Attention Mechanism-Based Method for Intrusion Target Recognition in Railway

The detection of foreign object intrusion is crucial for ensuring the safety of railway operations.To address challenges such as low efficiency,suboptimal detection accuracy,and slow detection speed inherent in conventional comprehensive video monitoring systems for railways,a railway foreign object intrusion recognition and detection system is conceived and implemented using edge computing and deep learning technologies.In a bid to raise detection accuracy,the convolutional block attention module(CBAM),including spatial and channel attention modules,is seamlessly integrated into the YOLOv5 model,giving rise to the CBAM-YOLOv5 model.Furthermore,the distance intersection-over-union_non-maximum suppression(DIo U_NMS)algorithm is employed in lieu of the weighted nonmaximum suppression algorithm,resulting in improved detection performance for intrusive targets.To accelerate detection speed,the model undergoes pruning based on the batch normalization(BN)layer,and Tensor RT inference acceleration techniques are employed,culminating in the successful deployment of the algorithm on edge devices.The CBAM-YOLOv5 model exhibits a notable 2.1%enhancement in detection accuracy when evaluated on a selfconstructed railway dataset,achieving 95.0%for mean average precision(m AP).Furthermore,the inference speed on edge devices attains a commendable 15 frame/s.

Numerical investigation on MHD Jeffery-Hamel nanofluid flow with different nanoparticles using fuzzy extension of generalized dual parametric homotopy algorithm

This study considers an MHD Jeffery-Hamel nanofluid flow with distinct nanoparticles such as copper,Al_(2)O_(3)and SiO_(2)between two rigid non-parallel plane walls with the fuzzy extension of the generalized dual parametric homotopy algorithm.The nanofluids have been formulated to enhance the thermophysical characteristics of fluids,including thermal diffusivity,conductivity,convective heat transfer coefficients and viscosity.Due to the presence of distinct nanofluids,a change in the value of volume fraction occurs that influences the velocity profiles of the flow.The short value of nanoparticles volume fraction is considered an uncertain parameter and represented in a triangular fuzzy number range among[0.0,0.1,0.2].A novel generalized dual parametric homotopy algorithm with fuzzy extension is used here to study the fuzzy velocities at various channel positions.Finally,the effectiveness of the proposed approach has been demonstrated through a comparison with the available results in the crisp case.

Effect of gas saturation on P-wave velocity in tight sandstone

By measuring the variation of the P-and S-wave velocities of tight sandstone samples under water saturation,it was confirmed that with the decrease in water saturation,the P-wave velocity first decreased and then increased.The variation in velocity was influenced by the sandstone’s porosity.The commonly used Gassmann equation based on fluid substitution theory was studied.Comparing the calculated results with the measured data,it was found that the Gassmann equation agreed well with the measured data at high water saturation,but it could not explain the bending phenomenon of P-wave velocity at low saturation.This indicated that these equations could not accurately describe the relationship between fluid content and rock acoustic velocity.The reasons for this phenomenon were discussed through Taylor’s expansion.The coefficients of the fitting formula were calculated and verified by fitting the measured acoustic velocity changes of the cores.The relationship between P-wave velocity and saturation was discussed,which provides experimental support for calculating saturation using seismic and acoustic logging data.

Influence of pier height on the safety of trains running on high-speed railway bridges during earthquakes

Sudden earthquakes pose a threat to the running safety of trains on high-speed railway bridges,and the stiffness of piers is one of the factors affecting the dynamic response of train-track-bridge system.In this paper,a experiment of a train running on a high-speed railway bridge is performed based on a dynamic experiment system,and the corresponding numerical model is established.The reliability of the numerical model is verified by experiments.Then,the experiment and numerical data are analyzed to reveal the pier height effects on the running safety of trains on bridges.The results show that when the pier height changes,the frequency of the bridge below the 30 m pier height changes greater;the increase of pier height causes the transverse fundamental frequency of the bridge close to that of the train,and the shaking angle and lateral displacement of the train are the largest for bridge with 50 m pier,which increases the risk of derailment;with the pier height increases from 8 m to 50 m,the derailment coefficient obtained by numerical simulations increases by 75% on average,and the spectral intensity obtained by experiments increases by 120% on average,two indicators exhibit logarithmic variation.

Research on the LA-UMamba Model for Asymmetric Modules with Added Auxiliary Information

Deep learning techniques are revolutionizing the developmentof medical image segmentation.With the advancement of Transformer models,especially ViT and Swin-Transformer,which enhances the remote-dependent modeling capability of the model through the self-attention mechanism,better segmentation performance can be achieve.Moreover,the high computational cost of Transformer has motivated researchers to explore more efficient models,such as the Mamba model based on state-space modeling(SSM),and for the field of medical segmentation,reducing the number of model parameters is also necessary.In this study,a novel asymmetric model called LA-UMamba was proposed,which integrates visual Mamba module to efficiently capture complex visual features and remote dependencies.The classical design of U-Net was adopted in the upsampling phase to help reduce the number of references and recover more details.To mitigate the information loss problem,an auxiliary U-Net downsampling layer was designed to focus on sizing without extracting features,thus enhancing the protection of input information while maintaining the efficiency of the model.The experiments were conducted on the ACDC MRI cardiac segmentation dataset,and the results showed that the proposed LA-UMamba achieves proved performance compared to the baseline model in several evaluation metrics,such as IoU,Accuracy,Precision,HD and ASD,which improved that the model is successful in optimizing the detail processing and reducing the complexity of the model,providing a new perspective for further optimization of medical image segmentation techniques.

A thermo-mechanical damage constitutive model for deep rock considering brittleness-ductility transition characteristics

This paper developed a statistical damage constitutive model for deep rock by considering the effects of external load and thermal treatment temperature based on the distortion energy.The model parameters were determined through the extremum features of stress−strain curve.Subsequently,the model predictions were compared with experimental results of marble samples.It is found that when the treatment temperature rises,the coupling damage evolution curve shows an S-shape and the slope of ascending branch gradually decreases during the coupling damage evolution process.At a constant temperature,confining pressure can suppress the expansion of micro-fractures.As the confining pressure increases the rock exhibits ductility characteristics,and the shape of coupling damage curve changes from an S-shape into a quasi-parabolic shape.This model can well characterize the influence of high temperature on the mechanical properties of deep rock and its brittleness-ductility transition characteristics under confining pressure.Also,it is suitable for sandstone and granite,especially in predicting the pre-peak stage and peak stress of stress−strain curve under the coupling action of confining pressure and high temperature.The relevant results can provide a reference for further research on the constitutive relationship of rock-like materials and their engineering applications.

Gated Neural Network-Based Unsteady Aerodynamic Modeling for Large Angles of Attack

Modeling of unsteady aerodynamic loads at high angles of attack using a small amount of experimental or simulation data to construct predictive models for unknown states can greatly improve the efficiency of aircraft unsteady aerodynamic design and flight dynamics analysis.In this paper,aiming at the problems of poor generalization of traditional aerodynamic models and intelligent models,an intelligent aerodynamic modeling method based on gated neural units is proposed.The time memory characteristics of the gated neural unit is fully utilized,thus the nonlinear flow field characterization ability of the learning and training process is enhanced,and the generalization ability of the whole prediction model is improved.The prediction and verification of the model are carried out under the maneuvering flight condition of NACA0015 airfoil.The results show that the model has good adaptability.In the interpolation prediction,the maximum prediction error of the lift and drag coefficients and the moment coefficient does not exceed 10%,which can basically represent the variation characteristics of the entire flow field.In the construction of extrapolation models,the training model based on the strong nonlinear data has good accuracy for weak nonlinear prediction.Furthermore,the error is larger,even exceeding 20%,which indicates that the extrapolation and generalization capabilities need to be further optimized by integrating physical models.Compared with the conventional state space equation model,the proposed method can improve the extrapolation accuracy and efficiency by 78%and 60%,respectively,which demonstrates the applied potential of this method in aerodynamic modeling.

The enlightenment of artificial intelligence large-scale model on the research of intelligent eye diagnosis in traditional Chinese medicine

Eye diagnosis is a method for inspecting systemic diseases and syndromes by observing the eyes.With the development of intelligent diagnosis in traditional Chinese medicine(TCM);artificial intelligence(AI)can improve the accuracy and efficiency of eye diagnosis.However;the research on intelligent eye diagnosis still faces many challenges;including the lack of standardized and precisely labeled data;multi-modal information analysis;and artificial in-telligence models for syndrome differentiation.The widespread application of AI models in medicine provides new insights and opportunities for the research of eye diagnosis intelli-gence.This study elaborates on the three key technologies of AI models in the intelligent ap-plication of TCM eye diagnosis;and explores the implications for the research of eye diagno-sis intelligence.First;a database concerning eye diagnosis was established based on self-su-pervised learning so as to solve the issues related to the lack of standardized and precisely la-beled data.Next;the cross-modal understanding and generation of deep neural network models to address the problem of lacking multi-modal information analysis.Last;the build-ing of data-driven models for eye diagnosis to tackle the issue of the absence of syndrome dif-ferentiation models.In summary;research on intelligent eye diagnosis has great potential to be applied the surge of AI model applications.

林语堂“歌戈鱼虞模”古读研究的贡献及局限

林语堂根据印欧语系的音读演变以及中国古音的阴阳对转情况,提出“歌戈读长ɑ,麻读短ɑ”的观点。他还根据历史语言学原则、古韵部之间的音转关系、梵汉对音的证据,提出“周秦汉魏时,鱼虞模读开o(或合o)”的见解。虽然在论述的过程中还存在不足,但他于研究材料、方法、结论皆有个人创见,为后来学者更深入的探讨提供了理论基础。

Non-cooperative Space Target Estimation Algorithm Without Prior Information Dependence Based on Temporal Line of Sight Constraint

Under single-satellite observation,the parameter estimation of the boost phase of high-precision space noncooperative targets requires prior information.To improve the accuracy without prior information,we propose a parameter estimation model of the boost phase based on trajectory plane parametric cutting.The use of the plane passing through the geo-center and the cutting sequence line of sight(LOS)generates the trajectory-cutting plane.With the coefficient of the trajectory cutting plane directly used as the parameter to be estimated,a motion parameter estimation model in space non-cooperative targets is established,and the Gauss-Newton iteration method is used to solve the flight parameters.The experimental results show that the estimation algorithm proposed in this paper weakly relies on prior information and has higher estimation accuracy,providing a practical new idea and method for the parameter estimation of space non-cooperative targets under single-satellite warning.

Real-Time Monitoring Method for Cow Rumination Behavior Based on Edge Computing and Improved MobileNet v3

[Objective]Real-time monitoring of cow ruminant behavior is of paramount importance for promptly obtaining relevant information about cow health and predicting cow diseases.Currently,various strategies have been proposed for monitoring cow ruminant behavior,including video surveillance,sound recognition,and sensor monitoring methods.How‐ever,the application of edge device gives rise to the issue of inadequate real-time performance.To reduce the volume of data transmission and cloud computing workload while achieving real-time monitoring of dairy cow rumination behavior,a real-time monitoring method was proposed for cow ruminant behavior based on edge computing.[Methods]Autono‐mously designed edge devices were utilized to collect and process six-axis acceleration signals from cows in real-time.Based on these six-axis data,two distinct strategies,federated edge intelligence and split edge intelligence,were investigat‐ed for the real-time recognition of cow ruminant behavior.Focused on the real-time recognition method for cow ruminant behavior leveraging federated edge intelligence,the CA-MobileNet v3 network was proposed by enhancing the MobileNet v3 network with a collaborative attention mechanism.Additionally,a federated edge intelligence model was designed uti‐lizing the CA-MobileNet v3 network and the FedAvg federated aggregation algorithm.In the study on split edge intelli‐gence,a split edge intelligence model named MobileNet-LSTM was designed by integrating the MobileNet v3 network with a fusion collaborative attention mechanism and the Bi-LSTM network.[Results and Discussions]Through compara‐tive experiments with MobileNet v3 and MobileNet-LSTM,the federated edge intelligence model based on CA-Mo‐bileNet v3 achieved an average Precision rate,Recall rate,F1-Score,Specificity,and Accuracy of 97.1%,97.9%,97.5%,98.3%,and 98.2%,respectively,yielding the best recognition performance.[Conclusions]It is provided a real-time and effective method for monitoring cow ruminant behavior,and the proposed federated edge intelligence model can be ap‐plied in practical settings.

Cross-section distortion and springback characteristics of double-cavity aluminum profile in force controlled stretch-bending

3D elastic-plastic FE model for simulating the force controlled stretch-bending process of double-cavity aluminum profile was established using hybrid explicit−implicit solvent method.Considering the computational accuracy and efficiency,the optimal choices of numerical parameters and algorithms in FE modelling were determined.The formation mechanisms of cross-section distortion and springback were revealed.The effects of pre-stretching,post-stretching,friction,and the addition of internal fillers on forming quality were investigated.The results show that the stress state of profile in stretch-bending is uniaxial with only a circumferential stress.The stress distribution along the length direction of profile is non-uniform and the maximum tensile stress is located at a certain distance away from the center of profile.As aluminum profile is gradually attached to bending die,the distribution characteristic of cross-section distortion along the length direction of profile changes from V-shape to W-shape.After unloading the forming tools,cross-section distortion decreases obviously due to the stress relaxation,with a maximum distortion difference of 13%before and after unloading.As pre-stretching and post-stretching forces increase,cross-section distortion increases gradually,while springback first decreases and then remains unchanged.With increasing friction between bending die and profile,cross-section distortion slightly decreases,while springback increases.Cross-section distortion decreases by 83%with adding PVC fillers into the cavities of profile,while springback increases by 192.2%.

Temperature-Induced Unfolding Pathway of Staphylococcal Enterotoxin B:Insights from Circular Dichroism and Molecular Dynamics Simulation

In this study,circular dichroism(CD)and molecular dynamics(MD)simulation were used to investigate the thermal unfolding pathway of staphylococcal enterotoxin B(SEB)at temperatures of 298–371 and 298–500 K,and the relationship between the experimental and simulation results were explored.Our computational findings on the secondary structure of SEB showed that at room temperature,the CD spectroscopic results were highly consistent with the MD results.Moreover,under heating conditions,the changing trends of helix,sheet and random coil obtained by CD spectral fitting were highly consistent with those obtained by MD.In order to gain a deeper understanding of the thermal stability mechanism of SEB,the MD trajectories were analyzed in terms of root mean square deviation(RMSD),secondary structure assignment(SSA),radius of gyration(R_(g)),free energy surfaces(FES),solvent-accessible surface area(SASA),hydrogen bonds and salt bridges.The results showed that at low heating temperature,domain Ⅰ without loops(omitting the mobile loop region)mainly relied on hydrophobic interaction to maintain its thermal stability,whereas the thermal stability of domain Ⅱ was mainly controlled by salt bridges and hydrogen bonds.Under high heating temperature conditions,the hydrophobic interactions in domain Ⅰ without loops were destroyed and the secondary structure was almost completely lost,while domain Ⅱ could still rely on salt bridges as molecular staples to barely maintain the stability of the secondary structure.These results help us to understand the thermodynamic and kinetic mechanisms that maintain the thermal stability of SEB at the molecular level,and provide a direction for establishing safer and more effective food sterilization processes.

Research on Multi-modal In-Vehicle Intelligent Personal Assistant Design

Intelligent personal assistants play a pivotal role in in-vehicle systems,significantly enhancing life efficiency,driving safety,and decision-making support.In this study,the multi-modal design elements of intelligent personal assistants within the context of visual,auditory,and somatosensory interactions with drivers were discussed.Their impact on the driver’s psychological state through various modes such as visual imagery,voice interaction,and gesture interaction were explored.The study also introduced innovative designs for in-vehicle intelligent personal assistants,incorporating design principles such as driver-centricity,prioritizing passenger safety,and utilizing timely feedback as a criterion.Additionally,the study employed design methods like driver behavior research and driving situation analysis to enhance the emotional connection between drivers and their vehicles,ultimately improving driver satisfaction and trust.

In-situ Micro-CT analysis of deformation behavior in sandwich-structured meta-stable beta Ti−35Nb alloy

Beta Ti−35Nb sandwich-structured composites with various reinforcing layers were designed and produced using additive manufacturing(AM)to achieve a balance between light weight and high strength.The impact of reinforcing layers on the compressive deformation behavior of porous composites was investigated through micro-computed tomography(Micro-CT)and finite element method(FEM)analyses.The results indicate that the addition of reinforcement layers to sandwich structures can significantly enhance the compressive yield strength and energy absorption capacity of porous metal structures;Micro-CT in-situ observation shows that the strain of the porous structure without the reinforcing layer is concentrated in the middle region,while the strain of the porous structure with the reinforcing layer is uniformly distributed;FEM analysis reveals that the reinforcing layers can alter stress distribution and reduce stress concentration,thereby promoting uniform deformation of the porous structure.The addition of reinforcing layer increases the compressive yield strength of sandwich-structured composite materials by 124%under the condition of limited reduction of porosity,and the yield strength increases from 4.6 to 10.3 MPa.

关于∧-模的一个准素子模定理

在有单位元 1的交换环∧上作成的模 M上 ,定义σ-子模 ,其中σ∈ H om∧ (M,M) ,从而得到准素子模定理。

我国大豆生产波动动因分析——基于省际面板模型的实证研究

在我国大豆生产波动频繁的情况下,找出影响大豆单产、种植面积的影响因素是实现“提单产、稳面积”的首要举措。研究采用1999~2020年21个省份面板数据,在分析大豆单产、种植面积的变化趋势基础上,根据大豆单产增减趋势,把大豆产区划分成增产区、减产区和波动区,采用改进的面板数据模型分别对三个产区的大豆单产和种植面积的影响因素进行实证分析。结果表明,机械化收割、产研结合、农药投入、小型拖拉机的投入是影响不同区域大豆单产的显著因素;机械化耕作是提升所有区域种植面积的显著因素,玉米种植面积、化肥投入、农药投入、灌溉条件、科研资金与人员投入、产研结合、机械化收割是影响不同区域大豆种植面积显著因素。根据分析,提出了相应政策建议。

模型水轮机初生空化的特征谱提取识别方法

目前水轮机模型初生空化识别方法仍采用人工识别法,这种方法获得结果的周期较长、主观性强、准确度及效率都较低。针对于此,对水轮机模型初生空化识别方法进行了研究与优化创新,提出了一种基于炮声谱与特殊脉动谱特征提取的水轮机空化智能识别方法,即多态泡音智能识别方法 PSVFR。该方法依据自主开发的水轮机空化噪声多态算法MTCSPC,对数据进行处理,通过采集初生空化音态特征向量,建立矩阵模型,与样本数据库中的定性矩阵进行特征比对、计算、判断,以帮助机器完成对模型水轮机空化噪声的学习和识别。与现有技术相比,该方法能够提高机器对水轮机初生空化现象的识别准确度和识别效率,识别效率可达80%。

基于YOLOv5改进的CBAM平贝母检测模型

目前的平贝母种植阶段中,覆土仍然依靠人工,这存在耗时、耗力、成本高等问题。故需要研制一款智能化的平贝母覆土机器,而实现智能化的第一步在于实现识别。以YOLOv5系列为基础训练出一系列模型,在其系列中挑出最优模型(YOLOv5s)进行改进;进而在骨干段加入CBAM注意力机制替换其C3结合,从而减少网络序列长度,减少计算复杂度。试验结果表明,改进后模型的准确度最高可达91.32%,召回率最高可达88.78%。本文模型的提出为下一步实现平贝母覆土机智能控制提供理论参考。

Large-scale complex physical modeling and precisionanalysis

Large-scale 3D physical models of complex structures can be used to simulate hydrocarbon exploration areas. The high-fidelity simulation of actual structures poses challenges to model building and quality control. Such models can be used to collect wideazimuth, multi-azimuth, and full-azimuth seismic data that can be used to verify various 3D processing and interpretation methods. Faced with nonideal imaging problems owing to the extensive complex surface conditions and subsurface structures in the oil-rich foreland basins of western China, we designed and built the KS physical model based on the complex subsurface structure. This is the largest and most complex 3D physical model built to date. The physical modeling technology advancements mainly involve 1） the model design method, 2） the model casting flow, and 3） data acquisition. A 3D velocity model of the physical model was obtained for the first time, and the model building precision was quantitatively analyzed. The absolute error was less than 3 mm, which satisfies the experimental requirements. The 3D velocity model obtained from 3D measurements of the model layers is the basis for testing various imaging methods. Furthermore, the model is considered a standard in seismic physical modeling technology.