27 results about "Hat matrix" patented technology

Auto-tuning covariances associated with a set of noise models for a variety of sensor modalities and / or perception components such that the covariances are leveled respective to one another may include whitening the covariances and / or error models and determining scalars to apply to the covariances. Determining these scalars may comprise using the residuals that result from generating the set of noise model (e.g., such as may be determined as part of least squares estimation) along with the hat matrix of the process model to determine the scalars. The covariances may iteratively be updated until the scalar adjustments converge or until another end condition is met.

The invention discloses a B5G base station T / R assembly health measurement method and system based on sparse projection and a hidden Markov model, and the method comprises the steps: collecting the multi-source operation data of a T / R assembly, and constructing a multi-dimensional operation parameter time sequence; a sparse projection matrix is constructed through covariance analysis and eigenvalue decomposition, and high-dimensional operation parameters are mapped into low-dimensional sparse health eigenvectors; training a hidden Markov model based on the feature sequence in the normal state, fitting an observation probability by using a Gaussian mixture model, and establishing a normal state reference model; calculating a KL distance between the current feature sequence distribution and the reference distribution, and mapping the KL distance into a normalized health degree index; and adaptively determining a state division threshold value by using a K-means clustering algorithm to realize health state grading of the T / R assembly. According to the method, environmental noise is effectively stripped through sparse projection, the dynamic reference model is utilized to adapt to complex working conditions, and online sensing and accurate evaluation of early weak degradation of the B5G base station assembly are realized.

The invention provides a non-stationary industrial process monitoring method and system. The method comprises an offline training stage: calculating a time Laplacian matrix and a space Laplacian matrix based on a historical data matrix; constructing an objective function of the stationary subspace analysis method, and adding a time constraint term of a time Laplacian matrix and a space constraint term of a space Laplacian matrix into the objective function; solving the objective function to obtain a stable projection matrix; calculating a stationary component and a monitoring index of each sample in the data matrix X in sequence; determining a control limit by using a kernel density estimation method; an online monitoring stage: based on the real-time operation data x, calculating a stationary component of the real-time operation data x and a corresponding real-time monitoring index according to the stationary projection matrix, and if the real-time monitoring index is greater than a control limit, judging that the operation of the non-stationary process has a fault; the monitoring accuracy can be improved.

A plant three-dimensional reconstruction method for a weak texture image and scale distortion comprises the following steps: step 1, recovering an initial Gaussian point cloud and camera parameters through an SfM method by using a multi-view RGB image; 2, differential rendering is executed in training iteration, a predicted image is compared with a real image, and a loss function containing various constraints is calculated to serve as a basis for gradient solving and parameter updating; 3, calculating a pixel weighted average gradient and an artifact suppression coefficient of each Gaussian in a visible view angle; 4, updating Gaussian parameters through back propagation, and adaptively triggering Gaussian cloning, splitting or deleting operation based on the pixel gradient, the covariance matrix and transparency; and 5, dynamically adjusting the Gaussian projection matrix according to the focal length and the resolution of the camera during rendering so as to maintain scale consistency, and performing anti-aliasing rendering in combination with pixel-level super-sampling. According to the method, the reconstruction precision and the structure reduction capability of the virtual plant in the weak texture region are remarkably improved.

The application discloses a near-field strong interference suppression method based on spherical wave deconvolution beam forming positioning, which comprises the following steps: performing fine grid division on a near-field scanning area of a linear array; performing spherical wave focusing beam forming on the near-field scanning area and two-dimensional deconvolution beam forming about distance-angle under the condition of unknown sound source position; calculating the distribution entropy characteristics of the deconvolution beam forming result in the angle dimension, and using a preset threshold to correlate the interference positioning of a target; constructing an array flow pattern matrix A of each near-field target l and generating a covariance matrix A; calculating the orthogonal complement space projection matrix of the covariance matrix A, and applying the orthogonal complement space projection matrix to the array element domain covariance matrix C to perform near-field interference suppression; using a far-field plane wave model to perform deconvolution beam forming calculation on the covariance matrix for suppressing near-field interference, so as to obtain passive detection results of the near-field strong interference suppression. The application can realize near-field interference positioning and interference suppression under the condition that the interference position is unknown.

Provided are a URA calculation method and device for a low-orbit navigation satellite, a computer storage medium and a computer program product. The URA calculation method comprises: in precise orbit determination and clock error calculation for a low-orbit satellite, a covariance submatrix of a predicted orbit and clock error parameter in a predicted arc segment is given by using a transfer matrix, denoted as a first covariance submatrix; a projection matrix from a geocentric geodetic coordinate system to a satellite orbit coordinate system is calculated; the first covariance submatrix is converted into a second covariance submatrix of the satellite orbit coordinate system according to the projection matrix; the satellite orbit prediction accuracy and the clock error prediction accuracy are calculated according to the second covariance submatrix; and the URA is calculated according to the satellite orbit prediction accuracy and the clock error prediction accuracy. The present disclosure proposes a URA calculation method for a low-orbit satellite based on a covariance matrix, which guarantees the real-time performance and reliability of integrity monitoring.

The application discloses a medical image class incremental learning method and system based on feature principal direction guided projection, and belongs to the technical field of medical image analysis; the method acquires medical image features through a pre-trained feature extractor; random projection is performed on the features to obtain first-view features; a covariance matrix is calculated based on a feature cache area and is decomposed to extract the first K principal directions, construct a guided projection matrix, and obtain second-view features; the double-view features are spliced to obtain final representation; the classifier weight is directly calculated through a ridge regression closed-form solution without backward propagation training; the application adaptively captures the core structure of the feature space through data-driven principal direction projection, retains the detailed information in combination with random projection, effectively improves the feature separability of fine-grained classes, adopts an analytical classifier updating mechanism, realizes stable class incremental learning without storing historical data, improves the model updating efficiency, and is suitable for efficient continuous learning tasks in various medical image scenes.

The invention discloses a circuit fault diagnosis method based on a fractional order observer, and the method comprises the steps: constructing a fractional order circuit model based on the Kirchhoff's law, and describing a capacitor end voltage and an inductive current as a Caputo fractional order derivative form; a double-disturbance decoupling mechanism is designed, total disturbance is decomposed into a decoupling disturbance component and a non-decoupling disturbance component, physical isolation of the decoupling disturbance is achieved through a projection matrix, and the influence of the non-decoupling disturbance is eliminated through a dynamic suppression coefficient. Constructing a state observer to generate a state estimation error and a fault error vector, and dynamically compensating a fault estimation value in combination with an adaptive gain matrix; and ensuring the system stability through a Lyapunov function, and proving that an error state is converged to a preset boundary domain under the condition of satisfying a characteristic value correlation matrix inequality. Compared with the prior art, rapid and accurate diagnosis of faults such as inductance saturation and capacitance aging can be realized.

The application discloses a network fault positioning fast response system based on distributed log analysis, relates to the technical field of computer network and operation and maintenance, and comprises the following modules: a data acquisition module, which generates time series data frames based on a sliding window; a state vector construction module, which extracts indexes and constructs standardized state vectors; a covariance matrix updating module, which recursively updates a dynamic covariance matrix by using a forgetting factor; a feature space decomposition module, which constructs a residual projection matrix based on dynamic principal component dimension parameters; an anomaly detection module, which calculates residual anomaly energy and determines a fault candidate set based on contribution degrees; and a parameter self-adaptive module, which calculates residual space entropy values and feeds back to adjust principal component dimension parameters of the next frame; and the application realizes adaptive and accurate bounding of gray faults through entropy feedback closed loop and orthogonal projection.

The invention relates to the technical field of natural language processing, in particular to a model prefix parameter and hyper-parameter joint optimization method and system. Comprising the following steps: S1, acquiring a high-dimensional prefix parameter and hyper-parameter joint optimization instruction; s2, generating a random projection matrix, and mapping the high-dimensional prefix parameters into low-dimensional prefix parameters through the projection matrix; s3, searching an optimal low-dimensional prefix parameter in a low-dimensional space through a covariance matrix adaptive evolution strategy, reconstructing the low-dimensional prefix parameter into a dynamic high-dimensional MLP parameter, generating a key-value prefix required by each layer of the model, and embedding and injecting the key-value prefix into the large model; s4, training hyper-parameters are dynamically adjusted based on a particle swarm algorithm and a cosine annealing strategy, and the hyper-parameters are transited step by step in the training process; s5, training the model based on the optimal prefix parameter and the hyper-parameter, and outputting the trained model; and S6, applying the trained model to the medical question and answer task. According to the method, the precision and stability of prefix tuning in the medical question and answer task can be improved.

The application discloses an industrial process anomaly detection method and device and a computer device, relates to the technical field of industrial control, and comprises the following steps: obtaining target working condition data of an industrial process, performing working condition identification on the target working condition data according to a target clustering algorithm, and obtaining a target working condition type corresponding to the target working condition data. According to a target projection matrix of the target working condition type, an inverse matrix of a covariance matrix of a hidden variable under the target working condition type and the target working condition data, T 2 statistics and SPE statistics are calculated. 2 Then, according to the comparison result of the target T 2 statistics and the target T 2 statistics and the comparison result of the SPE statistics and the target SPE statistics, whether an abnormal situation exists at the collection time of the target working condition data and the cause of the abnormal situation are determined. The embodiment of the application can realize accurate detection of the industrial process.

The invention discloses a network fault positioning quick response system based on distributed log analysis, and relates to the technical field of computer networks and operation and maintenance, and the system comprises a data acquisition module which generates a time sequence data frame based on a sliding window; the state vector construction module is used for extracting indexes and constructing standardized state vectors; the covariance matrix updating module is used for recursively updating the dynamic covariance matrix by using the forgetting factor; the feature space decomposition module is used for constructing a residual projection matrix based on the dynamic principal component dimension parameters; the anomaly detection module is used for calculating residual anomaly energy and determining a fault candidate set based on the contribution degree; the parameter self-adaptive module is used for calculating a residual spatial entropy value and feeding back and adjusting a principal component dimension parameter of a next frame; through entropy feedback closed loop and orthogonal projection, adaptive accurate delimitation of the gray scale fault is realized.

The present application is to solve the problems of spectrum leakage, large loss of useful signal and insufficient interference locking precision in the existing narrow-band interference suppression technology; a narrow-band interference suppression method for Beidou navigation is provided, comprising the following steps: S101: discrete signal sequence sampling and segmented caching; S102: constructing an adaptive signal covariance matrix; S103: performing fast eigenvalue decomposition on the covariance matrix; S104: determining the number of interference components based on the eigenvalue distribution; S105: constructing an interference subspace projection matrix; S106: performing orthogonal projection to suppress narrow-band interference; S107: signal gain compensation and output reconstruction; S108: detecting whether the interference environment has changed dramatically; all components pointing to the interference in the signal space are filtered through the orthogonal projection matrix, avoiding the convergence problem of time-domain adaptive filtering, and effectively dealing with the strong narrow-band interference scene with high power density.

The present application belongs to the technical field of multi-channel signal detection, and particularly relates to a self-adaptive subspace detection method and system suitable for a partially homogeneous environment, comprising the following steps: Step 1: constructing a target signal subspace matrix and an interference signal subspace matrix; Step 2: constructing a sample covariance matrix by using a training sample; Step 3: constructing a quasi-whitening matrix by using the sample covariance matrix, and then obtaining a whitening matrix by performing quasi-whitening processing; Step 4: constructing an orthogonal projection matrix of the target signal and a corresponding complementary projection matrix; Step 5: constructing an intermediate variable matrix and a non-zero eigenvalue decomposition, and obtaining a non-uniform parameter; Step 6: constructing a detection statistic by using a to-be-detected data matrix, the orthogonal projection matrix of the target signal, the orthogonal projection matrix of the interference signal and the non-uniform parameter; Step 7: determining a detection threshold; and Step 8: determining whether the target exists. The present application improves the detection performance of an airborne radar in a partially homogeneous environment.

The application provides a target angle measurement method, an electronic device, a storage medium and a program product, which can be used in the technical field of array signal processing. The method comprises the following steps: initializing an angle search grid of a detection device, and forming a plurality of angle combinations based on search angles in the angle search grid; wherein the angle combination comprises a first angle and a second angle, and the first angle is smaller than the second angle; constructing a sample covariance matrix based on amplitude-phase data of a single fast shot in all virtual channels of the detection device; for each angle combination, obtaining a maximum likelihood function corresponding to the angle combination based on a projection matrix of the angle combination and the sample covariance matrix; and obtaining a measurement angle of a target to be measured based on the angle combination and the maximum likelihood function. The method provided by the application does not need to know the number of targets, and can realize target angle measurement under a single fast shot, thereby significantly improving the real-time performance of target angle measurement.

This invention relates to a multi-beam array interference suppression method combining time-frequency analysis and symbol characteristics, belonging to the field of array interference suppression. The method includes: sampling and truncating the received signal, performing a Fast Fourier Transform (FFT) to transform it to the time-frequency domain and calculating the sampling covariance matrix; calculating the projection matrix and performing interference suppression; reallocating subcarriers and performing an inverse FFT; synchronizing the first and second paths of the desired signal respectively; extracting the two path signals using the cyclic characteristics of the cyclic prefix; performing OFDM demodulation and frequency domain channel estimation on the extracted signals; calculating weights according to the maximum signal-to-noise ratio (SNR) criterion; and weighting the two path signals to obtain the output signal. This invention proposes a multi-beam array interference suppression method combining time-frequency analysis and symbol characteristics. It uses an orthogonal projection algorithm in the time-frequency domain to suppress strong interference, extracts the two stronger paths using the cyclic characteristics of the cyclic prefix, and merges the two paths using the maximum SNR criterion, effectively improving the reliability of the receiver.

The application discloses a kind of based on airborne circular array rotation measurement fusion's joint direction finding and amplitude-phase error estimation method, comprising: the array receiving data model of constructing airborne circular array in amplitude-phase error environment, control airborne uniform circular array step rotation, obtain array receiving data, calculate covariance matrix, extract noise subspace;Noise subspace and the orthogonality of steering vector are used to construct joint orthogonal projection matrix and quadratic optimization objective function;Introduce constraint condition to eliminate scale ambiguity, deduce the closed-form analytical solution of the amplitude-phase error using Lagrange multiplier method;The analytical solution is substituted back into the quadratic optimization objective function, and the direction of arrival estimation value of radiation source is obtained by one-dimensional spectrum search;According to the direction of arrival estimation value, calculate total joint orthogonal projection matrix, and solve the closed-form analytical solution of full-channel unknown amplitude-phase error.The application effectively overcomes the underdetermined problem of static observation, and significantly improves the direction finding accuracy and calculation stability in strong error environment.

The invention discloses a random subspace modal parameter uncertainty quantitative accelerated calculation method based on data-driven random subspace recognition, and belongs to the technical field of structural health monitoring and operation modal analysis. The method comprises the following steps: collecting multi-channel response data and setting parameters; constructing a block Hankel matrix and calculating a sample correlation matrix, and extracting and multiplexing a group of invariant sub-block intermediate quantities corresponding to the data set from the correlation matrix; the SVD module is used for constructing a reference projection matrix and carrying out SVD decomposition to obtain a subspace base; according to the method, matrix-level first-order disturbance propagation is implemented based on block data in uncertainty quantization, and the variance / covariance of modal parameters is obtained by combining one-step eigenvector disturbance and non-iterative eigenvalue disturbance, so that repeated construction of a high-dimensional intermediate matrix and modal-by-modal disturbance solution in a traditional method are avoided, the calculation amount and storage requirements are remarkably reduced, and the method is suitable for large-scale popularization and application. The engineering expandability is improved; and meanwhile, the statistical consistency of modal parameters and uncertainty estimation of the modal parameters is kept.

This invention discloses a robust beamforming method based on interference signal steering vector estimation, comprising: determining a projection matrix of the signal subspace based on the sample covariance matrix of the echo signal; determining a projection matrix of the interference subspace based on the covariance matrix of the interference signal; projecting the hypothetical interference signal steering vector multiple times onto the signal subspace and the interference subspace using the projection matrices of the signal subspace and the interference subspace respectively, to obtain a final projected steering vector; estimating the estimated steering vector of the interference signal based on the final projected steering vector; and forming a robust beam based on the estimated steering vector of the interference signal. According to the method provided by this invention, by projecting the hypothetical interference signal steering vector multiple times onto the signal subspace and the interference subspace respectively, the influence of array correction errors such as array position errors and array structure errors, as well as array element amplitude and phase errors, on beamforming performance can be effectively reduced, achieving robust beamforming.

This invention discloses a kernelized inverse nearest neighbor discriminant analysis method, comprising the following steps: obtaining training image samples; mapping the input data to a high-dimensional space using a Gaussian kernel function; obtaining the representations of the intra-class and inter-class scatter matrices of the training image samples in the high-dimensional space using kernel tricks and the inverse nearest neighbor algorithm; learning the projection matrix that maximizes the inter-class scatter matrix and minimizes the intra-class scatter matrix through feature decomposition; extracting features from the training and test image samples using the projection matrix; and classifying the test samples using the nearest neighbor algorithm. This invention, for the first time, extends the inverse nearest neighbor linear discriminant analysis method to a high-dimensional space to solve the classification problem of nonlinear data. It proposes using a Gaussian kernel function for high-dimensional mapping, kernelizing the algorithm, and using kernel tricks for non-explicit mapping derivation. High-dimensional inverse nearest neighbors are established in the space after Gaussian function mapping to obtain the scatter matrix, wherein the kernel trick is used to realize the distance representation in the high-dimensional space.

This invention proposes a priori-enhanced multi-kernel canonical variable analysis method and device for monitoring blast furnace ironmaking. First, historical multivariate data is collected, standardized, and past and future Hankel matrices are constructed. Then, a CSI-MKCVA model is constructed, and an enhanced hybrid kernel function is built to simultaneously capture local, global, and time-related nonlinear features of the blast furnace ironmaking process, and the covariance matrix and Laplace matrix are reconstructed using weighted averages. Next, the projection matrices of past and future data are iteratively trained to extract priori spatiotemporal nonlinear features (CSNF) with conformal and discriminative capabilities. Finally, based on the extracted CSNF, the kernel density estimation method is used to calculate control limits, enabling real-time monitoring. This invention enhances the spatial correlation modeling and fault category identification capabilities in the blast furnace ironmaking process, significantly improving the anomaly monitoring accuracy under conditions of data nonlinearity, dynamism, and class imbalance, ensuring the safe and stable operation of blast furnace production.

The application provides a non-stationary industrial process monitoring method and system, and the method comprises an offline training stage: a time Laplacian matrix and a space Laplacian matrix are calculated based on a historical data matrix; a target function of a stationary subspace analysis method is constructed, and a time constraint term of the time Laplacian matrix and a space constraint term of the space Laplacian matrix are added to the target function; the target function is solved to obtain a stationary projection matrix; stationary components and monitoring indexes of each sample in the data matrix X are calculated in turn based on the data matrix X; a control limit is determined by using a kernel density estimation method; and an online monitoring stage: based on real-time running data x, stationary components of the real-time running data x and corresponding real-time monitoring indexes are calculated according to the stationary projection matrix, and if the real-time monitoring indexes are greater than the control limit, it is judged that a non-stationary process operation has a fault; and the application can improve monitoring accuracy.

The application discloses a satellite denial environment positioning method based on multi-source opportunity signal fusion, and relates to the technical field of satellite navigation and multi-source fusion positioning. In view of the sparse bias caused by the multipath effect in the urban canyon and the variance and scale mismatching problem existing in the GNSS, 5G and DTMB heterogeneous signal fusion, the application firstly constructs a unified heterogeneous observation model and explicitly introduces a sparse bias term; secondly, a diagonal pre-whitening matrix is constructed by using a measurement noise covariance matrix to multiply a residual error equation, so that the heterogeneous signals are unified to a comparable statistical scale; then, a projection matrix is introduced in the pre-whitening domain to eliminate state increment interference, and L1 regularization is combined to accurately solve a sparse bias estimation value; finally, the bias is deducted from the original observation to obtain pure innovation, and an extended Kalman filter is input to complete state updating. Without directly eliminating abnormal links, the application adaptively suppresses the non-line-of-sight bias, and significantly improves the positioning precision and robustness in a complex environment.

The invention relates to the technical field of video data processing, in particular to an AI-based video editing synthesis method and system, and the method comprises the steps: constructing an optical flow validity index through optical flow direction consistency statistics and an amplitude variation coefficient, and recognizing a cross-fragment optical flow failure boundary; determining an effective reference frame range of each fragment on two sides of the boundary by taking an accumulated value of the inter-frame normalized optical flow amplitude as a stop condition; constructing an influence factor of each reference frame according to the time sequence neighbor component and the motion quality component; constructing a weighted covariance matrix according to the influence factors, adaptively selecting effective structure feature dimensions through eigenvalue decomposition, and respectively forming projection matrixes of fragments on two sides; and under the joint constraint of the bilateral projection matrix, minimizing the bilateral projection residual error by taking the effective reference frame count as a weighting coefficient, and generating a transition frame. According to the method, the problems of reference frame range determination, reference frame contribution quantization and effective feature extraction in the AI-based transition frame generation process in the video editing and splicing process are solved.

The invention discloses an ovarian cancer multi-index combined diagnosis security federal learning method based on flexible access control, and belongs to the technical field of medical artificial intelligence and privacy calculation crossing. The method comprises three parts of a security initialization stage, federated principal component analysis and a federated genetic algorithm. In the security initialization stage, secret keys are securely distributed for a client and a server in federated learning through attribute-based encryption, homomorphic encryption and digital signature technologies, and a fine-grained access control strategy is set. In the federal principal component analysis stage, a client side locally calculates a covariance matrix, encrypts and signs the covariance matrix and then uploads the covariance matrix to a server for safe aggregation, and the server carries out characteristic decomposition and generates a disturbance projection matrix. In a federated genetic algorithm stage, a client side locally calculates fitness, a server safely aggregates and generates a new population, and a globally optimal solution is obtained through multiple rounds of iteration. According to the method, data privacy and model security are effectively protected through multiple encryption and signature mechanisms.

The application discloses an image recognition method based on quaternion generalized kernel sparse principal component analysis, and the method comprises the following steps: acquiring training and test sample images; extracting the entropy, red, green and blue four component information of each image, and performing quaternion matrix representation on the information to construct a corresponding quaternion real representation matrix; constructing a corresponding quaternion covariance kernel matrix and a quaternion p-norm Euclidean distance according to the quaternion real representation matrix, and then constructing a quaternion generalized kernel sparse principal component analysis optimization model; solving the optimization model, taking the calculated kernel sparse principal components of the training sample in the row and column directions as the final solution; calculating the projection matrix of the training and test sample covariance kernel matrix according to the final solution of the model in the row and column directions; and using the quaternion p-norm Euclidean distance to recognize the category to which the images in the test sample set belong, so that the recognition accuracy and robustness are improved.

The invention relates to the field of data processing, in particular to a lithium battery all-parameter real-time intelligent monitoring method and system, and the method comprises the steps: collecting all parameters of a lithium battery pack in real time, carrying out the correlation analysis of the all parameters and the hot-spot temperature, and constructing a feature vector; presetting a sliding window, correcting a covariance matrix of the sliding window based on mutual information of each feature vector and the hot spot temperature and a Pearson's correlation coefficient, presetting a temperature sensitive projection coefficient, maximizing the covariance matrix and the sum of squares of the preset temperature sensitive projection coefficient and the hot spot temperature covariance; obtaining a projection matrix and calculating a temperature sensitive projection coefficient; taking the temperature-sensitive projection coefficient as single control input, scheduling the rotating speed of an exhaust motor and calculating cooling income; and by taking the cooling income as a supervision signal, iteratively updating the projection matrix through a gradient rising method, so that the temperature-sensitive projection coefficient is continuously aligned with the optimal direction of energy-saving temperature control, explosion-proof air cooling millisecond response and hot spot early warning are realized, and the power consumption of a fan is reduced.