In this paper,we address the problem of multiple frequency-hopping(FH)signal parameters estimation in the presence of random missing observations.A space-time matrix with random missing observations is acquired by a u...In this paper,we address the problem of multiple frequency-hopping(FH)signal parameters estimation in the presence of random missing observations.A space-time matrix with random missing observations is acquired by a uniform linear array(ULA).We exploit the inherent incomplete data processing capability of atomic norm soft thresholding(AST)to analyze the space-time matrix and complete the accurate estimation of the hopping time and frequency of the received FH signals.The hopping time is obtained by the sudden changes of the spatial information,which is implemented as the boundary to divide the time domain signal so that each segment of the signal is a superposition of time-invariant multiple components.Then,the frequency of multiple signal components can be estimated precisely by AST within each segment.After obtaining the above two parameters of the hopping time and the frequency of signals,the direction of arrival(DOA)can be directly calculated by them,and the network sorting can be realized.Results of simulation show that the proposed method is superior to the existing technology.Even when a large portion of data observations is missing,as the number of array elements increases,the proposed method still achieves acceptable accuracy of multi-FH signal parameters estimation.展开更多
The accurate and intelligent identification of the working conditions of a sucker-rod pumping system is necessary. As onshore oil extraction gradually enters its mid-to late-stage, the cost required to train a deep le...The accurate and intelligent identification of the working conditions of a sucker-rod pumping system is necessary. As onshore oil extraction gradually enters its mid-to late-stage, the cost required to train a deep learning working condition recognition model for pumping wells by obtaining enough new working condition samples is expensive. For the few-shot problem and large calculation issues of new working conditions of oil wells, a working condition recognition method for pumping unit wells based on a 4-dimensional time-frequency signature (4D-TFS) and meta-learning convolutional shrinkage neural network (ML-CSNN) is proposed. First, the measured pumping unit well workup data are converted into 4D-TFS data, and the initial feature extraction task is performed while compressing the data. Subsequently, a convolutional shrinkage neural network (CSNN) with a specific structure that can ablate low-frequency features is designed to extract working conditions features. Finally, a meta-learning fine-tuning framework for learning the network parameters that are susceptible to task changes is merged into the CSNN to solve the few-shot issue. The results of the experiments demonstrate that the trained ML-CSNN has good recognition accuracy and generalization ability for few-shot working condition recognition. More specifically, in the case of lower computational complexity, only few-shot samples are needed to fine-tune the network parameters, and the model can be quickly adapted to new classes of well conditions.展开更多
Least-squares reverse-time migration(LSRTM) formulates reverse-time migration(RTM) in the leastsquares inversion framework to obtain the optimal reflectivity image. It can generate images with more accurate amplitudes...Least-squares reverse-time migration(LSRTM) formulates reverse-time migration(RTM) in the leastsquares inversion framework to obtain the optimal reflectivity image. It can generate images with more accurate amplitudes, higher resolution, and fewer artifacts than RTM. However, three problems still exist:(1) inversion can be dominated by strong events in the residual;(2) low-wavenumber artifacts in the gradient affect convergence speed and imaging results;(3) high-wavenumber noise is also amplified as iteration increases. To solve these three problems, we have improved LSRTM: firstly, we use Hubernorm as the objective function to emphasize the weak reflectors during the inversion;secondly, we adapt the de-primary imaging condition to remove the low-wavenumber artifacts above strong reflectors as well as the false high-wavenumber reflectors in the gradient;thirdly, we apply the L1-norm sparse constraint in the curvelet-domain as the regularization term to suppress the high-wavenumber migration noise. As the new inversion objective function contains the non-smooth L1-norm, we use a modified iterative soft thresholding(IST) method to update along the Polak-Ribie re conjugate-gradient direction by using a preconditioned non-linear conjugate-gradient(PNCG) method. The numerical examples,especially the Sigsbee2 A model, demonstrate that the Huber inversion-based RTM can generate highquality images by mitigating migration artifacts and improving the contribution of weak reflection events.展开更多
Specific emitter identification can distin-guish individual transmitters by analyzing received signals and extracting inherent features of hard-ware circuits.Feature extraction is a key part of traditional machine lea...Specific emitter identification can distin-guish individual transmitters by analyzing received signals and extracting inherent features of hard-ware circuits.Feature extraction is a key part of traditional machine learning-based methods,but manual extrac-tion is generally limited by prior professional knowl-edge.At the same time,it has been noted that the per-formance of most specific emitter identification meth-ods degrades in the low signal-to-noise ratio(SNR)environments.The deep residual shrinkage network(DRSN)is proposed for specific emitter identification,particularly in the low SNRs.The soft threshold can preserve more key features for the improvement of performance,and an identity shortcut can speed up the training process.We collect signals via the receiver to create a dataset in the actual environments.The DRSN is trained to automatically extract features and imple-ment the classification of transmitters.Experimental results show that DRSN obtains the best accuracy un-der different SNRs and has less running time,which demonstrates the effectiveness of DRSN in identify-ing specific emitters.展开更多
Purpose-Recently,the convolutional neural network(ConvNet)has a wide application in the classification of motor imagery EEG signals,However,the low sigalto-noise electroencephalogram(EEG)signals are ollectede under th...Purpose-Recently,the convolutional neural network(ConvNet)has a wide application in the classification of motor imagery EEG signals,However,the low sigalto-noise electroencephalogram(EEG)signals are ollectede under the interference of noises.However,the conventional ConvNet model cannot directly solve this problem.This study aims to discuss the aforementioned issues.Design/methodology/approach-To solve this problem,this paper adopted a novel residual shrinkage block(RSB)to construct the ComvNet model(RSBConvNet).During the feature extraction from EEG simnals,the proposed RSBConvNet prevented the noise component in EEG signals,and improved the classification accuracy of motor imagery.In the construction of RSBConvNet,the author applied the soft thresholding strategy to prevent the non-related.motor imagery features in EEG sigmals.The soft thresholding was inserted into the residual block(RB),and the suitable threshold for the curent EEG signals distribution can be learned by minimizing the loss function.Therefore,during the feature extraction of motor imagery,the proposed RSBConvNet de noised the EEG signals and improved the discriminative of dassifiation features.Findings-Comparative experiments and ablation studies were done on two public benchumark datasets.Compared with conventionalConvNet models,the proposed RSBConvNet model has olbvious improvements in motor imagery classification accuracy and Kappa officient.Ablation studies have also shown the de noised abilities of the RSBConvNet modeL Morbover,different parameters and computational methods of the RSBConvNet model have been tested om the dassificatiton of motor imagery.Originality/value-Based ou the experimental results,the RSBComvNet constructed in this paper has an excellent reogmition accuracy of M-BCI which can be used for further appications for the online MI-BCI.展开更多
Particles,particle aggregates,and reactor walls complicate the dynamic microstructures of circulating fluidized beds(CFBs).Using local solids concentration data from a 10-m-high and 76.2-mm-inner-diameter riser with F...Particles,particle aggregates,and reactor walls complicate the dynamic microstructures of circulating fluidized beds(CFBs).Using local solids concentration data from a 10-m-high and 76.2-mm-inner-diameter riser with FCC(Fluid Catalytic Cracking)particles(dp=67μm,ρp=1500 kg/m^3),this paper presents an improved denoising process for use before nonlinear chaos analysis.Using the soft-threshold denoising method in the wavelet domain with experimental empty bed signals as base data to estimate the denoising threshold,an efficient denoising algorithm was proposed and used for the dynamic signals in CFBs.Analysis shows that for the local solids concentration time series,high-frequency fluctuations may be one of the system properties,while noise interference can also make a low-frequency contribution.An exact denoising method is needed in such cases.The correlation dimension and Kolmogorov entropy were calculated using denoised data and the results showed that the particle behavior in the CFB is highly complex.Generally,two correlation dimensions coexist in a low-flux CFB.The first correlation dimension is low and corresponds to small-scale fluctuations that reveal a high-frequency pseudo-periodic movement,but the second correlation dimension is high and corresponds to large-scale fluctuations that indicate multi-frequency movements,including particle aggregation and breakage.At the same axial level,the first correlation dimensions change slightly with radial position,and the main tendency is high at the center but slightly lower near the wall.However,the second correlation dimensions show large changes along the radial direction,are again high in the core region,and after r/R≥0.6(r as radial position,R as radius of the riser),the dimensions clearly drop down.This indicates that the particle behavior is more complex and has higher degrees of freedom at the center,but clusters near the wall are restrained to some degree because of wall effects.展开更多
文摘In this paper,we address the problem of multiple frequency-hopping(FH)signal parameters estimation in the presence of random missing observations.A space-time matrix with random missing observations is acquired by a uniform linear array(ULA).We exploit the inherent incomplete data processing capability of atomic norm soft thresholding(AST)to analyze the space-time matrix and complete the accurate estimation of the hopping time and frequency of the received FH signals.The hopping time is obtained by the sudden changes of the spatial information,which is implemented as the boundary to divide the time domain signal so that each segment of the signal is a superposition of time-invariant multiple components.Then,the frequency of multiple signal components can be estimated precisely by AST within each segment.After obtaining the above two parameters of the hopping time and the frequency of signals,the direction of arrival(DOA)can be directly calculated by them,and the network sorting can be realized.Results of simulation show that the proposed method is superior to the existing technology.Even when a large portion of data observations is missing,as the number of array elements increases,the proposed method still achieves acceptable accuracy of multi-FH signal parameters estimation.
基金supported in part by the National Natural Science Foundation of China under Grant U1908212,62203432 and 92067205in part by the State Key Laboratory of Robotics of China under Grant 2023-Z03 and 2023-Z15in part by the Natural Science Foundation of Liaoning Province under Grant 2020-KF-11-02.
文摘The accurate and intelligent identification of the working conditions of a sucker-rod pumping system is necessary. As onshore oil extraction gradually enters its mid-to late-stage, the cost required to train a deep learning working condition recognition model for pumping wells by obtaining enough new working condition samples is expensive. For the few-shot problem and large calculation issues of new working conditions of oil wells, a working condition recognition method for pumping unit wells based on a 4-dimensional time-frequency signature (4D-TFS) and meta-learning convolutional shrinkage neural network (ML-CSNN) is proposed. First, the measured pumping unit well workup data are converted into 4D-TFS data, and the initial feature extraction task is performed while compressing the data. Subsequently, a convolutional shrinkage neural network (CSNN) with a specific structure that can ablate low-frequency features is designed to extract working conditions features. Finally, a meta-learning fine-tuning framework for learning the network parameters that are susceptible to task changes is merged into the CSNN to solve the few-shot issue. The results of the experiments demonstrate that the trained ML-CSNN has good recognition accuracy and generalization ability for few-shot working condition recognition. More specifically, in the case of lower computational complexity, only few-shot samples are needed to fine-tune the network parameters, and the model can be quickly adapted to new classes of well conditions.
基金supported by National Key R&D Program of China (No. 2018YFA0702502)NSFC (Grant No. 41974142, 42074129, and 41674114)+1 种基金Science Foundation of China University of Petroleum (Beijing) (Grant No. 2462020YXZZ005)State Key Laboratory of Petroleum Resources and Prospecting (Grant No. PRP/indep-42012)。
文摘Least-squares reverse-time migration(LSRTM) formulates reverse-time migration(RTM) in the leastsquares inversion framework to obtain the optimal reflectivity image. It can generate images with more accurate amplitudes, higher resolution, and fewer artifacts than RTM. However, three problems still exist:(1) inversion can be dominated by strong events in the residual;(2) low-wavenumber artifacts in the gradient affect convergence speed and imaging results;(3) high-wavenumber noise is also amplified as iteration increases. To solve these three problems, we have improved LSRTM: firstly, we use Hubernorm as the objective function to emphasize the weak reflectors during the inversion;secondly, we adapt the de-primary imaging condition to remove the low-wavenumber artifacts above strong reflectors as well as the false high-wavenumber reflectors in the gradient;thirdly, we apply the L1-norm sparse constraint in the curvelet-domain as the regularization term to suppress the high-wavenumber migration noise. As the new inversion objective function contains the non-smooth L1-norm, we use a modified iterative soft thresholding(IST) method to update along the Polak-Ribie re conjugate-gradient direction by using a preconditioned non-linear conjugate-gradient(PNCG) method. The numerical examples,especially the Sigsbee2 A model, demonstrate that the Huber inversion-based RTM can generate highquality images by mitigating migration artifacts and improving the contribution of weak reflection events.
基金the National Natural Science Foundation of China(No.U20B2038,No.61871398,NO.61901520 and No.61931011)the Natural Science Foundation for Distinguished Young Scholars of Jiangsu Province(No.BK20190030)the National Key R&D Program of China under Grant 2018YFB1801103.
文摘Specific emitter identification can distin-guish individual transmitters by analyzing received signals and extracting inherent features of hard-ware circuits.Feature extraction is a key part of traditional machine learning-based methods,but manual extrac-tion is generally limited by prior professional knowl-edge.At the same time,it has been noted that the per-formance of most specific emitter identification meth-ods degrades in the low signal-to-noise ratio(SNR)environments.The deep residual shrinkage network(DRSN)is proposed for specific emitter identification,particularly in the low SNRs.The soft threshold can preserve more key features for the improvement of performance,and an identity shortcut can speed up the training process.We collect signals via the receiver to create a dataset in the actual environments.The DRSN is trained to automatically extract features and imple-ment the classification of transmitters.Experimental results show that DRSN obtains the best accuracy un-der different SNRs and has less running time,which demonstrates the effectiveness of DRSN in identify-ing specific emitters.
基金This work was supported by the Education and Scientific Research Project for Young and Middle-aged Teachers in Fujian Province(No.JAT200581)。
文摘Purpose-Recently,the convolutional neural network(ConvNet)has a wide application in the classification of motor imagery EEG signals,However,the low sigalto-noise electroencephalogram(EEG)signals are ollectede under the interference of noises.However,the conventional ConvNet model cannot directly solve this problem.This study aims to discuss the aforementioned issues.Design/methodology/approach-To solve this problem,this paper adopted a novel residual shrinkage block(RSB)to construct the ComvNet model(RSBConvNet).During the feature extraction from EEG simnals,the proposed RSBConvNet prevented the noise component in EEG signals,and improved the classification accuracy of motor imagery.In the construction of RSBConvNet,the author applied the soft thresholding strategy to prevent the non-related.motor imagery features in EEG sigmals.The soft thresholding was inserted into the residual block(RB),and the suitable threshold for the curent EEG signals distribution can be learned by minimizing the loss function.Therefore,during the feature extraction of motor imagery,the proposed RSBConvNet de noised the EEG signals and improved the discriminative of dassifiation features.Findings-Comparative experiments and ablation studies were done on two public benchumark datasets.Compared with conventionalConvNet models,the proposed RSBConvNet model has olbvious improvements in motor imagery classification accuracy and Kappa officient.Ablation studies have also shown the de noised abilities of the RSBConvNet modeL Morbover,different parameters and computational methods of the RSBConvNet model have been tested om the dassificatiton of motor imagery.Originality/value-Based ou the experimental results,the RSBComvNet constructed in this paper has an excellent reogmition accuracy of M-BCI which can be used for further appications for the online MI-BCI.
文摘Particles,particle aggregates,and reactor walls complicate the dynamic microstructures of circulating fluidized beds(CFBs).Using local solids concentration data from a 10-m-high and 76.2-mm-inner-diameter riser with FCC(Fluid Catalytic Cracking)particles(dp=67μm,ρp=1500 kg/m^3),this paper presents an improved denoising process for use before nonlinear chaos analysis.Using the soft-threshold denoising method in the wavelet domain with experimental empty bed signals as base data to estimate the denoising threshold,an efficient denoising algorithm was proposed and used for the dynamic signals in CFBs.Analysis shows that for the local solids concentration time series,high-frequency fluctuations may be one of the system properties,while noise interference can also make a low-frequency contribution.An exact denoising method is needed in such cases.The correlation dimension and Kolmogorov entropy were calculated using denoised data and the results showed that the particle behavior in the CFB is highly complex.Generally,two correlation dimensions coexist in a low-flux CFB.The first correlation dimension is low and corresponds to small-scale fluctuations that reveal a high-frequency pseudo-periodic movement,but the second correlation dimension is high and corresponds to large-scale fluctuations that indicate multi-frequency movements,including particle aggregation and breakage.At the same axial level,the first correlation dimensions change slightly with radial position,and the main tendency is high at the center but slightly lower near the wall.However,the second correlation dimensions show large changes along the radial direction,are again high in the core region,and after r/R≥0.6(r as radial position,R as radius of the riser),the dimensions clearly drop down.This indicates that the particle behavior is more complex and has higher degrees of freedom at the center,but clusters near the wall are restrained to some degree because of wall effects.