As coal mining progresses to greater depths,controlling the stability of surrounding rock in deep roadways has become an increasingly complex challenge.Although four-dimensional(4D)support theoretically offers unique ...As coal mining progresses to greater depths,controlling the stability of surrounding rock in deep roadways has become an increasingly complex challenge.Although four-dimensional(4D)support theoretically offers unique advantages in maintaining the stability of rock mass,the disaster evolution processes and multi-source information response characteristics in deep roadways with 4D support remain unclear.Consequently,a large-scale physical model testing system and self-designed 4D support components were employed to conduct similarity model tests on the surrounding rock failure process under unsupported(U-1),traditional bolt-mesh-cable support(T-2),and 4D support(4D-R-3)conditions.Combined with multi-source monitoring techniques,including stress–strain,digital image correlation(DIC),acoustic emission(AE),microseismic(MS),parallel electric(PE),and electromagnetic radiation(EMR),the mechanical behavior and multi-source information responses were comprehensively analyzed.The results show that the peak stress and displacement of the models are positively correlated with the support strength.The multi-source information exhibits distinct response characteristics under different supports.The response frequency,energy,and fluctuationsof AE,MS,and EMR signals,along with the apparent resistivity(AR)high-resistivity zone,follow the trend U-1>T-2>4D-R-3.Furthermore,multi-source information exhibits significantdifferences in sensitivity across different phases.The AE,MS,and EMR signals exhibit active responses to rock mass activity at each phase.However,AR signals are only sensitive to the fracture propagation during the plastic yield and failure phases.In summary,the 4D support significantlyenhances the bearing capacity and plastic deformation of the models,while substantially reducing the frequency,energy,and fluctuationsof multi-source signals.展开更多
The spatial offset of bridge has a significant impact on the safety,comfort,and durability of high-speed railway(HSR)operations,so it is crucial to rapidly and effectively detect the spatial offset of operational HSR ...The spatial offset of bridge has a significant impact on the safety,comfort,and durability of high-speed railway(HSR)operations,so it is crucial to rapidly and effectively detect the spatial offset of operational HSR bridges.Drive-by monitoring of bridge uneven settlement demonstrates significant potential due to its practicality,cost-effectiveness,and efficiency.However,existing drive-by methods for detecting bridge offset have limitations such as reliance on a single data source,low detection accuracy,and the inability to identify lateral deformations of bridges.This paper proposes a novel drive-by inspection method for spatial offset of HSR bridge based on multi-source data fusion of comprehensive inspection train.Firstly,dung beetle optimizer-variational mode decomposition was employed to achieve adaptive decomposition of non-stationary dynamic signals,and explore the hidden temporal relationships in the data.Subsequently,a long short-term memory neural network was developed to achieve feature fusion of multi-source signal and accurate prediction of spatial settlement of HSR bridge.A dataset of track irregularities and CRH380A high-speed train responses was generated using a 3D train-track-bridge interaction model,and the accuracy and effectiveness of the proposed hybrid deep learning model were numerically validated.Finally,the reliability of the proposed drive-by inspection method was further validated by analyzing the actual measurement data obtained from comprehensive inspection train.The research findings indicate that the proposed approach enables rapid and accurate detection of spatial offset in HSR bridge,ensuring the long-term operational safety of HSR bridges.展开更多
This paper addresses the challenge of accurately and timely determining the position of a train,with specific consideration given to the integration of the global navigation satellite system(GNSS)and inertial navigati...This paper addresses the challenge of accurately and timely determining the position of a train,with specific consideration given to the integration of the global navigation satellite system(GNSS)and inertial navigation system(INS).To overcome the increasing errors in the INS during interruptions in GNSS signals,as well as the uncertainty associated with process and measurement noise,a deep learning-based method for train positioning is proposed.This method combines convolutional neural networks(CNN),long short-term memory(LSTM),and the invariant extended Kalman filter(IEKF)to enhance the perception of train positions.It effectively handles GNSS signal interruptions and mitigates the impact of noise.Experimental evaluation and comparisons with existing approaches are provided to illustrate the effectiveness and robustness of the proposed method.展开更多
Dempster-Shafer evidence theory is broadly employed in the research of multi-source information fusion.Nevertheless,when fusing highly conflicting evidence it may pro-duce counterintuitive outcomes.To address this iss...Dempster-Shafer evidence theory is broadly employed in the research of multi-source information fusion.Nevertheless,when fusing highly conflicting evidence it may pro-duce counterintuitive outcomes.To address this issue,a fusion approach based on a newly defined belief exponential diver-gence and Deng entropy is proposed.First,a belief exponential divergence is proposed as the conflict measurement between evidences.Then,the credibility of each evidence is calculated.Afterwards,the Deng entropy is used to calculate information volume to determine the uncertainty of evidence.Then,the weight of evidence is calculated by integrating the credibility and uncertainty of each evidence.Ultimately,initial evidences are amended and fused using Dempster’s rule of combination.The effectiveness of this approach in addressing the fusion of three typical conflict paradoxes is demonstrated by arithmetic exam-ples.Additionally,the proposed approach is applied to aerial tar-get recognition and iris dataset-based classification to validate its efficacy.Results indicate that the proposed approach can enhance the accuracy of target recognition and effectively address the issue of fusing conflicting evidences.展开更多
For milling tool life prediction and health management,accurate extraction and dimensionality reduction of its tool wear features are the key to reduce prediction errors.In this paper,we adopt multi-source information...For milling tool life prediction and health management,accurate extraction and dimensionality reduction of its tool wear features are the key to reduce prediction errors.In this paper,we adopt multi-source information fusion technology to extract and fuse the features of cutting vibration signal,cutting force signal and acoustic emission signal in time domain,frequency domain and time-frequency domain,and downscale the sample features by Pearson correlation coefficient to construct a sample data set;then we propose a tool life prediction model based on CNN-SVM optimized by genetic algorithm(GA),which uses CNN convolutional neural network as the feature learner and SVM support vector machine as the trainer for regression prediction.The results show that the improved model in this paper can effectively predict the tool life with better generalization ability,faster network fitting,and 99.85%prediction accuracy.And compared with the BP model,CNN model,SVM model and CNN-SVM model,the performance of the coefficient of determination R2 metric improved by 4.88%,2.96%,2.53%and 1.34%,respectively.展开更多
For a single-structure deep learning fault diagnosis model,its disadvantages are an insufficient feature extraction and weak fault classification capability.This paper proposes a multi-scale deep feature fusion intell...For a single-structure deep learning fault diagnosis model,its disadvantages are an insufficient feature extraction and weak fault classification capability.This paper proposes a multi-scale deep feature fusion intelligent fault diagnosis method based on information entropy.First,a normal autoencoder,denoising autoencoder,sparse autoencoder,and contractive autoencoder are used in parallel to construct a multi-scale deep neural network feature extraction structure.A deep feature fusion strategy based on information entropy is proposed to obtain low-dimensional features and ensure the robustness of the model and the quality of deep features.Finally,the advantage of the deep belief network probability model is used as the fault classifier to identify the faults.The effectiveness of the proposed method was verified by a gearbox test-bed.Experimental results show that,compared with traditional and existing intelligent fault diagnosis methods,the proposed method can obtain representative information and features from the raw data with higher classification accuracy.展开更多
For reservoirs with complex non-Gaussian geological characteristics,such as carbonate reservoirs or reservoirs with sedimentary facies distribution,it is difficult to implement history matching directly,especially for...For reservoirs with complex non-Gaussian geological characteristics,such as carbonate reservoirs or reservoirs with sedimentary facies distribution,it is difficult to implement history matching directly,especially for the ensemble-based data assimilation methods.In this paper,we propose a multi-source information fused generative adversarial network(MSIGAN)model,which is used for parameterization of the complex geologies.In MSIGAN,various information such as facies distribution,microseismic,and inter-well connectivity,can be integrated to learn the geological features.And two major generative models in deep learning,variational autoencoder(VAE)and generative adversarial network(GAN)are combined in our model.Then the proposed MSIGAN model is integrated into the ensemble smoother with multiple data assimilation(ESMDA)method to conduct history matching.We tested the proposed method on two reservoir models with fluvial facies.The experimental results show that the proposed MSIGAN model can effectively learn the complex geological features,which can promote the accuracy of history matching.展开更多
The learning status of learners directly affects the quality of learning.Compared with offline teachers,it is difficult for online teachers to capture the learning status of students in the whole class,and it is even ...The learning status of learners directly affects the quality of learning.Compared with offline teachers,it is difficult for online teachers to capture the learning status of students in the whole class,and it is even more difficult to continue to pay attention to studentswhile teaching.Therefore,this paper proposes an online learning state analysis model based on a convolutional neural network and multi-dimensional information fusion.Specifically,a facial expression recognition model and an eye state recognition model are constructed to detect students’emotions and fatigue,respectively.By integrating the detected data with the homework test score data after online learning,an analysis model of students’online learning status is constructed.According to the PAD model,the learning state is expressed as three dimensions of students’understanding,engagement and interest,and then analyzed from multiple perspectives.Finally,the proposed model is applied to actual teaching,and procedural analysis of 5 different types of online classroom learners is carried out,and the validity of the model is verified by comparing with the results of the manual analysis.展开更多
For some important object recognition applications such as intelligent robots and unmanned driving, images are collected on a consecutive basis and associated among themselves, besides, the scenes have steady prior fe...For some important object recognition applications such as intelligent robots and unmanned driving, images are collected on a consecutive basis and associated among themselves, besides, the scenes have steady prior features. Yet existing technologies do not take full advantage of this information. In order to take object recognition further than existing algorithms in the above application, an object recognition method that fuses temporal sequence with scene priori information is proposed. This method first employs YOLOv3 as the basic algorithm to recognize objects in single-frame images, then the DeepSort algorithm to establish association among potential objects recognized in images of different moments, and finally the confidence fusion method and temporal boundary processing method designed herein to fuse, at the decision level, temporal sequence information with scene priori information. Experiments using public datasets and self-built industrial scene datasets show that due to the expansion of information sources, the quality of single-frame images has less impact on the recognition results, whereby the object recognition is greatly improved. It is presented herein as a widely applicable framework for the fusion of information under multiple classes. All the object recognition algorithms that output object class, location information and recognition confidence at the same time can be integrated into this information fusion framework to improve performance.展开更多
Multi-source information fusion （MSIF） is imported into structural damage diagnosis methods to improve the validity of damage detection. After the introduction of the basic theory, the function model, classification...Multi-source information fusion （MSIF） is imported into structural damage diagnosis methods to improve the validity of damage detection. After the introduction of the basic theory, the function model, classifications and mathematical methods of MSIF, a structural damage detection method based on MSIF is presented, which is to fuse two or more damage character vectors from different structural damage diagnosis methods on the character-level. In an experiment of concrete plates, modal information is measured and analyzed. The structural damage detection method based on MSIF is taken to localize cracks of concrete plates and it is proved to be effective. Results of damage detection by the method based on MSIF are compared with those from the modal strain energy method and the flexibility method. Damage, which can hardly be detected by using the single damage identification method, can be diagnosed by the damage detection method based on the character-level MSIF technique. Meanwhile multi-location damage can be identified by the method based on MSIF. This method is sensitive to structural damage and different mathematical methods for MSIF have different preconditions and applicabilities for diversified structures. How to choose mathematical methods for MSIF should be discussed in detail in health monitoring systems of actual structures.展开更多
Near real-time maize phenology monitoring is crucial for field management,cropping system adjustments,and yield estimation.Most phenological monitoring methods are post-seasonal and heavily rely on high-frequency time...Near real-time maize phenology monitoring is crucial for field management,cropping system adjustments,and yield estimation.Most phenological monitoring methods are post-seasonal and heavily rely on high-frequency time-series data.These methods are not applicable on the unmanned aerial vehicle(UAV)platform due to the high cost of acquiring time-series UAV images and the shortage of UAV-based phenological monitoring methods.To address these challenges,we employed the Synthetic Minority Oversampling Technique(SMOTE)for sample augmentation,aiming to resolve the small sample modelling problem.Moreover,we utilized enhanced"separation"and"compactness"feature selection methods to identify input features from multiple data sources.In this process,we incorporated dynamic multi-source data fusion strategies,involving Vegetation index(VI),Color index(CI),and Texture features(TF).A two-stage neural network that combines Convolutional Neural Network(CNN)and Long Short-Term Memory Network(LSTM)is proposed to identify maize phenological stages(including sowing,seedling,jointing,trumpet,tasseling,maturity,and harvesting)on UAV platforms.The results indicate that the dataset generated by SMOTE closely resembles the measured dataset.Among dynamic data fusion strategies,the VI-TF combination proves to be most effective,with CI-TF and VI-CI combinations following behind.Notably,as more data sources are integrated,the model's demand for input features experiences a significant decline.In particular,the CNN-LSTM model,based on the fusion of three data sources,exhibited remarkable reliability when validating the three datasets.For Dataset 1(Beijing Xiaotangshan,2023:Data from 12 UAV Flight Missions),the model achieved an overall accuracy(OA)of 86.53%.Additionally,its precision(Pre),recall(Rec),F1 score(F1),false acceptance rate(FAR),and false rejection rate(FRR)were 0.89,0.89,0.87,0.11,and 0.11,respectively.The model also showed strong generalizability in Dataset 2(Beijing Xiaotangshan,2023:Data from 6 UAV Flight Missions)and Dataset 3(Beijing Xiaotangshan,2022:Data from 4 UAV Flight Missions),with OAs of 89.4%and 85%,respectively.Meanwhile,the model has a low demand for input featu res,requiring only 54.55%(99 of all featu res).The findings of this study not only offer novel insights into near real-time crop phenology monitoring,but also provide technical support for agricultural field management and cropping system adaptation.展开更多
Analyzing the service behavior of high dams and establishing early-warning systems for them have become increasingly important in ensuring their long-term service.Current analysis methods used to obtain safety monitor...Analyzing the service behavior of high dams and establishing early-warning systems for them have become increasingly important in ensuring their long-term service.Current analysis methods used to obtain safety monitoring data are suited only to single survey point data.Unreliable or even paradoxical results are inevitably obtained when processing large amounts of monitoring data,thereby causing difficulty in acquiring precise conclusions.Therefore,we have developed a new method based on multi-source information fusion for conducting a comprehensive analysis of prototype monitoring data of high dams.In addition,we propose the use of decision information entropy analysis for building a diagnosis and early-warning system for the long-term service of high dams.Data metrics reduction is achieved using information fusion at the data level.A Bayesian information fusion is then conducted at the decision level to obtain a comprehensive diagnosis.Early-warning outcomes can be released after sorting analysis results from multi-positions in the dam according to importance.A case study indicates that the new method can effectively handle large amounts of monitoring data from numerous survey points.It can likewise obtain precise real-time results and export comprehensive early-warning outcomes from multi-positions of high dams.展开更多
Visual question answering(VQA)is a multimodal task,involving a deep understanding of the image scene and the question’s meaning and capturing the relevant correlations between both modalities to infer the appropriate...Visual question answering(VQA)is a multimodal task,involving a deep understanding of the image scene and the question’s meaning and capturing the relevant correlations between both modalities to infer the appropriate answer.In this paper,we propose a VQA system intended to answer yes/no questions about real-world images,in Arabic.To support a robust VQA system,we work in two directions:(1)Using deep neural networks to semantically represent the given image and question in a fine-grainedmanner,namely ResNet-152 and Gated Recurrent Units(GRU).(2)Studying the role of the utilizedmultimodal bilinear pooling fusion technique in the trade-o.between the model complexity and the overall model performance.Some fusion techniques could significantly increase the model complexity,which seriously limits their applicability for VQA models.So far,there is no evidence of how efficient these multimodal bilinear pooling fusion techniques are for VQA systems dedicated to yes/no questions.Hence,a comparative analysis is conducted between eight bilinear pooling fusion techniques,in terms of their ability to reduce themodel complexity and improve themodel performance in this case of VQA systems.Experiments indicate that these multimodal bilinear pooling fusion techniques have improved the VQA model’s performance,until reaching the best performance of 89.25%.Further,experiments have proven that the number of answers in the developed VQA system is a critical factor that a.ects the effectiveness of these multimodal bilinear pooling techniques in achieving their main objective of reducing the model complexity.The Multimodal Local Perception Bilinear Pooling(MLPB)technique has shown the best balance between the model complexity and its performance,for VQA systems designed to answer yes/no questions.展开更多
Hydraulic presses are indispensable in automotive and aerospace manufacturing,with hydraulic cylinders serving as key components for operational safety and product quality.Internal leakage faults in hydraulic cylinder...Hydraulic presses are indispensable in automotive and aerospace manufacturing,with hydraulic cylinders serving as key components for operational safety and product quality.Internal leakage faults in hydraulic cylinders are difficult to diagnose due to the scarcity of labeled data,the complexity of fault mechanisms,and the limited representation capability of single-signal methods under variable operating conditions.To address these issues,a hybrid deep learning feature fusion model based on displacement error and pressure signal,including convolutional autoencoder,multi-head attention mechanism,residual network and bidirectional long short time series neural network(CAEMRAB),is proposed for the diagnosis and classification of leakage faults in hydraulic cylinders.A hydraulic cylinder test system simulates heavy load,variable speed,and nonlinear motion under actual operating conditions.Through the all-round deep feature decoupling of the proposed model,the multi-source signal representation ability in complex and multi-noise environments is enhanced,effectively extracting the local and global features of displacement error and pressure signal fault data and achieving efficient classification.Experimental results indicate that the proposed model achieves at least a 3.95%improvement in diagnostic accuracy compared with ablation models.In addition,it exhibits high diagnostic stability across other models,single-signal diagnosis,varying sample sizes,and complex noise conditions.These experiments fully validate the superior performance of the proposed method in terms of diagnostic accuracy,reliability,and robustness.展开更多
D-S evidence theory,as a general framework for reasoning with uncertainty,allows combining pieces of evidence from different information sources to derive a degree of belief function that is a type of fuzzy measure.Ho...D-S evidence theory,as a general framework for reasoning with uncertainty,allows combining pieces of evidence from different information sources to derive a degree of belief function that is a type of fuzzy measure.However,the mass assignments given by unknown information sources are disordered.How to measure the difference between the mass assignments has aroused people’s interest.In this paper,inspired by the information volume,a novel distance-based measure is proposed to measure the difference between mass assignments.The method can refine the uncertain information given by experts and compare the refined information to obtain the difference between mass assignments.At the same time,it is verified that the measure not only meets the properties of distance,but also proves the superiority of the proposed Information Volume Distance(IVD)through simulation experiments.Meanwhile,in the process of information fusion,the reliability of each source could be quantified through IVD.Therefore,based on IVD,a new multi-source information algorithm is proposed to solve the problem of multi-source information fusion.Moreover,algorithm is applied to decision-making problem and compare with other methods to verify the effectiveness.展开更多
Personality distinguishes individuals’ patterns of feeling, thinking,and behaving. Predicting personality from small video series is an excitingresearch area in computer vision. The majority of the existing research ...Personality distinguishes individuals’ patterns of feeling, thinking,and behaving. Predicting personality from small video series is an excitingresearch area in computer vision. The majority of the existing research concludespreliminary results to get immense knowledge from visual and Audio(sound) modality. To overcome the deficiency, we proposed the Deep BimodalFusion (DBF) approach to predict five traits of personality-agreeableness,extraversion, openness, conscientiousness and neuroticism. In the proposedframework, regarding visual modality, the modified convolution neural networks(CNN), more specifically Descriptor Aggregator Model (DAN) areused to attain significant visual modality. The proposed model extracts audiorepresentations for greater efficiency to construct the long short-termmemory(LSTM) for the audio modality. Moreover, employing modality-based neuralnetworks allows this framework to independently determine the traits beforecombining them with weighted fusion to achieve a conclusive prediction of thegiven traits. The proposed approach attains the optimal mean accuracy score,which is 0.9183. It is achieved based on the average of five personality traitsand is thus better than previously proposed frameworks.展开更多
With the emergence and development of social networks,people can stay in touch with friends,family,and colleagues more quickly and conveniently,regardless of their location.This ubiquitous digital internet environment...With the emergence and development of social networks,people can stay in touch with friends,family,and colleagues more quickly and conveniently,regardless of their location.This ubiquitous digital internet environment has also led to large-scale disclosure of personal privacy.Due to the complexity and subtlety of sensitive information,traditional sensitive information identification technologies cannot thoroughly address the characteristics of each piece of data,thus weakening the deep connections between text and images.In this context,this paper adopts the CLIP model as a modality discriminator.By using comparative learning between sensitive image descriptions and images,the similarity between the images and the sensitive descriptions is obtained to determine whether the images contain sensitive information.This provides the basis for identifying sensitive information using different modalities.Specifically,if the original data does not contain sensitive information,only single-modality text-sensitive information identification is performed;if the original data contains sensitive information,multimodality sensitive information identification is conducted.This approach allows for differentiated processing of each piece of data,thereby achieving more accurate sensitive information identification.The aforementioned modality discriminator can address the limitations of existing sensitive information identification technologies,making the identification of sensitive information from the original data more appropriate and precise.展开更多
During drilling operations,the low resolution of seismic data often limits the accurate characterization of small-scale geological bodies near the borehole and ahead of the drill bit.This study investigates high-resol...During drilling operations,the low resolution of seismic data often limits the accurate characterization of small-scale geological bodies near the borehole and ahead of the drill bit.This study investigates high-resolution seismic data processing technologies and methods tailored for drilling scenarios.The high-resolution processing of seismic data is divided into three stages:pre-drilling processing,post-drilling correction,and while-drilling updating.By integrating seismic data from different stages,spatial ranges,and frequencies,together with information from drilled wells and while-drilling data,and applying artificial intelligence modeling techniques,a progressive high-resolution processing technology of seismic data based on multi-source information fusion is developed,which performs simple and efficient seismic information updates during drilling.Case studies show that,with the gradual integration of multi-source information,the resolution and accuracy of seismic data are significantly improved,and thin-bed weak reflections are more clearly imaged.The updated seismic information while-drilling demonstrates high value in predicting geological bodies ahead of the drill bit.Validation using logging,mud logging,and drilling engineering data ensures the fidelity of the processing results of high-resolution seismic data.This provides clearer and more accurate stratigraphic information for drilling operations,enhancing both drilling safety and efficiency.展开更多
基金supported by the National Natural Science Foundation of China(Grant Nos.U22A20598 and 52104107)the"Qinglan Project"of Jiangsu Colleges and Universities,Young Elite Scientists Sponsorship Program of Jiangsu Province(Grant No.TJ-2023-086).
文摘As coal mining progresses to greater depths,controlling the stability of surrounding rock in deep roadways has become an increasingly complex challenge.Although four-dimensional(4D)support theoretically offers unique advantages in maintaining the stability of rock mass,the disaster evolution processes and multi-source information response characteristics in deep roadways with 4D support remain unclear.Consequently,a large-scale physical model testing system and self-designed 4D support components were employed to conduct similarity model tests on the surrounding rock failure process under unsupported(U-1),traditional bolt-mesh-cable support(T-2),and 4D support(4D-R-3)conditions.Combined with multi-source monitoring techniques,including stress–strain,digital image correlation(DIC),acoustic emission(AE),microseismic(MS),parallel electric(PE),and electromagnetic radiation(EMR),the mechanical behavior and multi-source information responses were comprehensively analyzed.The results show that the peak stress and displacement of the models are positively correlated with the support strength.The multi-source information exhibits distinct response characteristics under different supports.The response frequency,energy,and fluctuationsof AE,MS,and EMR signals,along with the apparent resistivity(AR)high-resistivity zone,follow the trend U-1>T-2>4D-R-3.Furthermore,multi-source information exhibits significantdifferences in sensitivity across different phases.The AE,MS,and EMR signals exhibit active responses to rock mass activity at each phase.However,AR signals are only sensitive to the fracture propagation during the plastic yield and failure phases.In summary,the 4D support significantlyenhances the bearing capacity and plastic deformation of the models,while substantially reducing the frequency,energy,and fluctuationsof multi-source signals.
基金sponsored by the National Natural Science Foundation of China(Grant No.52178100).
文摘The spatial offset of bridge has a significant impact on the safety,comfort,and durability of high-speed railway(HSR)operations,so it is crucial to rapidly and effectively detect the spatial offset of operational HSR bridges.Drive-by monitoring of bridge uneven settlement demonstrates significant potential due to its practicality,cost-effectiveness,and efficiency.However,existing drive-by methods for detecting bridge offset have limitations such as reliance on a single data source,low detection accuracy,and the inability to identify lateral deformations of bridges.This paper proposes a novel drive-by inspection method for spatial offset of HSR bridge based on multi-source data fusion of comprehensive inspection train.Firstly,dung beetle optimizer-variational mode decomposition was employed to achieve adaptive decomposition of non-stationary dynamic signals,and explore the hidden temporal relationships in the data.Subsequently,a long short-term memory neural network was developed to achieve feature fusion of multi-source signal and accurate prediction of spatial settlement of HSR bridge.A dataset of track irregularities and CRH380A high-speed train responses was generated using a 3D train-track-bridge interaction model,and the accuracy and effectiveness of the proposed hybrid deep learning model were numerically validated.Finally,the reliability of the proposed drive-by inspection method was further validated by analyzing the actual measurement data obtained from comprehensive inspection train.The research findings indicate that the proposed approach enables rapid and accurate detection of spatial offset in HSR bridge,ensuring the long-term operational safety of HSR bridges.
基金supported by the National Natural Science Foundation of China(Nos.61925302,62273027)the Beijing Natural Science Foundation(L211021).
文摘This paper addresses the challenge of accurately and timely determining the position of a train,with specific consideration given to the integration of the global navigation satellite system(GNSS)and inertial navigation system(INS).To overcome the increasing errors in the INS during interruptions in GNSS signals,as well as the uncertainty associated with process and measurement noise,a deep learning-based method for train positioning is proposed.This method combines convolutional neural networks(CNN),long short-term memory(LSTM),and the invariant extended Kalman filter(IEKF)to enhance the perception of train positions.It effectively handles GNSS signal interruptions and mitigates the impact of noise.Experimental evaluation and comparisons with existing approaches are provided to illustrate the effectiveness and robustness of the proposed method.
基金supported by the National Natural Science Foundation of China(61903305,62073267)the Fundamental Research Funds for the Central Universities(HXGJXM202214).
文摘Dempster-Shafer evidence theory is broadly employed in the research of multi-source information fusion.Nevertheless,when fusing highly conflicting evidence it may pro-duce counterintuitive outcomes.To address this issue,a fusion approach based on a newly defined belief exponential diver-gence and Deng entropy is proposed.First,a belief exponential divergence is proposed as the conflict measurement between evidences.Then,the credibility of each evidence is calculated.Afterwards,the Deng entropy is used to calculate information volume to determine the uncertainty of evidence.Then,the weight of evidence is calculated by integrating the credibility and uncertainty of each evidence.Ultimately,initial evidences are amended and fused using Dempster’s rule of combination.The effectiveness of this approach in addressing the fusion of three typical conflict paradoxes is demonstrated by arithmetic exam-ples.Additionally,the proposed approach is applied to aerial tar-get recognition and iris dataset-based classification to validate its efficacy.Results indicate that the proposed approach can enhance the accuracy of target recognition and effectively address the issue of fusing conflicting evidences.
基金financed with the means of Basic Scientific Research Youth Program of Education Department of Liaoning Province,No.LJKQZ2021185Yingkou Enterprise and Doctor Innovation Program (QB-2021-05).
文摘For milling tool life prediction and health management,accurate extraction and dimensionality reduction of its tool wear features are the key to reduce prediction errors.In this paper,we adopt multi-source information fusion technology to extract and fuse the features of cutting vibration signal,cutting force signal and acoustic emission signal in time domain,frequency domain and time-frequency domain,and downscale the sample features by Pearson correlation coefficient to construct a sample data set;then we propose a tool life prediction model based on CNN-SVM optimized by genetic algorithm(GA),which uses CNN convolutional neural network as the feature learner and SVM support vector machine as the trainer for regression prediction.The results show that the improved model in this paper can effectively predict the tool life with better generalization ability,faster network fitting,and 99.85%prediction accuracy.And compared with the BP model,CNN model,SVM model and CNN-SVM model,the performance of the coefficient of determination R2 metric improved by 4.88%,2.96%,2.53%and 1.34%,respectively.
基金Supported by National Natural Science Foundation of China and Civil Aviation Administration of China Joint Funded Project(Grant No.U1733108)Key Project of Tianjin Science and Technology Support Program(Grant No.16YFZCSY00860).
文摘For a single-structure deep learning fault diagnosis model,its disadvantages are an insufficient feature extraction and weak fault classification capability.This paper proposes a multi-scale deep feature fusion intelligent fault diagnosis method based on information entropy.First,a normal autoencoder,denoising autoencoder,sparse autoencoder,and contractive autoencoder are used in parallel to construct a multi-scale deep neural network feature extraction structure.A deep feature fusion strategy based on information entropy is proposed to obtain low-dimensional features and ensure the robustness of the model and the quality of deep features.Finally,the advantage of the deep belief network probability model is used as the fault classifier to identify the faults.The effectiveness of the proposed method was verified by a gearbox test-bed.Experimental results show that,compared with traditional and existing intelligent fault diagnosis methods,the proposed method can obtain representative information and features from the raw data with higher classification accuracy.
基金supported by the National Natural Science Foundation of China under Grant 51722406,52074340,and 51874335the Shandong Provincial Natural Science Foundation under Grant JQ201808+5 种基金The Fundamental Research Funds for the Central Universities under Grant 18CX02097Athe Major Scientific and Technological Projects of CNPC under Grant ZD2019-183-008the Science and Technology Support Plan for Youth Innovation of University in Shandong Province under Grant 2019KJH002the National Research Council of Science and Technology Major Project of China under Grant 2016ZX05025001-006111 Project under Grant B08028Sinopec Science and Technology Project under Grant P20050-1
文摘For reservoirs with complex non-Gaussian geological characteristics,such as carbonate reservoirs or reservoirs with sedimentary facies distribution,it is difficult to implement history matching directly,especially for the ensemble-based data assimilation methods.In this paper,we propose a multi-source information fused generative adversarial network(MSIGAN)model,which is used for parameterization of the complex geologies.In MSIGAN,various information such as facies distribution,microseismic,and inter-well connectivity,can be integrated to learn the geological features.And two major generative models in deep learning,variational autoencoder(VAE)and generative adversarial network(GAN)are combined in our model.Then the proposed MSIGAN model is integrated into the ensemble smoother with multiple data assimilation(ESMDA)method to conduct history matching.We tested the proposed method on two reservoir models with fluvial facies.The experimental results show that the proposed MSIGAN model can effectively learn the complex geological features,which can promote the accuracy of history matching.
基金supported by the Chongqing Normal University Graduate Scientific Research Innovation Project (Grants YZH21014 and YZH21010).
文摘The learning status of learners directly affects the quality of learning.Compared with offline teachers,it is difficult for online teachers to capture the learning status of students in the whole class,and it is even more difficult to continue to pay attention to studentswhile teaching.Therefore,this paper proposes an online learning state analysis model based on a convolutional neural network and multi-dimensional information fusion.Specifically,a facial expression recognition model and an eye state recognition model are constructed to detect students’emotions and fatigue,respectively.By integrating the detected data with the homework test score data after online learning,an analysis model of students’online learning status is constructed.According to the PAD model,the learning state is expressed as three dimensions of students’understanding,engagement and interest,and then analyzed from multiple perspectives.Finally,the proposed model is applied to actual teaching,and procedural analysis of 5 different types of online classroom learners is carried out,and the validity of the model is verified by comparing with the results of the manual analysis.
文摘For some important object recognition applications such as intelligent robots and unmanned driving, images are collected on a consecutive basis and associated among themselves, besides, the scenes have steady prior features. Yet existing technologies do not take full advantage of this information. In order to take object recognition further than existing algorithms in the above application, an object recognition method that fuses temporal sequence with scene priori information is proposed. This method first employs YOLOv3 as the basic algorithm to recognize objects in single-frame images, then the DeepSort algorithm to establish association among potential objects recognized in images of different moments, and finally the confidence fusion method and temporal boundary processing method designed herein to fuse, at the decision level, temporal sequence information with scene priori information. Experiments using public datasets and self-built industrial scene datasets show that due to the expansion of information sources, the quality of single-frame images has less impact on the recognition results, whereby the object recognition is greatly improved. It is presented herein as a widely applicable framework for the fusion of information under multiple classes. All the object recognition algorithms that output object class, location information and recognition confidence at the same time can be integrated into this information fusion framework to improve performance.
基金The National High Technology Research and Develop-ment Program of China(863Program)(No.2006AA04Z416)the Na-tional Science Fund for Distinguished Young Scholars(No.50725828)the Excellent Dissertation Program for Doctoral Degree of Southeast University(No.0705)
文摘Multi-source information fusion （MSIF） is imported into structural damage diagnosis methods to improve the validity of damage detection. After the introduction of the basic theory, the function model, classifications and mathematical methods of MSIF, a structural damage detection method based on MSIF is presented, which is to fuse two or more damage character vectors from different structural damage diagnosis methods on the character-level. In an experiment of concrete plates, modal information is measured and analyzed. The structural damage detection method based on MSIF is taken to localize cracks of concrete plates and it is proved to be effective. Results of damage detection by the method based on MSIF are compared with those from the modal strain energy method and the flexibility method. Damage, which can hardly be detected by using the single damage identification method, can be diagnosed by the damage detection method based on the character-level MSIF technique. Meanwhile multi-location damage can be identified by the method based on MSIF. This method is sensitive to structural damage and different mathematical methods for MSIF have different preconditions and applicabilities for diversified structures. How to choose mathematical methods for MSIF should be discussed in detail in health monitoring systems of actual structures.
基金supported by grants from the National Key Research and Development Program of China(2022YFD2001103)the National Natural Science Foundation of China(42371373)。
文摘Near real-time maize phenology monitoring is crucial for field management,cropping system adjustments,and yield estimation.Most phenological monitoring methods are post-seasonal and heavily rely on high-frequency time-series data.These methods are not applicable on the unmanned aerial vehicle(UAV)platform due to the high cost of acquiring time-series UAV images and the shortage of UAV-based phenological monitoring methods.To address these challenges,we employed the Synthetic Minority Oversampling Technique(SMOTE)for sample augmentation,aiming to resolve the small sample modelling problem.Moreover,we utilized enhanced"separation"and"compactness"feature selection methods to identify input features from multiple data sources.In this process,we incorporated dynamic multi-source data fusion strategies,involving Vegetation index(VI),Color index(CI),and Texture features(TF).A two-stage neural network that combines Convolutional Neural Network(CNN)and Long Short-Term Memory Network(LSTM)is proposed to identify maize phenological stages(including sowing,seedling,jointing,trumpet,tasseling,maturity,and harvesting)on UAV platforms.The results indicate that the dataset generated by SMOTE closely resembles the measured dataset.Among dynamic data fusion strategies,the VI-TF combination proves to be most effective,with CI-TF and VI-CI combinations following behind.Notably,as more data sources are integrated,the model's demand for input features experiences a significant decline.In particular,the CNN-LSTM model,based on the fusion of three data sources,exhibited remarkable reliability when validating the three datasets.For Dataset 1(Beijing Xiaotangshan,2023:Data from 12 UAV Flight Missions),the model achieved an overall accuracy(OA)of 86.53%.Additionally,its precision(Pre),recall(Rec),F1 score(F1),false acceptance rate(FAR),and false rejection rate(FRR)were 0.89,0.89,0.87,0.11,and 0.11,respectively.The model also showed strong generalizability in Dataset 2(Beijing Xiaotangshan,2023:Data from 6 UAV Flight Missions)and Dataset 3(Beijing Xiaotangshan,2022:Data from 4 UAV Flight Missions),with OAs of 89.4%and 85%,respectively.Meanwhile,the model has a low demand for input featu res,requiring only 54.55%(99 of all featu res).The findings of this study not only offer novel insights into near real-time crop phenology monitoring,but also provide technical support for agricultural field management and cropping system adaptation.
基金Project supported by the National Natural Science Foundation of China(Nos.51139001,51179066,51079046,and 50909041)
文摘Analyzing the service behavior of high dams and establishing early-warning systems for them have become increasingly important in ensuring their long-term service.Current analysis methods used to obtain safety monitoring data are suited only to single survey point data.Unreliable or even paradoxical results are inevitably obtained when processing large amounts of monitoring data,thereby causing difficulty in acquiring precise conclusions.Therefore,we have developed a new method based on multi-source information fusion for conducting a comprehensive analysis of prototype monitoring data of high dams.In addition,we propose the use of decision information entropy analysis for building a diagnosis and early-warning system for the long-term service of high dams.Data metrics reduction is achieved using information fusion at the data level.A Bayesian information fusion is then conducted at the decision level to obtain a comprehensive diagnosis.Early-warning outcomes can be released after sorting analysis results from multi-positions in the dam according to importance.A case study indicates that the new method can effectively handle large amounts of monitoring data from numerous survey points.It can likewise obtain precise real-time results and export comprehensive early-warning outcomes from multi-positions of high dams.
文摘Visual question answering(VQA)is a multimodal task,involving a deep understanding of the image scene and the question’s meaning and capturing the relevant correlations between both modalities to infer the appropriate answer.In this paper,we propose a VQA system intended to answer yes/no questions about real-world images,in Arabic.To support a robust VQA system,we work in two directions:(1)Using deep neural networks to semantically represent the given image and question in a fine-grainedmanner,namely ResNet-152 and Gated Recurrent Units(GRU).(2)Studying the role of the utilizedmultimodal bilinear pooling fusion technique in the trade-o.between the model complexity and the overall model performance.Some fusion techniques could significantly increase the model complexity,which seriously limits their applicability for VQA models.So far,there is no evidence of how efficient these multimodal bilinear pooling fusion techniques are for VQA systems dedicated to yes/no questions.Hence,a comparative analysis is conducted between eight bilinear pooling fusion techniques,in terms of their ability to reduce themodel complexity and improve themodel performance in this case of VQA systems.Experiments indicate that these multimodal bilinear pooling fusion techniques have improved the VQA model’s performance,until reaching the best performance of 89.25%.Further,experiments have proven that the number of answers in the developed VQA system is a critical factor that a.ects the effectiveness of these multimodal bilinear pooling techniques in achieving their main objective of reducing the model complexity.The Multimodal Local Perception Bilinear Pooling(MLPB)technique has shown the best balance between the model complexity and its performance,for VQA systems designed to answer yes/no questions.
基金the Scientific Research Foundation for High-level Talents of Anhui University of Science and Technology(2024yjrc73)R&D and industrialization of high-precision intelligent forging equipment for forming large-size light alloy components(202423i08050024)a large die forging press operation condition monitoring sensor and system application(2023YFB3210805)。
文摘Hydraulic presses are indispensable in automotive and aerospace manufacturing,with hydraulic cylinders serving as key components for operational safety and product quality.Internal leakage faults in hydraulic cylinders are difficult to diagnose due to the scarcity of labeled data,the complexity of fault mechanisms,and the limited representation capability of single-signal methods under variable operating conditions.To address these issues,a hybrid deep learning feature fusion model based on displacement error and pressure signal,including convolutional autoencoder,multi-head attention mechanism,residual network and bidirectional long short time series neural network(CAEMRAB),is proposed for the diagnosis and classification of leakage faults in hydraulic cylinders.A hydraulic cylinder test system simulates heavy load,variable speed,and nonlinear motion under actual operating conditions.Through the all-round deep feature decoupling of the proposed model,the multi-source signal representation ability in complex and multi-noise environments is enhanced,effectively extracting the local and global features of displacement error and pressure signal fault data and achieving efficient classification.Experimental results indicate that the proposed model achieves at least a 3.95%improvement in diagnostic accuracy compared with ablation models.In addition,it exhibits high diagnostic stability across other models,single-signal diagnosis,varying sample sizes,and complex noise conditions.These experiments fully validate the superior performance of the proposed method in terms of diagnostic accuracy,reliability,and robustness.
基金supported by the National Natural Science Foundation of China(No.62003280)Chongqing Talents:Exceptional Young Talents Project(No.cstc2022ycjhbgzxm0070)+1 种基金Natural Science Foundation of Chongqing,China(No.CSTB2022NSCQ-MSX0531)Chongqing Overseas Scholars Innovation Program(No.cx2022024).
文摘D-S evidence theory,as a general framework for reasoning with uncertainty,allows combining pieces of evidence from different information sources to derive a degree of belief function that is a type of fuzzy measure.However,the mass assignments given by unknown information sources are disordered.How to measure the difference between the mass assignments has aroused people’s interest.In this paper,inspired by the information volume,a novel distance-based measure is proposed to measure the difference between mass assignments.The method can refine the uncertain information given by experts and compare the refined information to obtain the difference between mass assignments.At the same time,it is verified that the measure not only meets the properties of distance,but also proves the superiority of the proposed Information Volume Distance(IVD)through simulation experiments.Meanwhile,in the process of information fusion,the reliability of each source could be quantified through IVD.Therefore,based on IVD,a new multi-source information algorithm is proposed to solve the problem of multi-source information fusion.Moreover,algorithm is applied to decision-making problem and compare with other methods to verify the effectiveness.
文摘Personality distinguishes individuals’ patterns of feeling, thinking,and behaving. Predicting personality from small video series is an excitingresearch area in computer vision. The majority of the existing research concludespreliminary results to get immense knowledge from visual and Audio(sound) modality. To overcome the deficiency, we proposed the Deep BimodalFusion (DBF) approach to predict five traits of personality-agreeableness,extraversion, openness, conscientiousness and neuroticism. In the proposedframework, regarding visual modality, the modified convolution neural networks(CNN), more specifically Descriptor Aggregator Model (DAN) areused to attain significant visual modality. The proposed model extracts audiorepresentations for greater efficiency to construct the long short-termmemory(LSTM) for the audio modality. Moreover, employing modality-based neuralnetworks allows this framework to independently determine the traits beforecombining them with weighted fusion to achieve a conclusive prediction of thegiven traits. The proposed approach attains the optimal mean accuracy score,which is 0.9183. It is achieved based on the average of five personality traitsand is thus better than previously proposed frameworks.
基金supported by the National Natural Science Foundation of China(No.62302540),with author Fangfang Shan for more information,please visit their website at https://www.nsfc.gov.cn/(accessed on 05 June 2024)Additionally,it is also funded by the Open Foundation of Henan Key Laboratory of Cyberspace Situation Awareness(No.HNTS2022020),where Fangfang Shan is an author.Further details can be found at http://xt.hnkjt.gov.cn/data/pingtai/(accessed on 05 June 2024)the Natural Science Foundation of Henan Province Youth Science Fund Project(No.232300420422),and for more information,you can visit https://kjt.henan.gov.cn(accessed on 05 June 2024).
文摘With the emergence and development of social networks,people can stay in touch with friends,family,and colleagues more quickly and conveniently,regardless of their location.This ubiquitous digital internet environment has also led to large-scale disclosure of personal privacy.Due to the complexity and subtlety of sensitive information,traditional sensitive information identification technologies cannot thoroughly address the characteristics of each piece of data,thus weakening the deep connections between text and images.In this context,this paper adopts the CLIP model as a modality discriminator.By using comparative learning between sensitive image descriptions and images,the similarity between the images and the sensitive descriptions is obtained to determine whether the images contain sensitive information.This provides the basis for identifying sensitive information using different modalities.Specifically,if the original data does not contain sensitive information,only single-modality text-sensitive information identification is performed;if the original data contains sensitive information,multimodality sensitive information identification is conducted.This approach allows for differentiated processing of each piece of data,thereby achieving more accurate sensitive information identification.The aforementioned modality discriminator can address the limitations of existing sensitive information identification technologies,making the identification of sensitive information from the original data more appropriate and precise.
基金Supported by the National Natural Science Foundation of China(U24B2031)National Key Research and Development Project(2018YFA0702504)"14th Five-Year Plan"Science and Technology Project of CNOOC(KJGG2022-0201)。
文摘During drilling operations,the low resolution of seismic data often limits the accurate characterization of small-scale geological bodies near the borehole and ahead of the drill bit.This study investigates high-resolution seismic data processing technologies and methods tailored for drilling scenarios.The high-resolution processing of seismic data is divided into three stages:pre-drilling processing,post-drilling correction,and while-drilling updating.By integrating seismic data from different stages,spatial ranges,and frequencies,together with information from drilled wells and while-drilling data,and applying artificial intelligence modeling techniques,a progressive high-resolution processing technology of seismic data based on multi-source information fusion is developed,which performs simple and efficient seismic information updates during drilling.Case studies show that,with the gradual integration of multi-source information,the resolution and accuracy of seismic data are significantly improved,and thin-bed weak reflections are more clearly imaged.The updated seismic information while-drilling demonstrates high value in predicting geological bodies ahead of the drill bit.Validation using logging,mud logging,and drilling engineering data ensures the fidelity of the processing results of high-resolution seismic data.This provides clearer and more accurate stratigraphic information for drilling operations,enhancing both drilling safety and efficiency.
