12 results about "Higher-order singular value decomposition" patented technology

The invention relates to the technical field of power equipment state monitoring, and particularly discloses a power cable electrical performance detection method and system, and the method comprises the steps: synchronously collecting a broadband electromagnetic signal, a mechanical vibration signal and a temperature signal at a cable monitoring point; calculating a wavelet coherence coefficient between the signals through continuous wavelet transform, and constructing a multi-modal coupling tensor fusing amplitude and cross-modal time-frequency correlation characteristics; performing time slicing and high-order singular value decomposition on the tensor to obtain a time-varying core tensor sequence, mapping the time-varying core tensor sequence into a high-dimensional manifold curve, and generating a system state fingerprint by calculating local curvature distribution and topology invariants of the curve; inputting the fingerprints into a pre-trained defect prediction model, and directly outputting defect inoculation probability and evolution stage judgment; according to the method, the limitation that early weak defect detection is not sensitive in a traditional method is broken through, and early warning and accurate diagnosis of cable insulation latent defects are achieved.

The invention discloses a vector hydrophone array orientation estimation method for performing tensor decomposition by using a propagation operator, relates to the technical field of vector hydrophone array orientation estimation, and discloses a vector hydrophone array orientation estimation method for performing tensor decomposition by using a propagation operator. The method comprises the following steps: firstly, constructing a three-dimensional array manifold tensor composed of an array direction matrix and a vector hydrophone output matrix; respectively expanding received signal tensors according to three modes, solving a propagation operator based on a column block covariance matrix, constructing a normalized signal subspace, establishing a spatial spectrum function with a noise subspace, and obtaining a pitch angle and an azimuth angle of a sound source through spectrum peak search; according to the method, high-order singular value decomposition is avoided, the operand is greatly reduced, meanwhile, high resolution and low sidelobe direction finding performance are kept, and the method is suitable for a real-time underwater acoustic direction finding system of a ship-borne platform, a buoy platform and an unmanned platform.

The invention discloses an early-stage arc risk prediction method based on an MSC-Informer algorithm, and the method comprises the steps: fusing the electrical quantity data collected by the edge end of a power system at high frequency and the multi-modal data collected by the cloud end at low frequency, and constructing a time-modal-feature three-order tensor; performing high-order singular value decomposition on the third-order tensor, decoupling an interaction relationship among time, modality and feature dimensions, and extracting multi-modal fusion features; the multi-modal fusion features are input to an MSC-Informer prediction model, the MSC-Informer prediction model comprises an MSC multi-scale convolution module and an Informer network with a probability sparse attention mechanism, and the MSC multi-scale convolution module extracts local mutation features in the electric signals by paralleling convolution kernels of different scales; the Informer network captures a long-range dependency relationship in the electric signal; and calculating a comprehensive risk score through weighted fusion on the basis of multiple arc early risk state probabilities output by the MSC-Informer prediction model, and dividing arc early risk grades according to a preset threshold.

This invention provides a method for synchronously extracting multi-channel composite fault features of bearings in the running gear of electric locomotives. The method mainly includes: Step S1, preprocessing the multi-channel vibration signals of composite faults in the bearings of electric locomotives using a tensor synchronous denoising method based on higher-order singular value decomposition; Step S2, adaptively filtering and decomposing the preprocessed signals using a multi-layer K-value MVMD algorithm; and Step S3, calculating the fault peak factor of the envelope spectrum of each channel and plotting a tower-shaped EC diagram. Based on the tower-shaped EC diagram, the optimal analysis result of the multi-channel vibration signal is selected and output, synchronously extracting the composite fault features of the bearings. This method overcomes the problems of severe noise interference and ineffective detection of composite faults encountered by traditional methods when processing multi-channel signals of composite faults in electric locomotive bearings. It achieves synchronous and visualized output of composite faults, providing a strong basis for the extraction and identification of weak and composite fault features in complex dynamic signals of the running gear.

The invention relates to the technical field of vehicle data processing, and provides a safety belt and air bag cooperative control method and system based on multi-mode signals. Vehicle data and user data are collected, and processing results obtained through standardization processing are combined to generate a multi-modal signal tensor; decomposing the signal tensor through high-order singular value decomposition to obtain a collision mode component representing collision dynamic characteristics and a user characteristic component representing user biomechanical characteristics; performing feature extraction on the two components through a long-short-term memory network to obtain a time-space dependent damage risk probability; and modeling the safety belt controller and the airbag controller based on the injury risk probability to generate a safety function, and solving the safety function to generate a cooperative control strategy. According to the scheme, the safety belt and the air bag can achieve the synergistic effect according to different conditions, and the reliability of the safety device in vehicle safety control and the safety of passengers in collision accidents are effectively improved.

The invention relates to the technical field of data processing, in particular to a multidimensional data analysis method based on machine learning, which comprises the steps of data preprocessing, feature extraction and fusion, model construction and optimization, and data analysis and decision making. During preprocessing, an improved isolated forest algorithm is used for removing abnormal values, a Bayesian network is used for supplementing missing values, minimum-maximum scaling and logarithm transformation are combined with normalization data, in feature extraction fusion, a high-order singular value decomposition tensor is combined with an attention mechanism for weighting fusion components, and during model construction optimization, a DDQN architecture is improved, and the probability that the model is optimized is lowered. Parameters are updated in combination with empirical regression and a strategy gradient algorithm, finally processed data are input, and decision suggestions are generated by using a multi-objective decision and a Pareto frontier analysis method; the objective of the invention is to solve the problems of insufficient processing precision of abnormal values and missing values during preprocessing of multi-dimensional data, incapability of dynamically capturing key information by feature extraction and fusion, weak model generalization ability and difficulty in processing multi-target conflicts.

This invention discloses a monitoring and evaluation system and method for conjugate soil slope restoration based on multi-source sensing, belonging to the field of civil engineering monitoring and safety assessment technology. The system includes a heterogeneous sensing layer, an edge fusion layer, a cloud evolution layer, and an autonomous decision-making layer. The method includes: acquiring multi-dimensional state information of the conjugate system through a biomimetic multi-source sensor network; constructing a four-dimensional time-series tensor model and using high-order singular value decomposition with confidence constraints for data fusion and restoration; dynamically evaluating the system's health based on a physical information element learning network; and predicting failure paths and generating graded early warnings using a micro-damage propagation graph network. This invention treats the restoration body and the original soil as a conjugate system, realizing a shift from phenomenon monitoring to mechanism inversion and from static evaluation to dynamic prediction, solving the problems of isolated data, static evaluation, and delayed early warning in traditional monitoring.

The application discloses a four-dimensional clutter suppression method and system for ultrasonic micro blood flow imaging, and belongs to the field of ultrasonic imaging. The method is applied to a four-dimensional data set after multi-frame multi-angle plane wave transmission and beam synthesis based on time accumulation, utilizes the angular coherence principle, that is, main lobe components have high coherence in angular space while side lobe components are opposite, and adopts high-order singular value decomposition for clutter suppression. The four-dimensional clutter filtering method based on angles expands the data dimension (increases the feature space), increases the information quantity, is beneficial to blood flow separation, can effectively suppress clutter with fewer frames, can effectively improve the micro blood flow imaging quality in small data sets, requires smaller multi-angle data quantity, has lower calculation cost, and solves the problems in the prior art.

The application relates to the technical field of power equipment state monitoring, and specifically discloses a power cable electric performance detection method and system, wherein wideband electromagnetic signals, mechanical vibration signals and temperature signals at cable monitoring points are synchronously collected; wavelet coherence coefficients among the signals are calculated through continuous wavelet transformation, and a multi-modal coupling tensor fusing amplitude and cross-modal time-frequency correlation characteristics is constructed; time slicing and high-order singular value decomposition are performed on the tensor to obtain a time-varying core tensor sequence, which is mapped into a high-dimensional manifold curve, a system state fingerprint is generated by calculating the local curvature distribution and topological invariant of the curve; and the fingerprint is input into a pre-trained defect prediction model to directly output the incubation probability and evolution stage of defects; the application breaks through the limitation of traditional methods that are not sensitive to early weak defects, and realizes early warning and accurate diagnosis of cable insulation latent defects.

The application discloses a kind of underwater target positioning methods based on underwater acoustic particle velocity polarization processing.The method includes: arranging vector hydrophone array and carrying out underwater acoustic signal acquisition, and write the vector received signal of vector hydrophone array into tensor form;The tensor received signal of vector hydrophone array is carried out high-order singular value decomposition and the noise subspace of each dimension information matrix is solved by truncation to n module expansion left singular matrix;Array manifold matrix of vector hydrophone array is constructed and is written into polarization array manifold tensor;Azimuth-polarization two-dimensional spectrum search is carried out using array manifold tensor and noise subspace, and the azimuth and polarization parameters of the detected target are obtained from two-dimensional spectrum peak value.The underwater target positioning method provided by the application can accurately estimate the polarization parameters and the angle of arrival of incident signal, which helps to complete the underwater acoustic signal information based on frequency, wave number, mode and other characteristics by means of polarization parameters, and significantly improves the underwater target detection capability.

This invention relates to a hyperspectral image denoising method based on tensor-based higher-order singular value decomposition (HDVDe), belonging to the field of hyperspectral image processing technology. To overcome the problems in existing technologies, this invention provides a hyperspectral image denoising method based on HDVDe, comprising: dividing the hyperspectral image tensor into blocks and merging similar blocks into clusters; establishing a hyperspectral image denoising optimization model based on the nonlocal self-similarity and global correlation in the spectral domain of the hyperspectral image; solving the hyperspectral image denoising optimization model using the ADM algorithm; and restoring a noise-free hyperspectral image by combining noise-free image blocks. This invention effectively improves the problem of imbalanced tensor expansion matrices by separating the nonlocal self-similarity domain information, spectral domain information, and spatial domain information of the hyperspectral image to apply low-rank constraints in the spectral domain, thus greatly utilizing the global and local correlations in the spectral domain of the hyperspectral image.

This invention discloses an underwater sound source matching field localization method based on higher-order singular value decomposition of tensor signals, comprising the following steps: Step 1, reconstructing the spatiotemporal frequency multidimensional data recorded by the vertical receiving array into tensor data; Step 2, calculating the tensor signal subspace based on higher-order singular value decomposition and singular matrix truncation. Since singular value decomposition is performed on the matrix expanded in each dimension of the tensor, noise can be further suppressed, resulting in a more accurate signal subspace and improving the accuracy of target localization; Step 3, specifying the receiving array configuration and inputting marine environmental information such as hydrological conditions, seabed topography, and seabed sediment. This invention, while accurately estimating the target location, has stronger background suppression capabilities and outperforms conventional matching field processing.