.

Based on the issues of short battery life and slow charging of mining explosion-proof lithium battery vehicles, the current status of battery replacement technology at home and abroad and the development of the entire industry were studied. Combined with the current market situation, a set of battery replacement technology suitable for mining explosion-proof lithium battery vehicles was designed using a large capacity explosion-proof lithium battery vehicle produced by Taiyuan Coal Science Institute as an example. The corresponding structural design, simulation analysis, and prototype testing of the battery exchange device were carried out, The feasibility and reliability of the scheme have been verified, providing some reference for the large-scale implementation of the mining explosion-proof electric vehicle power replacement technology in the future.

To investigate the connection between nozzle jet performance and structural characteristics (contraction angle θ, outlet diameter d, ratio of straight segment to outlet diameter L2/d), impact force studies were performed on nine nozzles with varied structures using a self-developed water jet experimental platform, with target distances of 20mm, 100 mm, 200 mm, and 300 mm with jet pressures of 0.1 MPa, 0.2 MPa, and 0.3 MPa. The impact force of a nozzle water jet grows dramatically as the outlet diameter increases. When the pressure of the water jet remains constant, the impact force increases as the target distance increases. The maximum water jet impact force is 6.1 KG when the d is 11 mm. The BP neural network, the PSO and the GA-BP neural networks were utilized to forecast and assess the nozzle impact force at a target distance of 300 mm, respectively. The results reveal that, when compared to the PSO and the BP neural network, the GA-BP neural network projected values are more consistent with the measured values, with a lower average error rate and greater predictive capacity.

The complexity of modern aircraft systems makes the causes of aircraft faults complex and diverse, resulting in difficulty in fault location. This paper elaborates the related research of fault diagnosis for the aircraft system and labors intelligent methods of fault diagnosis. In the fault diagnosis test based on the actual fault data reported by a certain type of aircraft, the fault location accuracy of the fusion method is generally better than that of the single method, and fusion diagnosis method based on rough set theory and BP network is the best. Obviously, it is an inevitable trend that different diagnosis methods are effectively fused to locate automatically the fault of aircraft system.

Due to the development and expansion of cascade power plants in the upper and lower reaches of the basin, the flow rate increased during the dry season, and the low-head operating conditions of the power plant increased significantly. The operating conditions of the unit changed and it was difficult to reach the rated output. In order to ensure the safe operation and economic benefits of the power plant, the runner of the bulb tubular turbine unit was technically modified and various stability tests and analyzes were conducted on the modified unit; the test results showed that all operating indicators met the requirements. The unit operated stably under the modified design conditions and achieved the expected purpose of the turbine technical transformation. Suggestions were made for the safe and stable operation of the unit based on the operation conditions.

Insulator defect detection of transmission lines is one of the important tasks in the operation of power systems. Therefore, the self-explosion defect of insulators has become an important task in power inspection. Therefore, this article proposes an improved YOLOv7 based algorithm for detecting self-exploding defects in insulators. Firstly, select the loss function WIoU to reduce the missed detection rate; Then, use the CARAFE module for feature recombination to improve the detection accuracy of the model; Finally, the CBAM attention mechanism is cited to further improve the model performance. Through experimental verification, the improved YOLOv7 model has good detection performance, with mAP@.5% and Recall increased by 1.3% and 2.4% respectively, making it more suitable for detecting self-exploding defects in insulators.

Robots have become an integral part of dangerous deep-sea exploration and operations. Aiming at the shortcomings of the existing Autonomous Underwater Vehicle (AUV) structure, such as limited equipment, high energy consumption and poor manoeuvrability, an improved method is proposed. This paper proposes an innovative underwater vehicle with a novel structural design that offers significant advantages, including improved manoeuvrability, higher sensitivity and relatively low energy consumption. Through a series of floating, diving and position displacement experiments, it is confirmed that the underwater robot exhibits basic mobility in different degrees of freedom, thus verifying the effectiveness of the improved method. Aiming at the existing problems, the innovative AUV design improves the manoeuvrability and sensitivity, and reduces the energy consumption by improving the shape structure of the AUV, and its effectiveness is verified by experiments.

Aiming at the need of health evaluation of spaceborne modulation and demodulation equipment, a method of load health evaluation based on order relation analysis and entropy weight method was proposed, and membership function was introduced to solve the matching problem of equipment weight and health level. Firstly, the health index system of spaceborne modulation and demodulation equipment is established, and then the indicators are uniformly processed. The subjective weights of each index are determined by the method of order relation analysis and expert experience. By adding entropy weight method to the data features, the objective weight of each index is obtained, and the comprehensive weight is obtained by Lagrange multiplier method. Then the health membership function is obtained based on fuzzy set theory to make the comprehensive weight of each index match the health status evaluation results. Finally, the feasibility of the index system and evaluation method is verified by simulation experiments.

The three-dimensional simulation model of torque sensor is established by finite element, the overall structure of the sensor is optimized, and the magnetic flux distribution and output voltage of the optimized structure are simulated, and the output data are obtained. Finally, a experimental platform is built to measure the static characteristics of the torque sensor. the sensitivity is 0.688 mV/Nm when loading, 0.597mV/Nm when unloading. The error of simulation output data is compared with the actual experimental data. When loading, the error is about 4.36% and the accuracy is 6.31%. During unloading, the error is about 5.14% and the precision is 6.59%. This paper shows that the torque sensor based on Galfenol alloy has a considerable effect, which provides reference for torque sensor.

The structure of a Galfenol torque sensor is designed based on the inverse magnetostriction effect, the thickness and grid width of the Galfenol patch are optimized through Comsol finite element simulation. The experimental platform is built according to the simulation results. the deformation of the sensitive element using a grid structure is increased by 9.82% compared to a thin plate structure; Compared to not widening the bottom of the magnetic choke, the stress on the bottom of the magnetic choke is reduced by 54%. The research in this paper shows that the sensor has high precision in small range of torque measurement, which provides reference for torque sensor.

Aiming at the influence of journal inclination Angle on water-lubricated bearings, a mechanical model of propulsion shafting bearing load was established by theoretical analysis and numerical simulation to analyze the bearing force. Then a water lubricated bearing model is constructed, and the distribution characteristics and rules of water film pressure are obtained by simulating the inclined water lubricated stern bearing. The change of vertical inclination Angle has no effect on the water film pressure distribution. At low speed, the change of inclination Angle has little influence on the water film pressure, while at high speed, the influence on the water film pressure is obvious. When the pressure reaches the peak, the water film breaks, which makes the bearing contact stress value more concentrated.

Due to the lag in temperature detection, non-uniform temperature distribution across various parts of components, and differences in phase transition temperatures among different materials, using the overall phase transition temperature of the repaired component as the boundary temperature for controlling component deformation still leads to significant thermal deformation. This paper explores the application of the golden ratio division method to phase transition temperature and utilizes this temperature as the control temperature for additive manufacturing. Based on this temperature, the entire additive manufacturing process is designed and experimentally verified. The experimental results demonstrate that the additive manufacturing process controlled by this temperature can significantly reduce deformation, enhance the quality of the workpiece, and reduce subsequent machining requirements.

In order to effectively solve the corrosion problem of the steel rail contact network in coal mines, the advantages and disadvantages of the existing physical rust removal, laser rust removal, and chemical rust removal methods were analyzed. A more suitable physical rust removal method was selected based on the actual operating conditions. At the same time, according to the design and functional requirements of the rust removal device, a rust removal device with automatic lifting, left and right deviation compensation, and adjustable pressure was developed. After on-site industrial testing, it was used, Successfully completed the rust removal work of the overhead contact system, verified the feasibility of the device, improved the efficiency of the rust removal work of the underground overhead contact system in the coal mine, and reduced the labor intensity.

With the rapid development of deep learning in the field of image processing, aircraft detection technology has become a hot spot in computer vision research. In this context, this paper proposes an airplane detection model based on improved Swin-Transformer. The model mainly integrates two advanced techniques: the attention mechanism of ECA (Efficient Channel Attention) and the multi-scale module of BiFPN (Bidirectional Feature Pyramid Network).The introduction of the ECA attention mechanism enables the model to more accurately capture and emphasize the aircraft features of key channel information, thus enhancing the recognition of aircraft targets. Meanwhile, the application of BiFPN optimizes the model's performance in dealing with aircraft targets at different scales, which is especially more effective in detecting small or long-range aircraft. A series of experiments on standard aircraft detection datasets show that our improved model achieves 98.3% and 97.6% in terms of accuracy and recall, respectively, compared to the traditional Swin-Transformer model. These results not only demonstrate the value of the ECA attention mechanism and BiFPN in aircraft detection, but also provide a new direction for future Transformer-based target detection research.

At present, how to manufacture lightweight and flexible wearable sensors suitable for human body has become an important scientific and technological problem. However, most of the existing manufacturing methods of portable flexible wearable sensors have the problems of complex manufacturing process and low design freedom. The laser induced graphene technology with high efficiency and customizable advantages is conducive to solving these problems. In this paper, the effects of flame retardants and fabric substrates on the properties of graphene were studied. The effects of sodium tetraborate and ammonium polyphosphate on the structural stability of fabric graphene were compared, and the effects of flame retardant concentration on the properties of graphene were analyzed. Ammonium polyphosphate solution with a concentration of 0.45 g/ml was selected as the flame retardant. The laser induced graphene fabrics prepared by different fabrics were compared and analyzed. Choose 120 cotton fabric as the substrate. Secondly, the effects of laser processing parameters on the properties of graphene were analyzed. Different combinations of laser power and scanning speed were used to analyze the mechanical and electrical properties of the prepared graphene. Determine that the laser power used is 3 W and the scanning speed is 125 mm/s. Finally, the feasibility of fabric graphene structure monitoring human ECG signal is demonstrated by comparing commercial ECG signal detection instruments. The manufacturing method of lightweight flexible fabric sensor based on laser-induced graphene has the characteristics of high efficiency and easy customization, which is expected to realize multi-modal monitoring of human health information, and has application prospects in medical health, human-computer interaction and other fields.

During the shock wave test, there are vibration noise interference problems due to factors such as the test equipment itself and the vibration of the environment, which affects the analysis of the test results. This article analyzes the mechanism and characteristics of vibration noise generated in shock wave testing, points out the direction of removing vibration noise interference, and introduces adaptive filters to remove vibration noise in data processing, achieving certain results.

Choosing an appropriate lubricating oil replacement strategy is crucial for the machine’s operation and maintenance. Based on the concept of condition-based maintenance (CBM), this article proposes a method for predicting the remaining useful life (RUL) of lubricating oil using lubrication condition monitoring (LCM) data and machine learning (ML) theory. Firstly, obtain lubricating oil samples through engine bench tests and quantitatively analyze the elemental content of the lubricating oil in use using atomic emission spectroscopy (AES). Then, a method for finding the optimal back propagation (BP) neural network was proposed to construct a lubricating oil RUL prediction model. The content of 12 elements in lubricating oil is used as the input variable, and the three states of lubricating oil are used as the output variable. Finally, by comparing with the lubricating oil RUL prediction model based on support vector machine (SVM), it is shown that the proposed optimal BP neural network model has better accuracy and robustness.

The utilization of Karman vortex shedding forces as the primary excitation source is proposed in this study for a piezoelectric energy harvesting device aimed at meeting the long-term self-sufficiency and power generation requirements of small rotary mechanisms. The piezoelectric layer material employed is zinc oxide (ZnO), and the device is affixed to a biased cantilever beam. Simulation analysis of the Karman vortex shedding phenomenon was executed using COMSOL Multiphysics software to derive force variation curves acting on the piezoelectric energy harvesting device at different rotational speeds. The optimal rotational speed was determined, and the force at this speed was applied to the piezoelectric energy harvesting device. Through adjustment of the geometric parameters of the device, the objective was to maximize the voltage output. The results indicate that an increase in the length of the film, a reduction in width, and an augmentation of the thickness of the piezoelectric layer can, to some extent, enhance the output voltage at both ends of the film. When the film length is set at 30 mm, the width at 3 mm, and the thickness at 0.002 mm, the maximum open-circuit voltage at both ends of the film can be achieved at 42.2 V.

The thin-walled structure of fully-wrapped carbon fiber reinforced hydrogen storage cylinders for unmanned aerial vehicles elevates the requirements for mechanical processing and non-destructive testing. Industrial computed tomography technology can offer valuable information on the distribution of the linear attenuation coefficient inside the tested object, enabling intuitive defect detection and size measurement. Here, we conducted defect detection and wall thickness analysis by industrial computed tomography on the hydrogen storage cylinder for unmanned aerial vehicles and found that the thickness of delamination defects at the cylinder neck and cylindrical shell was approximately equivalent. The wall thickness of the Al liner and carbon fiber layer exhibited relative minimum and maximum values at the Al liner abnormal area, respectively, likely resulting from that the sudden thickening of the carbon fiber layer caused the Al liner to shift inward under the additional pressure from the carbon fiber layer. These results proved that industrial computed tomography technology could effectively inspect the quality of hydrogen storage cylinders for unmanned aerial vehicles.

For the domestic wheel-turning repair standards of different rolling stock models, by analyzing the wheel diameter difference requirements after wheel-turning of the same axle, the same bogie, and the same carriage of different rolling stock models, and based on the measured wheel wear data of CRH380B rolling stock, and considering the different requirements of wheel diameter difference of the whole EMU and trailer of the rolling stock, we formulate the strategy of wheel-turning repair in line with the wheel-turning strategy of the whole EMU and trailer of the rolling stock. Because the wheel pair tread geometry is determined by the wheel rim thickness and wheel diameter, and is limited by the wheel diameter difference of the whole train set, the wheel pair turning strategy model based on the wheel rim thickness and wheel diameter difference of the whole train set is proposed, and the wheel pair real-time turning strategy of the wheel diameter difference and wheel rim thickness of the whole train set is designed, so as to achieve the purpose of optimizing the wheel pair turning of the whole train set, which is of great engineering significance and socio-economic significance.

During the unlocking process of the rotating-bolt locking special machinery head, there is a significant resistance. The aim is to select a motor for unmanned aerial combat vehicle (UACV) mounted external power special machinery drives, taking the special machinery unlocking force as the research subject. Adaptive analysis is conducted on the patented structure of the special machinery unlocking force measurement device previously designed. A virtual prototype model is established using the dynamic simulation software ADAMS to apply driving, constraints, and simulation parameters based on the model's principles. Addressing the phenomenon of the detachment of the guide rod slider from the tested bolt carrier during the operation of the special machinery unlocking force measurement device, it is hypothesized that the detachment time is related to the stiffness of the spring and the velocity of the lead screw slider. Through multiple comparative simulations, the correctness of the hypothesis is proven. It is demonstrated that increasing the velocity of the lead screw slider can completely eliminate the detachment time, whereas increasing the stiffness of the spring, although reducing the detachment time, still results in secondary detachment. Consequently, the conclusion is drawn that increasing the velocity of the lead screw slider is the way to resolve the detachment issue of the guide rod slider from the tested bolt carrier in the experiment.

Currently, lithium batteries are widely used in industry such as electric vehicles. It’s important to know the RUL (Remaining useful life). In this paper, we apply a frame of deep learning technology in order to predict the RUL of batteries. CNN method is provided to exact features, while LSTM method is to do the prediction. The NASA lithium batteries data are used for verifying the proposed algorithm. The result shows that the RUL prediction method of lithium batteries is accurate.

In order to evaluate and understand the launch performance and various aspects of the segmented enhanced four-rail electromagnetic launcher, its internal electromagnetic characteristics, current distribution and electromagnetic force were studied. On the basis of constructing a physical model, the calculation formulas for the magnetic induction intensity and electromagnetic force at any point in its internal space are proposed. Through the method of finite element analysis, the magnetic induction intensity and current distribution on the armature and main and auxiliary rails of the segmented enhanced four-rail electromagnetic transmitter were simulated, as well as the magnitude of the electromagnetic force experienced by the armature. The results show that on the segmented enhanced four-rail electromagnetic transmitter, the armature mainly experiences current concentration at the pivot-rail contact position, which may cause various problems. This situation can be further studied in the future.

The research focus of this paper is the on-board equipment of the CTCS 300T train operation control system, and a fault diagnosis method for train control on-board equipment based on Bayesian network is proposed. Firstly, to address the issue of imbalanced distribution of fault types in fault text, we have developed a Three-way Oversampling (3WOS) algorithm to automatically generate subclass text vector data. To tackle the problem of multiple synonyms and single semantics in fault text, we utilize Supervised Latent Dirichlet Allocation (SLDA) to conduct semantic clustering and feature analysis on the fault tracking table, and combine expert knowledge to establish a comprehensive fault information database. Then, we employ the K2 algorithm to train and integrate the collected fault information for building a Bayesian network. Finally, diagnostic reasoning is conducted using actual cases from high-speed railway operation sites of railway bureaus, and experimental results validate that our model exhibits high accuracy and feasibility.

The future products have a tendency to small batch, multi-variety development trend, machine tool as an indispensable part of the production line, to achieve rapid production change provides an important guarantee. From the point of view of how to realize rapid production change of machine tools, this paper constructs a production system of rapid production change of machine tools based on digital twin technology, analyzes the main technology of rapid production change of machine tools, and puts forward a tool path modeling method based on grid method to realize the programming of machining instructions of any part of machine tools. The results show that it has important research value to generate the tool path of parts machining and realize the rapid production change of machine tools fundamentally.

The model simulates the mechanism of a micro-pump without a valve. The simulation realizes unidirectional flow by introducing a vibrating pump mechanism into a microfluidic system and using a flexible microlobe to adjust the flow direction. During the down stroke, the liquid is introduced into the channel from the vertical chamber. The right lobe is then bent towards the bottom of the channel, while the left lobe is moved away from it. In this configuration makes it easier for liquid to flow out of the right outlet. During the upward motion, the liquid flows from the channel into the vertical chamber, while the flap flexes in the opposing direction. In this configuration, the right lobe has a greater ability to control the flow than the left lobe. As a result, the majority of the fluid is pulled into the vertical chamber through the outlet on the left side. This behavior leads to a flow of network traffic moving from the left side to the right side within the channel. Valveless micropump can be applied in microfluidic systems and can also be used to create a circulation system that pumps liquids into a continuously circulating loop to cool the microelectronic system.

In the working process of the four-track electromagnetic emitter, the uneven current distribution may lead to the orbital thermal damage, and then affect the service life of the electromagnetic emitter.comsol In order to alleviate the thermal damage, this paper presents the aluminum-copper four-track electromagnetic transmitter model, and adopts the finite element analysis method, using the software of aluminum-copper electromagnetic transmitter and the four-copper electromagnetic transmitter can reduce the temperature of the emission rail, improve the current distribution condition, and reduce the thermal damage effectively, so as to improve the service life of the electromagnetic transmitter^[1].

The analysis of spectral fatigue for wind turbine unit legs in this paper is grounded in dynamic analysis. The influence of weight in key areas, including the main hull, main deck, superstructure, cantilever, drilling floor, and others, is meticulously considered through the incorporation of mass elements. The calculation of Response Amplitude Operators (RAOs) for the legs is carried out using a frequency domain spectral analysis method. Critical fatigue locations are discerned by scrutinizing stress responses under design operating conditions, revealing vulnerabilities near the lower guides, spudcans, waterline, and specific chord and brace positions. The selected chords and braces, along with their respective locations, are visually depicted. Given that fatigue damage primarily results from cyclic load actions, this analysis excludes consideration of non-cyclic loads such as buoyancy and gravity. Additionally, the impact of wind load is not factored in; only fatigue resulting from wave and current loads is computed.

In the context of advancing technology and artificial intelligence, this paper introduces an innovative disinfection vehicle with autonomous disinfection, mask recognition, and spraying capabilities. Leveraging advanced obstacle avoidance and positioning technologies, the robot autonomously plans disinfection routes, adapting to various environments and minimizing dependence on human resources. The integration of mask recognition enhances public health safety during the disinfection process. The robot's multifunctionality extends to spraying operations, optimizing overall operational efficiency. Empirical validation underscores its effectiveness and practicality, positioning this vehicle as a robust support system for public health prevention and control efforts. This technological advancement promises efficient and secure disinfection services for public spaces, making a significant contribution to the evolution of public health measures.

A Unified Power Quality regulator (UPQC) can be used to regulate and control power quality issues that arise due to a high proportion of distributed renewable energy and power electronics.UPQC can manage power quality problems such as voltage fluctuations and harmonics caused by high proportion of new energy access and high proportion of power electronic equipment access. In this paper, the working principle and topological structure of the unified power quality control device are analyzed, and its key parameters are analyzed and designed. Secondly, the corresponding control strategy is proposed for the power quality compensation function of the parallel side and series side converters of the unified power quality regulation device.

With the popularization of distributed new energy and the promotion of green and low-carbon energy transformation to achieve carbon peak and carbon neutrality goals, a large number of new elements such as photovoltaics, electric vehicle charging facilities, and energy storage equipment are integrated into the distribution network substation, which will generate multiple harmonics in the substation area. For the high proportion of zero sequence harmonics in the transmission line, magnetic flux compensation zero sequence filters can effectively filter out and improve the harmonic environment. However, the adaptability of zero sequence filters may vary under various working conditions, and may even cause damage to the line or equipment. This article mainly analyses the principle of magnetic flux compensation and builds a filtering circuit model using MATLAB/SIMULINK, Analyzing the filtering effect of zero sequence filters under standard operating conditions and three-phase imbalance can be used as a basis for optimizing the parameters of zero sequence filters, which has great value and significance for the design of zero sequence filters.

This paper presents the design of a high-frequency test fixture for high-speed connectors. Specifically, after analyzing and designing the board material, laminate structure, trace structure, and signal via structure, the impedance consistency and bandwidth of the high-speed interconnect structure are verified through TDR and S-parameter simulations conducted in HFSS. a PCB layout is generated using Cadence. The test results confirm that the differential impedance of the fixture meets the requirements for high-speed connector testing. Within the frequency range of 0-30 GHz, the difference between S11 and S21 for the 2x-thru structure is greater than 0 dB, indicating a bandwidth capability of up to 30 GHz for the fixture.

In order to improve the security of the defended targets in our region and efficiently utilize various types of optoelectronic devices, the research is conducted on optoelectronic devices deployment within the region. With the objectives of maximizing the protection effectiveness and minimizing devices operational cost, a multi-objectives optimization model is established by considering constraints such as devices-target visibility conditions and device types, quantities, etc. Non-dominated sorting genetic algorithmⅡ (NSGA-Ⅱ) improved by Q-learning is designed to solve the model. To address the difficulty of fixed parameter settings adapting to dynamic changes, Q-learning is adopted to adaptively adjust the crossover probability and variation probability. In order to search for Pareto front solutions that are close to the global optimal more efficiently. The correctness of the model and the effectiveness of the algorithm are verified through simulation examples.

In the process of on orbit capture, there is a large contact and impact between the manipulator and the target. The flexibility of variable stiffness joints can effectively reduce the collision force and improve the capture reliability. In this paper, a design method of variable stiffness modular manipulator joint with large stiffness variation range is proposed, the mechanism design and variable stiffness principle of variable stiffness joint are described, and the ability of stiffness variation in large range is verified through experiments.

Due to the presence of false components in Hilbert Vibration Decomposition (HVD), it is unable to accurately identify on-site faults. To address this issue, the Refined Composite Multiscale Dispersion Entropy (RCMDE) is applied to calculate the entropy curve of the vibration signal of the condenser and perform fault classification and recognition. The processing effect of the fault vibration data of the BENTLYRK4 rotor experimental platform using this method shows that the proposed method can effectively identify and diagnose rotor faults. The application of this method can complete the diagnosis and identification of actual operational condenser faults.

Battery is an important part of the icing-monitoring terminal, which directly affects the operating state of the terminal. To study the correlation between the operating status of the battery of the ice-cover monitoring device and the environment, this paper analyzes the factors that may affect the operating status of the battery by counting the failure of the battery of the ice-cover monitoring terminal of EHV in recent years and using the monitoring data such as the voltage and environmental variables of the battery of the ice-cover monitoring terminal at different altitudes. The results show that the battery failure rate of the monitoring device at a low altitude is much higher than that of the monitoring device at a high altitude. The ambient temperature has a greater influence on the operating state of the battery of the monitoring device. The conclusion of the study can provide a reference for the differentiated configuration of the battery.

In order to reduce the serious consequences of pile leg puncture, a strain warning system with wireless data transmission is designed to improve the strength of puncture warning. Owing to the insufficient submarine foundation capacity and the influence of wind and waves, the pile legs are prone to puncture, resulting in platform collapse and serious consequences. Therefore, this paper builds and analyzes the Amesim model of the pile leg piercing phenomenon, and uses a land platform to simulate an ocean platform for experimental verification. The data is collected by wireless sensors and sent via Internet to analyze the data for warning.

As a collector ring material, carbon steel is widely used in domestic active hydraulic generators and high-speed current-carrying friction systems. However, during operation, the carbon brush/collector ring friction pair may lose contact due to the rotor's eccentric vibration, leading to electric sparks or ring fires at the contact interface. These occurrences result in mechanical wear and electrical corrosion, seriously endangering the normal operation of the unit. To clarify the underlying mechanism, this study investigated the evolution of electric spark strength at the interface of D172 carbon brushes and 45# steel, as well as the change in the surface ablation degree of 45# steel under various conditions of load, speed, and polarity. The results demonstrate that under the same load, the relative strength and discharge characteristics of electric sparks are influenced by polarity and rotation speed. Specifically, the arcing rate, maximum relative strength, average relative strength, and maximum duration of electric sparks exhibit positive correlations with rotation speed. As the rotating speed increases, the electric spark of carbon brushes connected to the positive pole gradually transitions from multiple points/columns to strips. On the other hand, the electric spark of carbon brushes connected to the negative pole gradually evolves into spherical/ellipsoidal shapes and shows splashing effects with the increasing rotating speed. Additionally, the degree of surface ablation increases with the velocity, and the ablation is greater in the case of carbon brushes connected to the negative pole compared to those connected to the positive pole.

Temperature variations during the welding process directly affect the melting and solidification process of welding, which has a significant impact on the quality and mechanical properties of the weld. Real-time prediction of the welding temperature field plays a crucial role in solving the problems posed by the instability and complexity of the welding process, and has developed into an important direction in welding technology. In this paper, the real-time temperature variations during the three-layer, three-pass MIG (melt inert-gas) welding of aluminum alloys were monitored in real time. On this basis, a welding temperature prediction model based on the long-short-term memory neural network (LSTM) was proposed to predict the temperature changes in the unwelded region by taking the temperature data of the torch-passed region during the welding process , by analyzing the prediction results in comparison with the actual measured temperature changes, the results show that the method can predict the welding temperature changes better. This method of over-prediction is important for welding quality control, defect prevention and process optimization.

This paper presents a method for compressing test response data using compressive sensing theory to improve chip testing efficiency. The proposed approach preprocesses the test response vectors, ensuring compatibility and eliminating duplicates. These preprocessed vectors are multiplied with sparse random matrices corresponding to the tested circuit. The test response vector with the highest coefficient is selected as the compressed result by comparing the absolute values of the resulting coefficients. Experimental results demonstrate that our approach achieves superior compression ratios and fault coverage compared to other methods. Our approach improves the compression rate by nearly 15% and maintains fault coverage above 90% with less than 1.5% loss of fault information. Overall, our proposed approach offers significant advantages over existing techniques.

NVH performance is one of the most important quality properties of a vehicle, especially for improving the comfort of new energy vehicles. This paper takes the NVH problem of a new energy vehicle as an example to explain how to deal with the noise problem. It is found that there is a howling noise in the starting process of the testing vehicle, which is an NVH problem. This will lead to reduced ride comfort, and ultimately affect vehicle quality and car sales. In order to find out the cause of the howling noise, this paper analyzes the fault problem by means of elimination analysis, bench test and the measurement of vibration noise of the real vehicle. It is found that there are inconsistencies in the initial electric Angle and other parameters of the motor between the loading motor and the pre-calibration motor, which leads to the howling noise. Finally, the starting control parameters of the motor are modified by the controller software to greatly reduce the howling noise. Through a series of analysis methods, the paper finds out the cause of the problem of howling noise, which to avoid the maintenance or replacement of hardware equipment of the vehicle and to improve the comfort of starting driving.

This paper aims to propose an obstacle avoidance trajectory planning and group collaborative control method for UAV based on intelligent cluster. Based on the RRT obstacle avoidance algorithm, the consistency theory algorithm of group collaborative control strategy is fused, and the coordination and collaboration between UAVs are realized through information exchange and distributed control methods, so as to maintain cluster cooperative flight while avoiding obstacles and complete specific patrol strike tasks. The experimental results show that the algorithm can effectively realize the obstacle avoidance trajectory planning and group collaborative control of UAV, and improve the safety, stability and maneuverability of UAV.

To reduce the impact of ship swaying motion on the on-board equipment during movement in complex sea conditions, and improve the performance of on-board equipment in the operation of the ship. Based on the theory of seakeeping, the relationship between the sailing state of a ship and its motion performance under specific sea conditions was studied to address the issue of the ship's sailing state (collectively referred to as the heading and speed) during functioning under sailing conditions; And the multi body dynamics simulation software was used to simulate and analyze the nozzle response characteristics and initial disturbance of on-board equipment during function under different sailing states. The research results indicate that adjusting the ship's navigation state during the on-board equipment in the operation can reduce the disturbance caused by ship sway, and provide guiding strategies for optimizing the ship's navigation state during the function of on-board equipment.

The current conventional cooperative control method of AC-DC hybrid distribution network oscillation mainly constructs the dispatching model by centralized dispatching, which leads to poor cooperative control performance due to the lack of analysis of oscillation influencing factors. In this regard, a cooperative control strategy of AC-DC hybrid distribution network oscillation based on mixed integer linear programming is proposed. The factors affecting the oscillation effect are analyzed from three aspects: power intensity, operating power and oscillation control parameters of the distribution network. The tide calculation equations of DC distribution network and AC distribution network are constructed respectively, and the cooperative control function is constructed and solved by distributed scheduling method. In the experiments, the performance of the proposed method is verified for the cooperative control. The experimental results show that when the proposed method is used to control the oscillation situation, the power exchanged in the distribution network is significantly reduced, and the cooperative control performance is more desirable.

In order to improve the braking energy recovery capacity of explosion-proof electric rubber-tyred vehicle, and reduce the impact of feedback braking on vehicle braking safety. A self-adaptive braking energy recovery control strategy for electric vehicles in coal mines based on road condition identification is proposed, and four braking feedback strategies are formulated to adapt to different road conditions. Through the real-time positioning technology for whole vehicles underground, the self-adaptive switching of braking energy recovery strategies in different coal field environment is realized, so as to better meet the driver's braking needs and intentions under different road conditions. In the simulation platform, the simulation model of the strategy is built according to the actual road conditions in the coal mine, and the results prove the feasibility of the strategy.

This project has developed a 633nm iodine steady frequency laser frequency evaluation system. The overall structure is mainly composed of the national length of the 633nm iodine steady frequency laser, new laser, frequency shooting monitoring and technical unit, and frequent evaluation system. Through design experiments, respectively In the three major measurement technology indicators of the three major measurement technical indicators of the peak frequency difference, the stability of the laser frequency, and the duplication of laser frequency, the experimental result was not certainty of the wavelength of the new system of 2.9 × 10^-11 to meet the relevant technical indicators.

A bandwidth intelligent tunable filter based on filter theory is designed in this article. By selecting a special interstage coupling structure and using a varactor circuit, dual frequency band tunable transmission could be achieved within the broadband range. The final design results show that the insertion loss of the filter is reduced to below 1.65dB and the return loss is greater than -15dB. In addition, the fixed frequency of the filter is 290MHz and the bandwidth is greater than 5MHz. The adjustable operating frequency ranges from 228-275.7MHz, the bandwidth is greater than 4.5MHz and the insertion loss is not greater than 1.2dB. The above parameters meet the filter requirements. This filter could achieve effective signal filtering in communication systems, which could ensuring system electromagnetic compatibility.

In view of the automotive comfort under high-speed following conditions, a semi-active suspension control method based on model predictive control is proposed. The performance function of the controller consists of the predicted output of the vehicle, the output force of the actuator, and their respective weights. Usually, the selection of the weights requires sufficient engineering experience and a large number of experiments to determine. There are mainly three working conditions for constant speed, acceleration and braking conditions, and different weight groups are proposed for different driving conditions. Therefore, the Particle swarm algorithm (PSO) is used to optimize the parameters of the MPC controller, and in order to obtain three groups of weights under different driving conditions. The simulation results show that compared with the passive suspension system, the MPC controller based on the particle swarm algorithm reduces the RMS value of the vertical acceleration of the vehicle under the three working conditions, so the suspension control strategy effectively improves the ride comfort of passengers.

A risk assessment method based on improved fuzzy FMEA is proposed to solve the problem of frequent failures caused by complex working environment of aircraft anti-skid braking system. On the basis of analyzing the application of FMEA, the fuzzy language of fault evaluation is established according to fuzzy set theory, and the standardized evaluation matrix of failure mode is established. is established according to fuzzy set theory, and the standardized evaluation matrix of failure mode is constructed by quantitatively describing expert experience through triangular fuzzy numbers; Then the grey relational-TOPSIS model is nested to calculate the comprehensive closeness degree of various failure modes, and the grey relational-TOPSIS model is nested to calculate the comprehensive closeness degree of various failure modes. Then the grey relational-TOPSIS model is nested to calculate the comprehensive closeness degree of various failure modes, and the final risk assessment conclusion is obtained. The research results show that the improved fuzzy FMEA is better than the The research results show that the improved fuzzy FMEA is better than the FMEA in evaluating risk of aircraft anti-skid braking system,and has good practical application value.

Finite control set model predictive control (FCS-MPC) has been widely used in permanent magnet synchronous motor (PMSM) control. However, the number of voltage vectors (VVs) of the conventional FCS-MPC is 8, which can't satisfy the requirement of high control performance, so a discrete space vector MPC (DSV-MPC) is proposed. The VV plane is firstly divided uniformly, and then the optimal VV is searched quickly with the dichotomy. Considering the limited computational capability of the DSP, the number of iterative searches and the error of the optimal VVs are analysed in detail, which is used to give the number of iterations as a reference. Finally, the effectiveness of the proposed DSV-MPC method is verified by simulation.

In order to study the vibration characteristics and fault characteristics of double-ring herringbone planetary gear transmission system under different gear tooth broken faults. The 3D model of the system is established using SolidWorks, and based on ADAMS virtual prototype and Hertz contact theory, considering the friction coefficient. The angular acceleration spectrum characteristics of the system under the condition of health and fault are studied. The results show that the fault characteristics of the system are obvious when the gear tooth broken fault occurs. Along with the tooth broken failure to deepen, the side-band frequencies on both sides of the fundamental frequency and harmonic frequency of the system are gradually dense. The sun gear has obvious response to the system asymmetry caused by the gear tooth broken fault. The results of the planetary gear system fault diagnosis research and design has a certain reference value.

To achieve parameterized modeling of gears, an involute spur cylindrical gear tooth model was established in Pro/E. Utilizing laser surface texturing and finite element theory, MSC. Nastran was employed to analyze the transient frequency response of both ordinary gears and laser surface-textured gears. Laser engraving technology was applied for surface texturing on the teeth. Both actual bench tests and finite element simulation results demonstrate that, under the same internal excitation, the response of ordinary gears differs from that of laser surface-textured gears. Notably, laser surface-textured gears exhibit the least deformation, effectively enhancing the fatigue performance of gears and bolstering their reliability.

In view of the low accuracy of the existing force feedback control model of steer-by-wire system and the difficulty of obtaining some system parameters, a joint simulation model of steer-by-wire system was established based on CarSim software. Aiming at the safety problem of variable Angle ratio based on constant yaw Angle gain, a scheme is designed to adopt constant Angle ratio at high speed and low speed and constant yaw Angle gain at medium speed. The steering rack force is obtained by measuring the working current of steering motor with high precision sensor, and the feedback force control algorithm is designed with PID control. Finally, the correctness of the control algorithm is verified by the central area steering test and steering portability test, and the performance of ESP system is compared. The results show that the feedback force control algorithm can greatly improve the control accuracy of the feedback torque and make the driver get a good road sense.

High-voltage pulse power supplies with simple structures and high energy transmission efficiency are widely used in the field of electric pulse comminution. To achieve a pulse power supply with a simple structure and high energy transfer efficiency, this paper adopts the structure of a high-voltage pulse generator that combines resonant charging of a pulse transformer with a single-stage Marx generator, and it conducts simulation analysis and performance testing on this pulse power supply. A 10kV pulse transformer was developed, and tests showed its coupling coefficient to be 0.999. Pill fragmentation experiments were conducted using needle-needle electrodes. The results indicate that the average charging efficiency of the high-voltage pulse power supply is 89%; the output pulse voltage peak is 11.2kV, with a rise time of 20 ns. Through experiments, the pulse power supply developed in this paper successfully fragmented a 1.57mm thick pill with only a single pulse application.

With the development of science and technology, people have higher and higher requirements for the stability of high-precision servo systems. In this paper, the principle of dual-motor anti-backlash is studied. By modeling the mathematical characteristics of the whole servo system, the drive control block diagram of the dual-motor is designed. Finally, the servo system model of the dual-motor anti-backlash transmission is established by MATLAB. Based on the servo system model, the control algorithm of the system is designed, and the fuzzy-PID control algorithm is designed in the position loop. According to the threshold of the system, a more suitable control algorithm is selected to improve the stability and robustness of the system. A new anti-backlash method for improving the constant bias torque is designed, and the model of the anti-backlash module is designed. Finally, through simulation verification, it is concluded that the fuzzy-PID control strategy can effectively reduce the overshoot of the system, and the added anti-backlash module significantly improves the stability of the system.

This paper introduces an algorithm for improving the accuracy of the ADC acquisition system based on pruning neural network, which can calibrate the errors caused by the analog-to-digital conversion module and other parts in the acquisition system, and effectively improve the accuracy of the ADC acquisition system. By using techniques such as neuron pruning, weight clustering, and parameter quantization, the network we trained greatly reduces hardware resource consumption while achieving the calibration effect of a fully connected neural network. It enables this network easier to deploy in embedded systems. The simulation results show that in the case of a signal input close to the Nyquist frequency, for a 12- bit 12.5MS/s ADC acquisition system, the ENOB can be increased from 5.31 to 8.83, and the SFDR can be increased from 46.3dB to 66.4dB.

Modern civil aircraft heavily rely on air data sensors to precisely measure crucial parameters, including static pressure, total pressure, angle of attack, and total temperature of the airflow. These data are then processed by advanced air data computers to derive essential flight information such as altitude and speed. However, it is imperative to acknowledge that any slight inaccuracies in the installation position of the static pressure ports can directly impact the accuracy of flight altitude and speed measurements. Hence, this comprehensive research article focuses on calculating and analyzing the potential errors associated with the positioning of these pivotal static pressure ports. By providing a solid theoretical foundation, this study aims to facilitate the determination of an optimal positioning degree for static pressure ports on civil aircraft, thereby enhancing flight safety and performance.

Large scale wind turbines replace traditional generators and are integrated into the power grid. The increase in wind power penetration rate leads to a decrease in the overall inertia of the power system. To alleviate the impact of wind power grid connection on power system stability, a comprehensive frequency regulation control strategy is proposed for wind turbines and energy storage to coordinate and participate in system frequency regulation, in order to achieve adaptive adjustment of active power with system frequency changes. A wind power grid connection simulation model is built using Matlab/Simulink software. The experimental results show that this method can spontaneously respond to frequency difference signals, timely release rotational kinetic energy to provide active compensation, effectively avoiding secondary frequency drops and improving the stability of wind power grid connection frequency.

Aiming at the common defects detection in the composite materials, a tapping probe was designed by using electromagnet and piezoelectric sensors. A percussion detection system for composite materials was built based on LabVIEW. The system includes double-axis plane scanning module, knocking head, driving module, signal acquisition module, signal processing and storage module, C-scan real-time display and alarm module. The system can display signal intensity in C-scan diagram and analyze the signal waveform. The test results of honeycomb core structure and helicopter rotor foam core composites show that when there are defects in the composite structures, the waveform of the defective position will be wider and the amplitude will be reduced. The experiment proves that the system has high detection efficiency and reliability, and can realize automatic detection.

The offshore anti-rolling crane has the anti-rolling device consisting of four connecting rods with a rotating joint at the end of crane arm. Under the control of the motion control system, the anti-rolling device can compensate for the spatial displacement of the crane's lifting point caused by swinging, thereby reducing the rolling of the load. In order to calculate the spatial displacement of the lifting point before and after the swing, this paper establishes the kinematic model of the crane arm and the anti-rolling device. Since running in swing condition, the motion control system faces several challenges, such as load sudden change, the given signal time-varying, external signal interference. In order to improve the control performance, this paper builds the simulation model of servo motor control system of the anti-rolling device applying Fuzzy-PID control in MATLAB/Simulink. The simulation results show that the Fuzzy-PID control is more suitable for the motion control system of the offshore anti-rolling crane in swing environment, which significantly improves the control performance, such as stability, following ability and robustness.

At present, most of the individual UAVs^[1] are mainly flying, and the concealment effect is not good when approaching the enemy in field combat, and it is easy to be found and shot down, and the rotor of the UAV is relatively fragile, and it is also more dangerous to fly in a closed and narrow environment such as indoors. The land-air amphibious individual drone can flexibly switch combat modes, and can carry explosives to conceal itself by changing modes when approaching the enemy in the field. According to a land-air amphibious UAV structure, By establishing the overall dynamical system model, the components of each part of the dynamical system are designed and selected, so that the air mode dynamical system and the ground mode dynamical system can be coordinated and converted, and the specific performance parameters are calculated., and provide a reference for the coordinated design of the ground dynamical system and the air dynamical system of individual UAV in the future.

In order to study the seismic performance of 550 kV composite external insulation bushing for UHV substation in high altitude area, the shaking table test of seismic simulation was carried out on a real bushing and the numerical simulation analysis was carried out. By inputting white noise random wave and artificial wave, the dynamic characteristics of the specimen and the acceleration, strain and displacement responses of the key parts under the peak acceleration of 0.2g, 0.6g and 0.8g earthquakes are obtained. Through numerical simulation analysis, the mechanical response of the casing under wind load is verified. The results show that the first-order frequency of 550 kV composite external insulation bushing is 3.69 Hz ; under the action of the standard time history wave with the basic seismic acceleration of 0.8g, the strength of the casing composite component meets the seismic requirements, and the displacement response is obvious. The strength of the casing itself meets the mechanical requirements under wind load.

With the development of technology, the pressure sensor in MEMS system is required to be small in size and high in accuracy, but the experiment is difficult and the error is large. In this paper, the finite element method is used to analyze a silicon-based pressure sensor model composed of a square film with an edge length of 1 mm, a thickness of 20 m, a p-type doping density of 1.321019 cm-3 and a thickness of 400 nm, and the output voltage sensitivity of the silicon-based pressure sensor model is 0.6mV/kPa at 0-100 Kpa, which meets the requirements of high sensitivity and linearity of MEMS pressure sensor.

With the breakthrough progress of deep learning technology in various fields, its application in fault diagnosis and prediction of electrical systems has received more and more attention. In this paper, a deep learning-based fault diagnosis and prediction model is proposed for the complex nonlinear characteristics in electrical systems. First, key features are automatically extracted from a large amount of electrical system operation data using a multilayer autoencoder (MLAE). These features are fed into a deep neural network (DNN) for fault classification and prediction. In order to improve the robustness and accuracy of the model, an attention mechanism is introduced so that the model pays more attention to the key features related to faults during the learning process. The method demonstrates high accuracy and reliability on multiple electrical system fault datasets compared to traditional electrical system fault diagnosis methods. In addition, this study explores the interpretability.

In order to solve the problems of low energy conversion efficiency, high peak current in the main circuit, and weak adaptability to the shape of the projectile in traditional external drive responsive emitters, an electromagnetic emitter with an internal drive direct fed armature structure is proposed. By establishing the electromagnetic mechanical coupling models of the direct fed armature emitter and the induction armature emitter, the circuit and motion characteristics of the two emitters were solved using numerical analysis methods, and the correctness of the model was verified using the finite element simulation results of field circuit coupling. The solution results show that the energy conversion efficiency of the internal drive direct fed armature electromagnetic transmitter is 51.12% higher than that of the internal drive induction armature electromagnetic transmitter, and its peak circuit current is about 7 kA lower than that of the induction transmitter. It has been proven that the internally driven direct fed armature electromagnetic transmitter has higher energy conversion efficiency and lower loop current peak.

With the emergence of the smart factory concept, Automated Guided Vehicles have been widely utilized in job shop transportation systems, drawing increased attention to the job shop scheduling problem integrated with multiple AGVs. In this problem, the sequential-logic constraint between transportation and processing tasks in each operation complicates the problem's decoupling compared to traditional job shop scheduling problems. In this paper, a deep reinforcement learning-based method is proposed to address this integrated scheduling problem. The problem is initially formulated as a Markov Decision Process, with state features designed from the perspectives of jobs and AGVs. Furthermore, a novel reward function is designed to maximize job shop resource utilization. The primary objective of the scheduling is to obtain an optimal solution that maximizes resource utilization, thereby enhancing the core competitiveness of the factory. Benchmark experiments demonstrate the effectiveness and competitiveness of this method in obtaining solutions for the integrated problems.

This study focuses on the nonlinear coupled dynamic analysis of deepwater top Tension riser systems. Using OrcaFlex software, we developed a model of a dual-casing top Tension riser equipped with centralizers. Sensitivity analyses of key parameters, such as centralizer height and top tension factor, revealed the coupling mechanisms and dynamic response characteristics of the dual-casing top-Tension riser. Additionally, the study examined variations in riser displacement amplitudes and pipe-to-pipe contact forces at different water depths. The model and methodologies proposed in this study provide crucial theoretical and practical guidance for the design and operation of deepwater top-Tension riser systems. At the same time, the LSTM neural network model was used to train a dynamic response model of a deep-water top Tension riser systems.

Aiming at the problems of high energy consumption and large tracking error in the dual servo valve drive control of fully automatic hydraulic heavy-duty manipulator, a dual servo valve drive control method for fully automatic hydraulic heavy-duty manipulator was studied. Based on the two port theory, generate a dual servo valve drive control strategy for a fully automatic hydraulic heavy-duty robotic arm. Taking control accuracy and control response time as necessary conditions for servo control, control inference simulation, design a dual servo valve drive controller for fully automatic hydraulic heavy-duty manipulator, and achieve dual servo valve drive control for fully automatic hydraulic heavy-duty manipulator. The experimental results show that the average control tracking errors of the A-level and B-level dual servo valve drivers of the proposed method are 0.1386m and 0.1434m, respectively, and the average servo drive energy consumption is 1.4257×10⁴J and 1.5257×10⁴J has good control performance, can reduce driving control energy consumption, and reduce tracking errors.

In the macro context of promoting the green and low-carbon transformation of energy to achieve carbon peak and carbon neutrality goals, a large number of new elements such as distributed photovoltaics, electric vehicle charging facilities, and energy storage equipment are integrated into the distribution network substation, which has a complex and far-reaching impact on the harmonic environment of the substation. This article mainly analyses the harmonic characteristics of common power electronic converters in the substation. The substation is modelled using PSCAD software, and complex harmonics in the substation are simulated using single-phase uncontrolled rectification capacitive load. The harmonic current content of the neutral line is analyzed, and the conclusion that the high-frequency zero sequence current on the neutral line is three times the zero-sequence current of each phase is verified.

The research object of this paper is the automatic obstacle avoidance tracking car, which can prolong the black track when the car detects the black track, stop when it encounters an obstacle, and continue to delay the track when the obstacle leaves the front of the direction of the car. Realize the function of automatic obstacle avoidance and tracking. The hardware design includes track recognition module circuit, obstacle identification module circuit, DC motor drive module circuit, single chip microcomputer minimum system and power module design. Software design includes main program design, initialization program design, motor driver design, CCD data processing program design. After hardware and software design and debugging, the automatic obstacle avoidance and tracking trolley control system has basically achieved the desired design goals and functions.

At present, the motion control accuracy of cutting arm is not high, and the automatic forming of section cutting has the problems of poor forming quality and large forming error, combined with the serious time-varying and uncertainty of the load change of the heading cutting arm mechanism, this paper proposes a method of using fuzzy neural PID controller to realize the motion control system of the cutting arm, the controller takes the working current deviation E and the variation rate EC of the electro-hydraulic proportional valve as the system control input variables, and carries on the simulation test to the electro-hydraulic proportional valve model, the experiment shows that the controller has strong robustness and can adjust the PID parameters adaptively when the parameters of the controlled plant change greatly, which meets the requirements of the motion control of the heading cutting arm, the accurate motion control of the cutting arm is realized, and the quality of the section cutting automatic forming is effectively improved.

This study aims to study intelligent control of automotive active suspension based on the BP neuronal control model. By analyzing the deficiency of the traditional active suspension system, the BP neuronal control model is introduced to improve the control accuracy and robustness of the system. In this study, the BP neuronal control model is used to simulate the dynamic behavior of an automotive active suspension system by building a neural network. By adjusting the parameters and structure of the neural network, the performance of the active suspension system is optimized. Experimental results show that intelligent control of active vehicle suspension based on BP neuronal control model can effectively improve vehicle driving stability and comfort. Compared to the traditional active suspension system, the new model offers significant advantages in terms of control accuracy and robustness. In addition, the experimental data also show that the neural control model can adjust the parameters of the suspension system in real time, to adapt to different road conditions and vehicle load changes. This research provides a new solution for intelligent control of automotive active suspension. Intelligent control of automotive active suspension based on BP neuronal control model has high control accuracy and robustness, and can adapt to different road conditions and vehicle load changes. Future research can further optimize the neural control model and improve its performance in practical applications.

As a frontier component in capturing wind energy, the wind turbine blade (WTB) plays a key role in the efficient performance of a wind turbine. The WTB will suffer various failures or different degrees of damage in its operation due to the harsh environmental conditions, thus the WTB fatigue reliability study is a critical topic. The WTB is a complex structure, composed of leading edge, trailing edge, leaf root, web, middle, main beam, etc., and the failures of each element is correlated. This paper proposes a WTB fatigue reliability modelling and assessment method with consideration of failure correlation. In particular, based on finite element method, the time to failure of each element is obtained and the lifetime distribution is determined. Then, Copula function is introduced to quantify and characterize the failure correlation of the elements, and a fatigue reliability model is constructed for the WTB. The effectiveness of the proposed model is explained by a contrastive analysis with the results of independence assumption.

A novel method for suppressing non-uniform reflection in differential confocal microimaging is proposed in this paper. The method is based on the differential confocal optical path and aims to eliminate the influence of non-uniform reflection perturbation on the surface contour imaging of highly stepped samples. This is achieved by dividing the squared difference between the pre-focus and post-focus signals of the differential confocal by the larger value of the squared pre-focus and post-focus signals. The effectiveness of the proposed method is demonstrated through theoretical analysis, simulation, and experimental verification. The results show that the method can successfully realize height imaging of height step samples. It is found that non-uniform reflection perturbation only affects the edge overshoot and position of the height profile imaging curve. This method can achieve high-precision, high-efficiency, and non-contact measurement of step samples with non-uniform reflection disturbances. This has significant implications for various applications requiring accurate surface contour imaging.

To automatically reconstruct the engineering drawings of shaft parts, a complete 3D reconstruction algorithm is proposed using DXF files as input. The algorithm proposes the data analysis method based on the view centerline, which can effectively analyze the engineering drawings of shaft parts containing flat keyways and through-holes, including view separation, outer contour extraction, and keyway information extraction and classification. Experiments prove that based on this data analysis method, the 3D reconstruction algorithm can automatically reconstruct the engineering drawings of shaft parts containing flat keyways, grooves, or through-holes compared with the existing methods.

Aiming at the texture processing and noise disturbance problems of simultaneous localization and mapping algorithm for dynamic scenes under the strong static assumption theory, Improved dynamic object tracking based on ant colony clustering (IACDOT) was proposed. The algorithm combines fractional differential and sparse optical flow algorithm to make full use of the weak texture gradient of the image. A dynamic feature search and selection strategy is designed to obtain ant colony clustering, which reduces motion interference and mismatching of dynamic and static features. The experimental results show that the algorithm not only realizes the adaptive selection of pixel gradient order, but also has a better ability to distinguish dynamic disturbance through the clustering of feature selection. It can effectively distinguish motion and static information while retaining more details of weak gradient feature optical flow. The algorithm has a good application prospect in simultaneous localization and mapping system.

When a ship explodes, it will cause serious damage to cabin equipment and shipboard personnel. It is one of the effective measures for shock wave protection to arrange water media in the closed cabin to form a water body. In order to analyse the attenuation effect of water medium on explosion shock wave, a model of shock wave propagation in a cabin with water body is established, and Euler method is used to describe the deformation and fragmentation of water body under the action of explosion shock wave. The calculation results show that water medium can effectively reduce the overpressure peak and quasi-static pressure of explosion shock wave, and the reduction effect increases with the increase of water thickness.

Path planning technology is one of the current hotspots of research in many intelligent technology fields, which has a broad application prospect and research value. With the rapid improvement of computer arithmetic power and the widespread popularization of automation, automatic driving has gradually become a hot topic of people's attention. Among them, path planning algorithms play a central role in the field of automatic driving. In this paper, firstly, it will summarize and introduce the development history and basic principles of the two basic path algorithms, Dijkstra algorithm and A* algorithm, list the advantages and disadvantages of these two path planning methods based on graph search and the application scenarios, and then analyze and compare Dijkstra algorithm and A* algorithm, discuss their advantages and disadvantages in the field of intelligent driving, and finally make an outlook of future development directions of the path planning technology in the field of intelligent driving. The research in this paper will provide certain reference value to the researchers of path planning algorithms for intelligent vehicles. It will be of great value to the research and application of new path planning algorithms based on Dijkstra and A* algorithms in the future.

The object features from the drone's perspective are poor, there is a lot of noise, the object scale changes drastically, and densely arranged objects, which brings huge challenges to object detection, and there are problems such as missed detection and false detection. Based on this, this article proposes an improved YOLOv5 UAV image object detection method. This method first introduces the RepVGG structure to deeply mine and enrich the semantic information of different features to alleviate the problems of poor object features and excessive noise in complex backgrounds; then it introduces the SKAttention attention mechanism to improve feature differences such as object occlusion, background interference, and multi-scale objects. representation ability under various circumstances; finally, the F-EIOU regression loss function is introduced to improve the regression speed and improve the recognition accuracy of noisy objects. Extensive experiments were conducted on the VisDrone2019 UAV aerial photography data set. Experimental results show that the average accuracy of the improved YOLOv5 (mAP@0.5) increased by 5.4%, and mAP@0.5:0.9 increased by 3.4%.

To address the problem of excessive dependence on the main control unit and poor adaptability of the mechanical claw movement, the paper proposes a feasible solution to the end-point visual perception of the mechanical arm, which improves the adaptability and intelligence of the robot. Based on the AI terminal Kendryte K210 module, the paper constructed an image sample training set, completed model training, and then integrated the module into the mechanical arm endpoint to achieve target recognition and synchronous mechanical arm movements. The experiments show that the mechanical claw designed with this solution can perceive the spatial position of the target object based on target recognition, achieving the goal of "not dropping heavy objects and not breaking fragile objects," making target recognition more flexible in the mechanical hand.

The rapid development of the energy internet, the deep integration of the power system and information system, and the emergence of various new services in the distribution network have led to an explosion of service data. Traditional routing optimization methods are inadequate in satisfying the efficiency and reliability requirements of the power communication network for data transmission. This paper presents a collaborative optimization method for enhancing the traffic in the power communication network based on segmentation learning. Firstly, we propose a power service data transmission routing architecture in EPCN, where multiple routes between the source and destination nodes are available. Secondly, a segmentation learning-based traffic collaboration optimization algorithm for EPCN is proposed, which divides the traffic to explore the transmission performance of multiple routes within one optimization, thereby reducing data congestion at critical nodes. Finally, simulation results demonstrate that the proposed algorithm outperforms in terms of data transmission delay and routing optimization speed.

UAV aerial image object detection is of great significance for intelligent target identification and tracking, but the target under the UAV viewpoint is subject to large changes in target scale due to the influence of light, and there are cases of occlusion, low target resolution, etc., which lead to low model detection accuracy, misdetection, leakage and other problems. To address the above problems, an improved object detection method for UAV aerial images is proposed based on the YOLOv5. The method introduces Space-to-depth Convolution (SPD-Conv), Normalization-based Attention Module (NAM) and regression loss function, and conducts a large number of experiments on Visdrone2019 dataset. The experimental results show that the improved YOLOv5 algorithm improves the mean accuracy percentage (mAP) by 6.1% and the mAP@0.5:0.95 by 5%.

Anomaly detection of telemetry data is a promising way for ensuring the reliability of spacecraft and success of space mission. This study proposes a reconstruction model with adaptive weighting for unsupervised anomaly detection of polluted telemetry data. The reconstruction model integrates the advantages of typical reconstruction models and can generate high-quality results. An adaptive weighting module is designed to against the data pollution issue by assigning different weights to training data samples. Experiments are conducted on two public telemetry datasets, and the results verify the effectiveness of our method.

With the promotion of policies and the continuous development of artificial intelligence, autonomous vehicle have gained more and more attention. The environment perception system is the key to realize the real-time interaction between the vehicle and the external environment. It is also the first step to realize automatic driving, and plays an important role in the safe driving of autonomous vehicle. The development and application of the environmental perception algorithms cannot be separated from testing and evaluation. The test evaluation can effectively verify the accuracy and stability of the environment perception algorithm, and provide reliable input for the decision-making and control of the auto drive system. This article establishes a perception algorithm testing system based on data playback, and proposes a testing and evaluation method for perception algorithms based on the environmental information requirements of decision control systems, in order to improve the development efficiency of environmental perception algorithms.

With the rapid development of intelligence, the function of the whole vehicle is becoming more and more complex. More and more cameras, millimeter-wave radars, laser radars, ultrasonic sensors, and various other data acquisition devices are equipped on vehicles to support the evolution of autonomous driving capabilities. Compared with the wave of computing chip research and development in full swing, the development of computing chip testing and certification capabilities has seriously lagged behind. In this paper, a set of computing power test system is built based on TDA4VM development board. This method measures the actual computing power of the chip under the premise that the AI computing unit of the computing chip is squeezed to the maximum extent. It provides a new method for the calculation force test of the car-level computing chip. It forms an independent and controllable test and evaluation ability of the vehicle computing chip, and supports the chip selection of the car manufacturers.

The dual-phase-lag model has attracted increasing attention on account of its better heat transfer performance among many existing non-Fourier models. In this paper, we focus on the dual-phase-lag heat transfer process in a three-dimensional medium heated by a moving volumetric laser heat source. A series solution for the temperature distribution has been derived analytically by Green’s function approach. According to this solution, the effect of the phase lag parameters and the heat source moving speed on the temperature distribution are investigated. The present results show the temperature variation not only on the top surface but also inside the medium, which can help us better understand the obvious non- Fourier temperature response of the three-dimensional medium.

Piezotronics and piezo-phototronics are new emerging fields that have already found applications in nanogenerators, piezoelectric transistors, strain sensors and LEDs. The strain-induced piezoelectric potential plays a crucial role in modulating charge-carrier transport properties in piezoelectric semiconductor materials, especially for two-dimensional (2D) materials. In this work, we theoretically study the piezotronic effect, circularly polarized light and Zeeman field’s impact on the modulation of valley and spin properties in monolayer (ML) MoS₂ nanoribbon. We create two ferromagnetic regions in the ML MoS₂ nanoribbon to simultaneously manipulate the transport properties of valley and spin using externally applied polarized light and strain. In the proposed junction of ML MoS₂, which consists of two barriers modulated by strain, light and Zeeman field, an intriguing coexistence of perfect valley and spin polarization emerges. Our investigation reveals that among these three effects, the huge piezoelectric effect plays a leading role.

In the manufacturing of printed circuit boards, due to production processes and other issues that can easily lead to defects in the circuit board. In order to improve the efficiency of circuit board defect detection, a defect detection algorithm for bare PCB based on improved YOLOv7-tiny is proposed. First, a new ELAN structure, New-ELAN, is proposed to replace the ELAN structure in the Head section, and the three detection heads in the Head section are reduced to two. Next, reconnecting the Neck structure and reducing the number of channels to reduce computation. The experimental results show that: under certain training conditions, the improved YOLOv7-tiny's mAP value reaches 93.9%, which is 4.8% higher than the original model. In addition, the speed and size of the improved model remain essentially the same. The improved model has better detection results.

Graph neural network technology is widely used in social recommendation system to learn the embedded representation of users and items in the process of user interest graph and user social graph propagation. To solve the existing problems, this paper proposes a depth graph convolution recommendation algorithm (DGCR) that integrates social and residual errors. It uses lightweight graph convolution to carry out convolution propagation for users and items respectively in two graphs about users, and introduces the idea of residual network for deep propagation to prevent over-smoothing. At the same time, a more flexible multi-layer perceptron is used for image fusion to reduce the information gap. Compared with the baseline algorithm on LastFM and Ciao datasets, the experimental results show that DGCR algorithm has a significant improvement in recommendation effectiveness and cold start problems.

Ship target detection at sea has important strategic significance in military activities, maritime security, and other aspects. Traditional image processing algorithms struggle to capture the various scales of ship features. In this paper, we propose an algorithm for ship target detection based on CBAM-YOLOv8. Firstly, spatial-to-depth convolution is used in the model's downsampling section instead of cross-stride convolution to improve feature utilization. Secondly, the CBAM (Convolutional Block Attention Module) attention mechanism is added to the deep layers of the model to fuse spatial and channel feature information. Finally, MPDIOU is used to replace the CIOU loss function, enhancing the extraction accuracy of detection boxes. Experimental results on a maritime target dataset show that the detection algorithm achieves a mAP value of 93.16% and a detection speed of approximately 134 FPS, meeting the requirements of real-time ship detection at sea and providing an effective technical reference for various maritime activities and tasks.

A 3D reconstruction network, TRC-MVSNet, is proposed to address the issue of poor 3D reconstruction of infrared multi-view. It combines Transformer feature matching and RC-MVSNet. To enhance the quality of the infrared image, a moving average filter method is employed to mitigate noise by averaging signals in a sliding window. Additionally, the Transformer network is utilized to improve the matching quality of infrared multi-view information by matching the information of multiple infrared views in the feature matching stage. Experimental results demonstrate that the TRCMVSNet 3D reconstruction model enhances accuracy by 0.03 and synthesis by 0.013 compared to the RC-MVSNet 3D reconstruction model on the DTU dataset. The TRC-MVSNet achieves high accuracy, efficiency, and minimal noise in the 3D reconstruction of infrared targets on the SYLU-IMD dataset, outperforming other methods.

This paper presents a design of Brushless motor control system based on FOC algorithm. The hardware is composed of Main Control Board, drive board, voltage stabilizing circuit, full-bridge Circuit of inverter and current sampling circuit. Through coordinate transformation, SVPWM and extended Kalman filter, the closed-loop control of the motor is completed. The system not only avoids the shortcomings of the traditional Brushless motor control algorithm, but also improves the operating efficiency of the motor system. In addition, FOC combined with no position sensor algorithm improves control efficiency, reduces process cost and expands application range. This case is integrated into the teaching process of motor, which improves students' interest in learning and engineering practice ability.

This article mainly studies the stress field distribution of the die-cutting machine under no-load conditions, and uses numerical simulation methods to couple and mesh the main body of the die-cutting machine. By setting boundary and initial conditions, and applying working load on the transmission shaft, the stress field distribution of the main part of the die-cutting machine was numerically simulated using COMSOL software. The results show that the stress in the main body of the die-cutting machine is mainly concentrated on key components such as the transmission shaft, bearings, and gears, and is far below the yield strength of the material. In addition, the four upper swing rods of the die-cutting machine bear significant stress under no-load conditions. By studying the stress field distribution of die-cutting machines, strong support can be provided for the optimization design and fault diagnosis of die-cutting machines, ensuring their safe operation and efficient production.

For the detection of hard hats, the human key point detection model and YOLO3 algorithm based on deep learning are adopted. In this paper, we will introduce a series of work required for the design model of helmet recognition, including the construction of model training environment, data set preparation, model training, model visualization operation and model recognition results. The identification model of safety hat in electric power construction site will be developed based on PyTorch. The following framework depicts the structure of a deep learning model framework based on PyTorch.

This study explores the application of long and short-term memory neural network based on sparrow search algorithm in gas monitoring technology. The SSA-LSTM model is proposed for quantitative monitoring of multi-component gases, and its effectiveness and superiority are verified by utilizing the new algorithm and through experiments, and effective data support is provided for timely warning in the military industry.

This paper introduces a support vector machine (SVM) multicomponent gas qualitative correction algorithm based on the whale optimization algorithm (WOA), which combines the advantages of WOA and SVM and aims to solve the cross-talk problem in multicomponent gas analysis. Through the steps of data preprocessing, feature extraction, training SVM model and prediction, the algorithm can effectively reduce the influence of cross-talk and improve the accuracy of measurement results. The experimental results show that the WOA-SVM algorithm has good robustness and generalization ability, and is an effective tool for cross-talk correction in multicomponent gas analysis.

The paper aims to enhance the recall rate of Webshell detection methods based on machine learning algorithms. It covers a broad research scope on Webshells, with a focus on Webshells as the research object. The research employs a combination of the Naive Bayes and Decision Tree methods, extracting and analyzing static features of Webshells. Significant conclusions regarding the improvement of recall rates are drawn from this analysis. Experimental validation conducted on publicly available datasets demonstrates that the proposed method significantly increases the recall rate of the Webshell detection model while maintaining a low false-positive rate. This improvement holds crucial significance for enhancing the security of web applications and servers, offering a feasible research approach in the field of Webshell detection.

In complex road environments, traditional algorithms for detecting objects often face issues such as misclassification and omission of distant targets and small objects like pedestrians. To overcome these challenges, a new object detection algorithm called SCK-YOLO has been introduced in this paper. This algorithm is an improvement over the model of YOLOv5s network and includes a tiny object Detection Laye. It replaces the computationally intensive C3 module with a more lightweight C2f module and uses the K-means algorithm for anchor box selection, replacing the original anchor box selection method. Experimental comparisons show that the proposed algorithm performs better on the Kitti dataset compared to the original YOLOv5s algorithm, with an increase of 1.59% in mAP@0.5 and a 3.99 % enhancement in mAP@0.5:0.95.

Reciprocating compressors are widely used in industrial fields, and the stable operation of their bearing components is crucial to the overall performance of the machine. However, as bearings are one of the components most prone to failure, their fault diagnosis is particularly important. The challenge in accurate diagnosis arises due to the fact that bearings typically operate in a stable state, resulting in a scarcity of abnormal data samples. This study focuses on the fault diagnosis of bearings in reciprocating compressors and proposes a method based on Generative Adversarial Networks (GAN). By simulating real fault data, GAN can generate a large number of synthetic fault samples, addressing the issue of data imbalance. These synthetic samples are combined with real normal operating data samples to form a more balanced dataset for training a neural network classifier. Experimental results validate the effectiveness of this method in enhancing the fault diagnosis performance of reciprocating compressor bearings, demonstrating its immense potential in industrial applications.

TrueGrid and LS-DYNA are used to study water pressure blasting, and a shallow borehole blasting rock model is established. The influence of water in the borehole on the rock blasting effect is analyzed by damage coefficient cloud map and stress-time diagram. The results show that adding water to the borehole can increase the degree of rock breakage and reduce the cost of rock blasting effectively.

When the telemetry data noise level is high, the traditional integral moving average processing model has a general effect on noise reduction. In order to improve the noise reduction ability and obtain higher load identification accuracy, a new noise reduction model is constructed based on the Chebyshev orthogonal polynomial and the idea of piecewise overlap. The impact load is simulated by half sine wave. Under the condition of 15 % noise, the error values of load identification results are 1.94 % and 0.76 % respectively after the treatment of integral sliding noise reduction model and Chebyshev noise reduction model. In contrast, the new Chebyshev noise reduction model under high-level noise conditions has better accuracy and wider application.

As a common battery for new energy vehicles at home and abroad, the capacity loss of lithium-ion battery has always been the focus of attention. Accurately identifying the maximum available capacity of the battery in the actual use process is a key point and difficulty in the current development of power battery technology. In this paper, the influence of the voltage, current and temperature parameters of power battery on the available capacity of lithium-ion battery is explained, and predicted by ARIMA model. After obtaining lithium battery measurement data set in NASA, ADF single root test and difference method were used to stabilize the raw data of lithium battery capacity, and to evaluate the effectiveness of various estimated parameters. After strict evaluation, ARIMA (1,1,1) was identified as the best fit model, and relatively accurate prediction results were obtained. The results show that the ARIMA model prediction method can more accurately predict the available capacity of lithium battery in electric vehicles.

With the rapid development of the energy Internet, the power industry has ushered in the era of power big data. Through dynamic monitoring and analysis of all kinds of data in power grid operation, potential risks can be found effectively and timely, and the safety of power grid operation can be improved. As an important part of power big data, user-side power consumption data provides strong support for managers to visually display users' abnormal power consumption behaviors. It provides a theoretical basis for better and faster solution to the behavior of electricity leakage due to equipment failure or personal behavior. Based on this background, this paper first uses k-means clustering algorithm to classify electricity consumption data according to user usage habits. This method can effectively avoid misjudgment and missing judgment of electricity data. Finally, the method of combining LOF and isolation forest is used to realize the anomaly detection of user side power consumption data. By comparing the single algorithm (LOF and isolation forest), it is proved that the anomaly monitoring model combined with the two algorithms can better identify the abnormal power consumption data at the user side.

In data processing, the removal of random noise is an essential step in order to improve quality. This article presents a detailed study on the removal of random noise in seismic signals, aiming to seek a more efficient and practical denoising method. This paper proposes a seismic data denoising method based on the K-SVD algorithm. By performing sparse decomposition and dictionary learning on seismic signals, the method achieves effective denoising of seismic signals. The results show that this method has a good denoising effect on seismic data.

Previous means of controlling pests and diseases were relatively outdated and did not improve the yield and efficiency of farmland very much. In order to be able to monitor the degree of pests and diseases of planted crops such as wheat and corn in real time and to increase robustness, this paper proposes a pest detection model YOLOv5s-ECA based on the improved YOLOv5 algorithm, which introduces an attention mechanism, ECA, into the backbone network of YOLOv5s in order to achieve the enhancement of the network's capability of extracting image features while there is no increase in the model's parameters and volume. Thus, the purpose of improving the accuracy of detecting the target is achieved. In order to verify the performance of the model, we built a pest and disease dataset, and the training results show that YOLOv5-ECA improves the detection mAP by 3.1% and Precision by 4.1%, and the detection results are better than other detection algorithms.

Deep learning has brought about a revolutionary transformation in the field of human pose estimation, offering significant advantages. Traditional approaches often seek performance improvement by expanding and deepening networks, resulting in increased parameters and complexity. In response to this challenge, we introduce DNNet, a novel framework rooted in HRNet. Its unique basic blocks incorporate feature mapping, channel weighting, and the fusion of different receptive fields. In comparison to HRNet, DNNet stands out with fewer parameters, lower computational complexity, and more resilient lightweight features. By leveraging features at various resolutions, DNNet's performance is further enhanced. Extensive validation on both the COCO dataset and a dataset focused on dangerous driving behavior reveals that DNNet's accuracy is comparable to that of HRNet. Notably, DNNet achieves a remarkable 69% reduction in parameters and a 59% decrease in computational complexity under similar accuracy conditions. In practical real-world applications, when applied to a dataset addressing dangerous behavior, DNNet outperforms both ShuffleNet and MobileNet in accuracy, highlighting its adaptability and efficacy in diverse scenarios. DNNet's unique features position it as a promising solution in the field of human pose estimation, providing a balanced trade-off between accuracy, efficiency, practical applicability, and the organic fusion of different receptive fields.

Removing noise from images is a challenging task in front-end computer vision applications, especially in snowy, rainy, foggy, and underwater environments, which pose significant difficulties for various visual tasks. Existing image restoration methods often struggle to achieve robust generalization and are burdened with immense computational complexity. This paper introduces a lightweight image restoration network based on polarized attention mechanisms and efficient feature extraction, capable of addressing tasks like rain removal, snow removal, fog removal, and underwater image enhancement. Initially, a polarized self-attention mechanism is proposed to intricately learn weather noise, followed by an efficient noise removal module complementing the initial network. This addition compensates for any noise information overlooked by the polarized self-attention mechanism. Furthermore, residual connections are integrated into each module of the network, along with grouped convolutions, to prevent network degradation and reduce computational load. In comparison to existing models, this network not only learns noise with similar shapes but also adapts to noise with significant shape variations. Additionally, the introduced residual structure ensures the network's stability, avoiding performance degradation before achieving expected results. The collaborative use of the efficient noise removal module and polarized self-attention mechanism sets a high standard for noise learning in this network. Experimental results demonstrate the efficacy of this algorithm on publicly available datasets including Snow100K, CSD, Rain200H, Rain200L, Rain800, RSID, and EUVP. The proposed method efficiently removes various image noises, effectively achieving the desired image restoration objectives.

Metaverse is a virtual blockchain technology-based world that combines real-world data and digital assets with a virtual world. The emergence of Metaverse provides new methods and ideas for the physical design of fusion ignition. This paper focuses on the concept, characteristics, and application of meta-universe in the physical design of fusion ignition. This paper firstly introduces the basic principles and challenges of fusion ignition, then details the concepts and technical characteristics of meta-universe, and finally describes the application of meta-universe in Z-FFR (Z-pinch driven fusion-fission hybrid reactor) physical design, including data sharing, virtual simulation, intelligent contracts, and other aspects. In this study, the performance of the device was significantly improved by a meta-universe-driven design. The design cycle was dramatically shortened, the device durability was increased by a factor of 10, 000, and the delay jitter was reduced to 1-2 ns. At the same time, the neutron yield was increased by more than two orders of magnitude in terms of fusion physics. This paper argues that the emergence of meta-universe will further promote the physical design of fusion ignition and improve design efficiency and reliability.

The calibration methods of robotic arms generally have problems such as high cost and low efficiency. Taking the end position measurement link of the robotic arm, which is the most time-consuming and costly among them, as an entry point, a high-precision and real-time visual monitoring method of end position based on deep learning is proposed. Firstly, the influence of the complex error sources of the vision system on the end target position measurement results is analyzed. Secondly, a neural network structure is established based on this, whose input feature is the target position containing the measurement error, and the output label is the accurately computed end joint position. Finally, the training samples and the test samples are generated in the neighborhood space of the specified monitoring point, respectively. Simulation results show that the trained network structure can achieve a prediction accuracy of 0.025 mm on over 99% of the test samples, which is comparable to the laser tracker ranging accuracy. The new method combines the powerful learning capability of neural networks with the complex error sources of vision systems to enable low-cost machine vision methods to achieve high accuracy measurements.

In view of the fact that most of the existing research on the pitting failure of ball screw focuses on the vibration signal experiment and the establishment of the degradation model, and the use of more intuitive visual aspects is less, the deep learning is studied in the pitting detection of the surface of the ball screw. The application of Faster R-CNN and Mask R-CNN two network models were built, and the two were compared and analyzed through experiments. The results show that both Faster R-CNN and Mask R-CNN can guarantee high classification accuracy under different learning rates, and both can excellently complete the detection task of pitting on the surface of the ball screw. While locating the eclipse, the mask of the pitting is output synchronously, which is helpful in the face of the subsequent task of describing the size of the pitting, and has more advantages in the face of small-area pitting, and there are fewer missed detections.

Since the 21st century, unmanned intelligent equipment has been widely used in local battlefields, replacing humans to complete "dangerous, repetitive and boring" combat and support tasks, fully demonstrating the advantages of flexibility, convenience, low cost and high efficiency, and achieving good results. In order to promote the development of unmanned intelligent support equipment of Chinese army, the development and application of unmanned intelligent support equipment of military powers are systematically reviewed, and the development experience is summarized. Suggestions for the development of unmanned intelligent support equipment are put forward from four aspects: carrying out theoretical research and test verification, formulating strategic planning and standards, deepening civil-military integration and technology joint research, and strengthening technical security and information confidentiality.

In robotics, achieving stable control of unicycles presents a significant challenge. However, standard balance control methods concentrate on regulating the position and static balance of the robot. Such techniques exhibit diminished performance in the presence of external disruptions or uncertain terrain. To enhance the stability and resilience of unicycles in dynamic settings, this research presents a double-flywheel unicycle robot that utilizes the conservation of angular momentum in a dual-gyroscope system to maintain its balance. Through the integration of mechanical design, electronic design, and controller design, a prototype of the unicycle robot is fabricated. For adaptive balance control, a cascaded PID controller based on MATLAB dynamic simulation is developed. The cascaded PID controller comprises angle, velocity, and angular velocity loops, enabling non-linear and coupled balance control. Moreover, using MECHANICS EXPLORERS, the disturbance rejection capacity of the designed cascaded PID controller is confirmed through pulse interference simulation. This study uses exercises that traverse ramps and maneuver corners to demonstrate the unicycle robot's dynamic balance. This exemplified the resilience of the entire balance system, demonstrating its capacity to recuperate from disruptions and uphold stability. The unicycle robot, in conjunction with the cascaded PID controller, can enhance response speed and achieve high-precision balance control performance, enabling stable motion in complex environments. This study significantly guides the development of newfound intelligent, flexible, and adaptable unicycle robots, providing novel ideas and approaches for future robot design and control.

Many factors cause false or missed detection of small UAV objects, such as large changes in object scale due to illumination, dense objects, complex backgrounds and occlusions that lead to low model detection accuracy. To solve the above problems, an improved UAV small object detection method is proposed, based on the YOLOv5.Replace the original conv2d detection head with Adaptive Spatial Feature Fusion, add Attentional Convolutional Mixtures, and replace the original regression loss function with F-EIOU. Extensive experiments are conducted on the Visdrone2019 dataset. The experimental results show that the improved YOLOv5 increases the mAP@0.5 by 6.1%, the mAP@0.5:0.95 by 2.7%, the recall by 5.7% and the precision by 3.2% on the Visdrone2019 dataset, meeting the practical needs of UAV small object detection in complex scenarios.

When an underwater robot performs a task, the propeller is most likely to malfunction, such as being entangled by foreign objects or the blades are damaged. At present, its fault diagnosis methods have problems such as relying on manual feature extraction and using neural networks with low accuracy. Therefore, this paper proposes an integration based on an improved one-dimensional convolutional neural network (1D-CNN) and a long short-term memory network (LSTM). Thruster fault diagnosis method. By analyzing thruster data, accurate diagnosis of four different thruster faults can be achieved. A comparative experiment was conducted between the proposed model and some traditional algorithm models. The results show that the proposed method has greatly improved the test accuracy, and this method can effectively diagnose underwater robot thruster faults.

Analyzing the characteristics of anti-armor operations involving EFP warhead-equipped rotary-wing unmanned drones, this study focuses on the dynamic factors affecting the dispersion accuracy of EFP warhead destruction element. It establishes a dynamic impact point model based on the seeker system detection error, drone's spatial attitude disturbances, warhead destruction element convected motion deviation, and system latency. This model aims to utilizes experimental flight data to analyze the impact of each factor on the point of impact. Simulation and analysis results reveal that, under certain disturbances, the primary factor causing aiming error is the drone's body attitude disturbances. The study concludes by determining the attitude control requirements necessary to achieve the desired hit rate on typical armored targets.

Due to the constraints of crane positioning, its behavior accuracy is relatively low. Therefore, the research on automatic positioning of crane for offshore heavy parts based on BIM and RFID technology is proposed. This paper makes a comprehensive analysis of the positioning constraints of offshore heavy lifting cranes from three perspectives: environmental constraints, operational constraints and safety constraints, and determines the corresponding control elements. In the positioning process, BIM and RFID technology are introduced, and the actual running state data of the hoisting crane is collected with the help of RFID technology. Based on the analysis results of the positioning constraint control elements of the hoisting crane with heavy offshore parts, the identification tag of RFID technology is set up. BIM model is used to decompose the scope of construction area and hoisting tasks in the early stage of hoisting, which is coupled with the data acquisition results of RFID technology to realize the location of specific positions. In the test results, the positioning error of the offshore heavy lifting crane in the X direction is stable within 0.4m, and the positioning error of the offshore heavy lifting crane in the Y direction is stable within 0.3m, which has high accuracy.

This paper presents a trajectory tracking approach for a three-wheel independent steering (3WIS) robot using a fuzzy PID controller. The kinematic model of the 3WIS robot is introduced, highlighting its capability for omnidirectional motion control. A comparison is made between traditional PID controllers and fuzzy PID controllers, emphasizing the advantages of the latter in handling nonlinear systems. The concept of fuzzy PID control is explained, and fuzzy rules for parameter adjustment are proposed. The accuracy of the kinematic model and the effectiveness of the fuzzy PID controller are verified through simulation using the Webots platform and real-world experiment. Results indicate successful trajectory tracking, although slight fluctuations in the heading angle are observed due to frictional torque and gear backlash. The findings validate the accuracy of the kinematic model and the effectiveness of fuzzy PID controller in trajectory tracking.

In order to improve the efficiency and accuracy of fault diagnosis of electric vehicle drive motor, this paper proposes a study on fault diagnosis model of electric vehicle based on learning algorithm. The system uses sensors to collect the data of motor operation state, and after fuzzy processing, the knowledge base and database are mobilized according to fuzzy rules to diagnose the motor operation state. The test results show that the accuracy of the system in motor fault diagnosis is 99.8%, and the response time is controlled within 200μs, which greatly improves the performance of the traditional system. In the four tests, the fault diagnosis time was less than 200μs, which was about 100μs shorter. In order to test the accuracy of system fault diagnosis, 1000 simulation tests were carried out in this test. The test results show that only two diagnosis results are deviated, and the problem appears in the speed data, which leads to the wrong simulation diagnosis results. Therefore, the current diagnostic accuracy of the system is 99.9%, which is significantly improved compared with the traditional diagnostic system. The validity and reliability of the method proposed in this paper are verified.

As a biometric recognition technology, face recognition has the characteristics of universality, high reliability and strong individual differences, and has broad application prospects in the field of smart security. According to the needs of the access control and attendance system in the construction of the school's smart campus, this paper applies the face recognition algorithm based on deep learning to the face recognition access control and attendance system, and makes lightweight improvements to the algorithm to address the common problem of large amounts of calculation. This paper uses the improved RetainNet face detection model to design and implement an automatic attendance system based on face recognition for the classroom scene. Based on the classroom surveillance video stream, face detection is first performed, and the face filtering method is used after obtaining the face image set. Eliminate face images with low face quality and successfully recognized positions, then perform face super-resolution and face alignment, and finally send them to the face recognition model for face comparison to complete attendance. Experiments show that the improvements proposed in this article effectively improve the accuracy of classroom face detection.

This paper proposes a design and implementation method for power patrol robot based on virtual orbits to address issues such as the risk of dislocation during autonomous navigation and the difficulty in accurately limiting areas through remote temperature measurement. Using Three-dimensional accurate mapping is used to establish the precise operating environment of power patrol robot. Robot operation safety zone is established, robot operation virtual track is formed, and safety zone determination algorithm of power patrol robot is formed based on robot appearance and walking path. According to substation three-dimensional simulation and robot wireless localization measurement, the power patrol robot proposed in this paper can form a virtual track on the basis of wireless localization, and can reliably monitor the operation safety under 3D accurate mapping environment, which greatly reduces the risk of dislocation during the autonomous navigation.

Historical arch bridges in China have experienced a gradual deterioration in structural integrity due to prolonged use and increasing traffic loads. Ensuring the safety of the arch bridge as well as the pedestrian demands precise health assessment of these bridges, which normally requires finite element (FE) modeling and updating. To facilitate the FE modeling with limited information, a point cloud-based automatic FE modeling strategy is explored in this study. The unmanned aerial vehicle is first employed to do an on-site survey via images and videos to provide the geometric data of the historical arch bridge. They are then transformed into point cloud data to reconstruct a 3D solid model of the bridge. The FE mesh is finally implemented with element properties assigned to obtain the targeted FE model. An ambient vibration test is also conducted to extract the structural modal parameters, for further verification of the constructed FE model and possibly FE model updating. The proposed strategy is based on commonly available software, and thus easy to implement. It offers a rapid means of building the FE model, laying the groundwork for the condition assessment of the historical arch bridges.

Manual detection of insulators on UHV lines is a significant security risk. Therefore, it is practical and valuable to develop an insulator inspection robot. At present, there are few reports on the application of robots in the detection of UHV lines. Based on the live detection test of 1000kV transmission line insulators, the communication system test, the detection system test, the image transmission system test, and the appearance arc test were conducted in the high-voltage laboratory. The paper studied the insulation level test of the robot and explored the influence of probe length and position on arc discharge. A safe area for live robot operation was obtained.

Under the influence of unstructured environmental constraints, the spatial morphological changes in slender soft robots are primarily determined by the robot's force situation. Therefore, this paper focused on theoretically modeling and analyzing the static equilibrium problem of slender soft robots. The static mechanical model of the robot under unstructured elastic constraints was simplified, established, and computationally solved based on the theory of elastic rods. Simulation calculations for the model under General plane elastic constraints were conducted, providing an initial verification of the accuracy.

Aiming at the problem of poor flexibility of waist rehabilitation robot in the process of rehabilitation movement, we analyzed the characteristics of waist movement and rehabilitation training, and proposed a rigid-flexible coupling configuration of waist rehabilitation robot with certain flexibility by integrating the theory of sports rehabilitation. The structure of the waist training device is introduced and the role of the waist in the waist training device is analyzed, and the kinematic model of the waist training device is established. The correctness of the kinematic analysis results is verified through simulation, and based on the kinematic simulation results of the lumbar training device, a rigid-flexible coupled lumbar rehabilitation robot test platform is further designed and constructed.

In the aviation field, accurate aircraft detection is of great significance for traffic monitoring, safety assurance, and military reconnaissance. Although traditional convolutional neural networks have achieved significant success in image recognition and object detection, they still face challenges in processing aerial images containing multi-scale and complex backgrounds. To address these issues, this study proposes a Vision Transformer based model that incorporates SKAttention (Selective Kernel Attention) and MSCAM (Multi Scale Channel Attention Module) technologies to improve the accuracy and efficiency of aircraft detection.

SKAttention technology effectively enhances the flexibility and accuracy of the model in processing aircraft images of different scales by adaptively selecting the most suitable convolution kernel size. MSCAM, on the other hand, optimizes the model's ability to process aircraft details and background information by enhancing channel attention at different scales. By combining these two methods into the Vision Transformer architecture, our model achieved accuracy and recall of 98.7% and 98.2%, respectively. These results validate the effectiveness of SKAttention and MSCAM in improving the performance of aviation aircraft detection based on Vision Transformer, providing new technological approaches and research directions for aviation image processing and object detection.

As a key component of circuit breaker switch, the dynamic characteristics of spring mechanism are directly related to the response speed and stability of high voltage circuit breaker switch. Therefore, based on deep learning technology, this paper makes a dynamic simulation analysis of the switch spring mechanism of high voltage circuit breaker. Firstly, the operational data is decomposed into multiple IMF modal components using Empirical Mode Decomposition (IMF) method, and correlation coefficients are calculated to extract features. Secondly, using the Long Short Term Memory Network (LSTM) model, the feature inputs are learned and trained, and the dynamic mathematical model results of the spring mechanism are output. The forget gate and input gate in the model respectively handle information retention and update, achieving accurate simulation of dynamic simulation. The update of weights adopts the adaptive momentum estimation gradient optimization algorithm, which improves the accuracy of the model. By comparing the simulation results of traditional methods and deep learning methods, the superiority of deep learning algorithms in the dynamic simulation of high-voltage circuit breaker switch spring mechanisms was verified. By comparing the simulation results of traditional methods and deep learning methods, the superiority of deep learning algorithms in the dynamic simulation of high-voltage circuit breaker switch spring mechanisms was verified.

In order to improve the prediction accuracy of high slope deformation data, a high slope deformation data prediction model combining autoregressive integrated moving average model (ARIMA), particle swarm optimization algorithm (PSO) and generalized regression neural network (GRNN) is proposed. Considering the nonlinearity and complexity of high slope deformation data, the model uses the ARIMA model for linear prediction, and the PSO-GRNN model corrects the residuals of the ARIMA model for nonlinear prediction. The results show that, by comparing with many prediction models, the residual-corrected ARIMA-PSO-GRNN model has the highest prediction accuracy, with MSE, MAE, and MRE of 0.0500, 0.1373, and 0.8285%, respectively. Using this model for prediction in practical work can provide scientific basis and decision support for related personnel.

The accurate measurement of rail wear is critical in track quality inspection. Usually, the wear is obtained by matching the measured profile with the standard one, and comparing their railhead differences. However, in the complex field environment, the process becomes difficult with lots of outliers mixed in the profile. The outliers not only mislead the location of rail waist, also could cause serious measurement errors. Considering that the global geometry of rail profile is prior and stable, a hybrid model including a residual dilated Convolutional Neural Network (CNN) and a bidirectional Recurrent Neural Network (RNN) is constructed in this paper. The former is used for extracting the local features to detect sparse outliers, and the latter is used for extracting the global geometry to detect dense outlier segments. The efficiency and superiority of the proposed method were verified by numerous experiments. The results show that the average F1-score for outlier detection reaches 99.54%, which outperforms some classical neutral network models obviously. Meanwhile, it omits the registration process, and the real-time performance is also improved obviously.

To address the issues of manual feeding and placement in the processing of fresh corn ears, this study develops a fresh corn ear detection system by using machine vision and deep learning techniques. The hardware setup and acquisition of clear image data were completed. The G channel image in the RGB channel was selected for subsequent image processing. Contrast enhancement and image sharpening techniques were employed to improve the clarity of the ear region. Additionally, a median filtering method was used to eliminate noise caused by lighting conditions. Morphological processing and Otsu threshold segmentation techniques were applied to extract complete corn ear regions. Based on the contour feature, three kinds of feature extraction of ear shape, gray value, and ear texture are selected. By comparison, the ear texture feature recognition algorithm with higher recognition accuracy and shorter detection time was selected. The algorithm was used to detect the ear handle region, and the cutting position was predicted in this region. This paper studied the method of predicting the cutting position by using the slope change feature of the spikelet region contour and used C# and Halcon mixed programming to realize the automatic detection of the ear. Experimental results show that the detection accuracy of the spikelet region can reach 98.23% by using the texture feature detection algorithm, and the detection time of a single ear is 23.90 ms.

Due to the multiple factors that affect the lifespan of electronic components, it is difficult to accurately predict the lifespan of electronic components by directly analyzing a single factor. Therefore, a study on the lifespan prediction method of electronic components based on dual channel multimodal convolutional networks is proposed. Based on the failure mechanism of electronic components, degradation models of electronic component electrical characteristic parameters were constructed from the perspectives of resistance and capacitance. Two parallel convolutional channels are used to extract continuous samples with uniform sampling time intervals. The degradation characteristics of electronic component electrical characteristic parameters at different time points extracted by the two parallel convolutional channels are fused, so that the final extracted electronic component features have temporal information. Multimodal fitting is performed on the electronic component features at the output of the convolutional network to obtain the final life prediction result. In the test results, the prediction error of the design method is stable within 6.0%, the time cost is stable within 21.0s, and fast convergence can be achieved.

Neural Radiance Field (NeRF), a compelling technique in the field of Computer Vision, represents a novel approach to view synthesis using implicit scene representation. By learning to represent 3D scenes from images, its goal is to render photorealistic images of the scene from unobserved viewpoints, showcasing the immense potential of neural volumetric representations. As a novel view synthesis and 3D reconstruction method, NeRF models find applications in robotics, urban mapping, autonomous navigation, virtual reality augmented reality, and more. In this article, we introduce a new 3D surface representation method called Signed Distance Function Field (SDF). We have developed a new volume rendering technique for training a neural SDF representation. In our research process, we noticed that traditional volume rendering methods have poor imaging performance in complex structures and self occluding images during surface reconstruction. Therefore, we have designed a new neural network to reduce the impact of complex structures and self occluding objects on 3D reconstruction. This results in more precise surface reconstruction, even in the absence of mask supervision. Our experiments, conducted on both the DTU dataset, demonstrate that this superiority is particularly evident when dealing with objects and scenes characterized by intricate structures and self-occlusion.

In response to the development needs of future unmanned platform cluster collaborative operation mode, with the goal of improving the collaborative operation capability of two-domain unmanned platforms, an analysis of collaborative perception technology for two-domain unmanned operation platform clusters has been carried out. This article proposes a dynamic target tracking algorithm based on BACF algorithm, which combines weighted fusion of multiple features and model adaptive updating. Assign weights based on the contribution of multiple features and perform weighted fusion. Use the fused features to detect the tracked target. By introducing a one-dimensional scale filter, achieve adaptive changes in target scale; This algorithm can solve the problems of interference such as water mist, high noise, occlusion, and lighting in two-domain work environments, as well as the problem of insufficient robustness and real-time requirements in the tracking process. It can effectively improve the cluster operation ability of two-domain unmanned platforms.

With the advancement of science and technology, the rapid development of the autonomous driving industry has put forward higher requirements for related algorithms. To improve the reliability and intelligence of autonomous vehicles, it is necessary to have robust and reliable decision-making module, which depends on accurate trajectory prediction. In this paper, research on multi-modal trajectory prediction and decision-making for autonomous driving based on deep learning is carried out. Firstly, a multi-modal trajectory prediction model is constructed based on graph neural network to obtain the predicted trajectories of vehicles around the autonomous vehicle. Based on the prediction results, a decision-making network model of the autonomous vehicle is constructed, and the optimal decision-making results satisfying the multi-objective requirements are obtained by combining the multi-objective optimization method. The experimental results verified the feasibility of the method and its good performance.

Visual inspection was an important technology for the roadside perception of vehicle-road cooperative. For the difficulty of keeping balance between detection precision and computation efficiency by common visual perception algorithms, a new visual processing method based on YOLOX and KPP-DeepSORT was presented. First, YOLOX was used for multi-channel image recognition, then considering the orderliness of the traffic stream, K-Means++ was introduced to preprocess DeepSORT for reducing tracking delay. Data showed that the proposed method had high accuracy for detecting and tracking the pedestrians and vehicles. Especially in rush-hour crossroad, it performed more efficiently than some other common algorithms, so that had extensive application prospect in Internet of Vehicles.

In embedded systems, precision MEMS inclinometers often experience temperature drift due to environmental changes during operation. This paper introduces Kalman filtering as a preprocessing step and combines it with neural network-based temperature compensation methods. Experimental verification shows that within the range of 0-50 degrees Celsius, the signal quality retention rate is 97.2%, the signal-to-noise ratio reaches 21.68 dB, and the temperature drift phenomenon is reduced by 85.96%, ensuring the effectiveness and feasibility of this method.

This study delves into the application of artificial intelligence in the stability analysis of loess tunnels, with a focus on the principles, structure, advantages, and disadvantages of the BP neural network model and the optimization of this network using genetic algorithms. The research employs a genetic algorithm-optimized BP neural network method, utilizing numerical simulation results as training samples. Key factors such as tunnel radius, distance between tunnels, angle, deformation modulus of the surrounding rock, cohesion, internal friction angle, Poisson's ratio, and density are selected as input parameters for the neural network. The study successfully constructs a GA-BP neural network prediction model, which demonstrates excellent performance in convergence speed and prediction accuracy. This achievement not only validates the effectiveness of genetic algorithm-optimized BP neural networks in loess tunnel stability analysis but also offers a new analytical and predictive tool for related fields. The application of this model allows for more accurate prediction and analysis of tunnel stability, providing scientific decision support for tunnel design and construction, thereby enhancing the safety and reliability of tunnel engineering.