56 results about "Symmetric matrix" patented technology

The invention discloses a digital human sentiment classification method and device, a computer program product and electronic equipment. The method comprises the steps that a to-be-classified emotion data set is acquired, and the emotion data set comprises emotion data of a plurality of digital people in different situations; a double random symmetric matrix is constructed based on the multiple pieces of emotion data in the emotion data set, and the sum of matrix elements of each row and each column in the double random symmetric matrix is 1; clustering the plurality of sentiment data in the double random symmetric matrix by using a target clustering algorithm to obtain a plurality of clusters; and determining the emotion type of each cluster, and taking the emotion type of each cluster as the emotion type of each digital person in the cluster in different situations. According to the method and the device, the technical problem that the accuracy of classifying the complex and high-dimensional sentiment data of the digital human is relatively low in related technologies is solved.

The invention discloses a data processing method and device, equipment, a storage medium and a program product, and belongs to the technical field of data science. The method comprises the following steps: normalizing an original data set to a preset numerical value interval to obtain a candidate data set; performing covariance matrix construction on the candidate data set to obtain a covariance matrix; constructing a transformation symmetric matrix based on the covariance matrix, and generating a transformation covariance matrix based on the covariance matrix and the transformation symmetric matrix; adding noise to the transformation covariance matrix by adopting a differential privacy algorithm to obtain a noise-added matrix, and carrying out inverse operation on the noise-added matrix to obtain a noise-added inverse matrix; and obtaining a target inverse covariance matrix according to the noise-added inverse matrix and the transformation symmetric matrix. According to the embodiment of the invention, the accuracy and stability of the inverse matrix of the covariance matrix can be improved.

The embodiment of the invention provides a linear regression method based on privacy protection. The method comprises the steps that a current participant receives an inverse decomposition factor matrix of a regularization symmetric matrix or a transformation form of the inverse decomposition factor matrix of the regularization symmetric matrix influenced by a column encrypted data transformation encryption matrix corresponding to private original data of a former participant; the current participant and the former participant cooperate to determine respective product matrixes or transformation forms thereof; the current participant sends an inverse decomposition factor matrix or a transformation form thereof to a subsequent participant; or, the current participant obtains and utilizes the product matrix of the front participant and the current participant or the transformation form of the product matrix to cooperate with the front participant and the label data owner to determine the multiplication encryption linear regression coefficient vector; multiplying the encrypted linear regression coefficient vector to be equal to the product of the inverse decomposition factor matrix containing the front participant and the current participant and the first product vector; according to the method provided by the invention, the implementation complexity can be reduced, and actual application requirements are met.

The invention discloses a system recommendation method based on Lanczos algorithm orthogonality. The system recommendation method comprises the following steps: firstly, modeling a problem of searching a community structure in a network structure into a problem of solving first k minimum feature pairs of a network Laplacian matrix; secondly, projecting a Laplacian matrix in a high-dimensional Euclidean space into a symmetric three-diagonal matrix in a low-dimensional Krylov subspace; then, using a feature value convergence criterion to screen out converged feature pairs of the symmetric tridiagonal matrix, and calculating feature vectors corresponding to the converged feature pairs; thirdly, calculating a feature vector of a Laplacian matrix according to the feature vector of the symmetric matrix; and finally, carrying out de-orthogonalization on the convergent feature vector of the Laplacian matrix and the Lancozs vector, storing the convergent feature vector of the Laplacian, and using the convergent feature vector of the Laplacian to detect the community structure through a standard k-means algorithm. According to the method, the real community detection data set is used as a drive, the Lanczos algorithm is used as a model, and the community structure contained in the graph network can be accurately detected, so that the recommendation quality of the system is improved.

The application provides a parallel computation method and system for multiplication of a symmetric matrix and a vector, and the method comprises the following steps: obtaining the symmetric matrix and the vector, and determining the number of threads; dividing a sub-region; in each sub-region, marking a first axis and / or a second axis parallel to the symmetric axis; calculating the symmetric axis region and the vector, and writing into a register; performing twice different first multiplication calculation on the first axis, obtaining two groups of different data and writing into the register respectively; performing twice different second multiplication calculation on the second axis, obtaining two groups or four groups of different data and writing into the register respectively, and obtaining the calculation result of the multiplication of the symmetric matrix and the vector. According to the scheme, the multiplication calculation of the symmetric matrix and the vector can be accurately and quickly realized, the order of element processing in the matrix is reasonably planned, resources of the threads are reasonably distributed, the write conflict can be effectively avoided, the multi-thread parallel calculation is realized, the calculation efficiency is high, the load of each thread is balanced, and the thread resource distribution is reasonable.

A computer-implemented method for determining a robust steering control of a steering system controller is disclosed, comprising: determining a solution to a system of statements comprising an equals sign, a greater-than sign, a greater-than or equal-to sign, a less-than sign and / or a less-than or equal-to sign, wherein at least a first part of the statements can be written equivalently or substantially equivalently as a first symmetric matrix greater than or equal to zero, wherein the first symmetric matrix is ​​based on system parameters of a system of one or more differential equations for modeling a steering of the steering system or their discretization, wherein the first symmetric matrix is ​​based on one or more perturbation parameters acting on the system of one or more differential equations; and calculating the steering control based on the solution.A further method for manufacturing and / or dataing a controller of a steering system is disclosed, comprising calculating the steering control; and adjusting the controller based on the steering control, in particular wherein the steering control is adjusted in the controller.

This invention discloses a method for extracting orthogonal radiative eigenmodes in lossy structures. For electromagnetic structures containing lossy materials, resistive loading, and other dissipation mechanisms, this invention obtains the scattering T-matrix of the structure under a preset incident mode basis. Utilizing the property that the T-matrix is ​​a complex symmetric matrix under reciprocity conditions, the Autonne-Takagi decomposition method is used to decompose this complex symmetric matrix, yielding a set of mutually orthogonal radiative eigenmodes in the sense of radiated power and their corresponding radiated intensity parameters. The obtained radiative eigenmodes can be used for mode decomposition and physical interpretation of the radiation and scattering characteristics of lossy structures, and are applicable to the design and optimization of various practical lossy electromagnetic structures such as resistive-loaded antennas, scatterers containing dielectric losses, and absorbing material structures.

The embodiment of the application relates to an algorithm and a device of a Bayesian classifier of a non-uniform background, which are applied to an underwater active sonar system, the algorithm comprising: obtaining underwater data to be measured and auxiliary data through the active sonar system; the number K of the auxiliary data is greater than 0; modeling classification of the unknown covariance matrix structure into a binary hypothesis testing problem; hypotheses of the binary hypothesis testing problem comprise H i Wherein i=0, 1, H0 is a case that the unknown covariance matrix is a complex conjugate symmetric matrix; H1 is a case that the unknown covariance matrix is a real symmetric matrix; a Bayesian model is set, the Bayesian model comprising a complex inverse Wishart random matrix and a real inverse Wishart random matrix; a classifier for distinguishing the two hypotheses is obtained by using a minimum error probability criterion.

The invention belongs to the field of GPU multi-card interconnection communication, and relates to a full-exchange GPU communication method and system. According to the method, the data stored in each GPU is averagely divided into N blocks, and each block of data corresponds to a specific data block needing to be sent to the target GPU; constructing a data symmetric matrix, and determining a total communication round according to the total number N of the GPUs; in each round of communication, two-way data transmission is carried out between the two GPUs based on a pairing rule of a symmetric matrix; and through multi-round iteration, complete exchange transmission of all data blocks is completed, so that each GPU obtains a complete data set from all GPUs. According to the method, a connection topology scheme of the GPUs does not need to be changed at all, the symmetric matrix of the GPUs only needs to be generated according to the number of the GPUs, and the condition that the GPUs communicate with other GPUs at the same time is reduced under the condition that the data transmission quantity is not increased, so that the congestion possibility of the GPUs is reduced, and the communication performance is improved.

A computing network comprising an orchestrator controller and multiple nodes is used to execute a distributed application across these nodes, each node including a node controller. The orchestrator controller determines whether a matrix is ​​symmetric, the matrix being divided into partial matrices, each partial matrix residing on one of the multiple nodes. The orchestrator controller causes each node controller to apply an antisymmetric function to the cells of its partial matrices and aggregates the results of the antisymmetric function application, thereby providing a sum of the partial matrices. The orchestrator controller determines whether the sum of all partial matrix sums is zero; if it is zero, the matrix is ​​determined to be symmetric. This computing network reduces the amount of communication between the multiple nodes required to determine the symmetry of the matrix.

The invention provides a heterogeneous symmetric matrix eigenvalue decomposition method, system and device based on GPU parallel acceleration, and belongs to the technical field of high-performance computing. According to the method, the design that five sub-processes can only be strictly and serially executed in the traditional symmetric matrix eigenvalue decomposition process is broken through, the problem that CPU and GPU computing resources are idle in a serial algorithm is solved, and the computing efficiency is greatly improved. Meanwhile, aiming at the BC-Backk stage, the invention provides a novel BC-Backk method realized based on a GPU (Graphics Processing Unit), the BC-Backk process is transferred from a CPU (Central Processing Unit) with limited parallel capability to a GPU capable of realizing large-scale parallel, and the parallel efficiency of the BC-Backk stage is greatly improved. According to the method, heterogeneous architecture characteristics and GPU parallelism are fully utilized, parallel between symmetric matrix eigenvalue solving sub-processes and internal parallel of BC-Back are achieved, CPU and GPU resources are fully utilized, and the calculation performance of the interior of a computer for the symmetric matrix eigenvalue solving process is greatly improved.

The invention provides a graph data processing method and device, equipment and a medium, relates to the technical field of artificial intelligence, can be applied to scenes such as cloud technology, artificial intelligence, intelligent traffic and auxiliary driving, and comprises the following steps: if a data structure of to-be-processed graph data is a directed graph, obtaining first matrix data and first graph attribute characteristics of the to-be-processed graph data; the first matrix data comprises an out-degree matrix and an in-degree matrix, elements of the out-degree matrix are used for representing an outward connection relationship between a corresponding node in the to-be-processed graph data and other nodes, and elements of the in-degree matrix are used for representing a connection relationship between the corresponding node in the to-be-processed graph data and the other nodes; the out-degree matrix and the in-degree matrix are symmetrical matrixes; performing data analysis of the first matrix data and the first graph attribute features based on the target graph neural network to obtain a graph analysis result; by means of the conversion method, direction information of the directed graph data can be reserved, and meanwhile complex conversion is not introduced.

The embodiment of the invention discloses a matrix operation method and device, computer equipment and a storage medium, and belongs to the technical field of computers.The method comprises the steps that block processing is conducted on a transformation matrix to obtain a plurality of transformation blocks, and block processing is conducted on a triangular matrix to obtain a plurality of initial full-element blocks and a plurality of initial triangular blocks; obtaining a transposed full-element block corresponding to each initial full-element block, and obtaining a restored full-element block corresponding to each initial triangular block; according to a preset matrix block position index, determining restored full-element blocks and initial full-element blocks with block positions having an index row-column symmetric relationship, and transposing the full-element blocks into the same target processing group; and performing matrix operation on each transformation block and the processing block at the corresponding block position in the corresponding target processing group, and determining a target matrix operation result according to a matrix block operation result. According to the invention, the matrix operation efficiency, especially the matrix operation efficiency for the symmetric matrix, can be improved.

The invention discloses a privacy protection approximate average consensus method of a networked system. According to the method, a first mask and a second mask are randomly generated, and the requirements that the difference between the first mask and a common constant is absolutely summable, the second mask is absolutely summable, and the common constant is not zero are met; the product of the first mask and the current state is calculated, the product and the second mask are added to obtain the outward sending amount, and the outward sending amount is sent to the neighbor node; weighted summation is carried out on the received outward sending amount and the outward sending amount sent by the network system according to a preset weight, the preset weight of the weighted summation is larger than 0, the sum is 1, the weighted summation result is divided by the public constant, the updating amount is obtained, and a weight matrix of the network system is a symmetric matrix; weighted summation is carried out on the update quantity and the initial state value, an updated state value is obtained, the weight of weighted summation is a fixed value and is larger than 0, and the sum is 1. The method can realize approximate average consensus, and is high in privacy protection capability and low in operand.

A random access method based on binary subspace chirp codes and generalized approximate message passing on a fast fading channel is applicable to fast fading channel scenarios. The method comprises: firstly, setting the symmetric matrix and vector constituting the binary chirp part to 0 to obtain a new codeword construction formula; then, based on the principle that the subspace has no intersection except the zero vector, selecting matrices constituting the subspace; substituting the matrices constituting the subspace into the new codeword construction formula to obtain a new codebook; then, allocating the new codebook to users as a transmitted signal; then, transmitting the received signal through the fast fading channel; and executing a generalized approximate message passing algorithm to estimate active users; iterating at the receiving end; and finally, judging the termination of iteration based on the active user set and the number of iterations to obtain an estimated active user set. The present invention reduces computational complexity while also reducing the missed detection rate.

Techniques for securing a digital ecosystem are disclosed. In embodiments, a method includes storing a master QNA object comprising s symmetric matrices, each having d rows. The master QNA is structured to allocate correlated QNA objects to cohorts of the digital ecosystem based on credentials of the cohorts. The method includes receiving a unique identifier of a cohort being admitted to the ecosystem; determining a set of s selection values based on the unique identifier of the new digital cohort and a selection function; and allocating a new QNA object comprising s vectors to the cohort based on the s selection values and the s matrices. Each selection value is between 1 and d and corresponds to a respective composite matrix of the s composite matrices such that each selection value indicates a specific row of the respective matrix to which the selection value corresponds.

The invention provides a construction method of an anisotropic rock mass equivalent degradation mechanical model, which comprises the following steps: S1, determining a mechanical model: based on a generalized Hooke's law, the basic expression form of the rock mass mechanical model can use an incremental expression of a stress-strain relationship; s2, determining an elastic stiffness matrix: constructing an elastic stiffness matrix # imgabs0 # by adopting a symmetric matrix method in an inner product form; and S3, determining a plastic strain increment: determining a yield function # imgabs1 #, obtaining a dilatation angle # imgabs2 #, replacing an internal friction angle # imgabs5 # in the yield function # imgabs4 # with the dilatation angle # imgabs3 # to obtain a plastic potential function # imgabs6 #, and solving an increment # imgabs8 # of a plastic strain tensor with the plastic potential function # imgabs7 #. On the basis of the equivalent elastic-plastic theory, the anisotropic rock mass equivalent degradation mechanical model considering the synergistic effect of the intermediate principal stress and the structural plane occurrence is established, so that the model can reflect the brittle-ductile degree, the capacity expansion characteristic and the failure mechanism of a material and the influence mechanism of the intermediate principal stress and the structural plane.

The application provides a QC-LDPC code construction method and device based on symmetric extension and electronic equipment, relates to the technical field of communication channel coding and decoding, and comprises the following steps: determining a target base matrix of a QC-LDPC code based on a random code method, wherein the target base matrix is full rank and does not have a short ring with a length of 4; generating a longitudinal symmetry matrix according to the target base matrix, a diagonal line offset matrix, an extension factor N and an integer k coprime with the extension factor N; synthesizing an extended base matrix according to the target base matrix, the longitudinal symmetry matrix, a modular inverse matrix of the target base matrix and a modular inverse matrix of the longitudinal symmetry matrix; performing extension on the extended base matrix to obtain a quasi-cyclic check matrix of the QC-LDPC code, wherein the quasi-cyclic check matrix has an extension dimension N relative to the extended base matrix; and generating the QC-LDPC code based on the quasi-cyclic check matrix. The construction method has the advantages of low complexity and flexible adaptation to various code rates.

The invention relates to a protein binding conformation prediction method and system. The protein binding conformation-based prediction method comprises the following steps: acquiring a static structure file of a target protein, and extracting atomic information to construct a residue index table; performing standardization processing on the atomic coordinates in the residue index table, and calculating a distance matrix between the residue pairs based on the standardized atomic coordinates; based on the conformation sampling data and the residue index table, obtaining a residue motion freedom degree through adaptive weighting calculation and correlation correction; a dynamic distance threshold value is obtained based on the residue motion freedom degree; converting the distance matrix into a binary contact graph based on the dynamic distance threshold, obtaining the contact frequency of the residue pair, and constructing a contact frequency matrix; extending the residue motion degree of freedom into a symmetric matrix, and fusing the contact frequency matrix and the symmetric matrix to obtain an initial dynamic probability matrix; performing normalization processing and probability propagation on the initial probability matrix in sequence to obtain a dynamic probability matrix; protein binding conformations are predicted based on the dynamic probability matrix.

The invention relates to the technical field of intersecting line track calibration, in particular to a circular pipe intersecting line welding line coordinate system calibration method which comprises the steps that an intersecting line track mathematical model is established, and the coordinate relation of intersecting lines in a main pipe coordinate system and a branch pipe coordinate system is determined; based on the transformation relation between a robot base coordinate system and a master coordinate system, an anti-symmetric matrix is adopted to express an attitude matrix, and a calibration constraint equation for eliminating a translation vector is established; and randomly selecting three teaching points on the intersecting line track, substituting the coordinates of the three teaching points in the two coordinate systems into the constraint equation, solving to obtain an attitude matrix and a translation vector, completing the solving of a conversion matrix, and realizing the calibration of the main coordinate system. According to the method, calibration can be completed only through any three teaching points, auxiliary equipment is not needed, the method is adaptive to all types of intersecting line models, the method has the advantages of being easy and convenient to operate, high in efficiency and precision and wide in application range, and an accurate coordinate system conversion basis can be provided for track planning and interpolation calculation of scenes such as robot intersecting line welding and cutting.

The invention discloses a method for quickly estimating the upper and lower bounds of the maximum magnetic force of a micro-nano robot in a double-synchronous rotating magnetic field. The method comprises the following steps: acquiring position vectors, common magnetic moment vectors, common rotating shaft unit vectors and to-be-estimated position points of two permanent magnets; calculating and superposing a total magnetic field generated by each permanent magnet at the to-be-estimated position point and a magnetic field rotating shaft; the magnetic gradient force borne by the micro-nano robot is expressed as a calculation formula related to the magnetic moment direction of the permanent magnet and the magnetic moment direction of the micro-nano robot; constructing a symmetric matrix related to a common rotating shaft of the permanent magnets and a magnetic field rotating shaft at the position point to be estimated, and taking a characteristic value of the symmetric matrix as a candidate upper bound of the normalized maximum magnetic force; constructing an auxiliary matrix based on the eigenvectors, and taking eigenvalues of the auxiliary matrix as a candidate lower bound of the normalized maximum magnetic force; and obtaining an estimation interval of the normalized magnetic gradient force maximum square according to the candidate upper and lower bounds, and outputting a modulus length estimation range. According to the method, rapid estimation of the maximum magnetic force is realized on the premise of ensuring the calculation efficiency.

Techniques for securing a digital ecosystem are disclosed. In embodiments, a method includes storing a master QNA object comprising s symmetric matrices, each having d rows. The master QNA is structured to allocate correlated QNA objects to cohorts of the digital ecosystem based on credentials of the cohorts. The method includes receiving a unique identifier of a cohort being admitted to the ecosystem; determining a set of s selection values based on the unique identifier of the new digital cohort and a selection function; and allocating a new QNA object comprising s vectors to the cohort based on the s selection values and the s matrices. Each selection value is between 1 and d and corresponds to a respective composite matrix of the s composite matrices such that each selection value indicates a specific row of the respective matrix to which the selection value corresponds.

This invention discloses a sensor-computer integrated visual positioning method and system for robot target localization. Image data is acquired by a sensor board equipped with sensors and sent to a core processor. Image processing and target edge extraction are performed at the programmable logic unit (PL), and the data is transmitted to the processing system unit (PS). The PS reads edge points, performs arc segment extraction to obtain elliptical arc segments, and writes them to the PL. Based on this, the PL combines this with a symmetric matrix to optimize the pulsation array and achieve ellipse fitting to complete target localization. An interface board microcontroller controls the power supply timing and manages external interfaces. The core processor (PS) and peripheral interfaces are responsible for overall system control and Ethernet communication. Integrating image sensing and edge computing into a single hardware component, through collaborative processing, hardware acceleration of key algorithms, optimized data transmission, and integrated management, this method achieves low latency, high accuracy, and high robustness in robot target localization, and possesses advantages such as compact structure, low power consumption, high adaptability, and rapid deployment.

Embodiments of the present application provide a linear regression method based on privacy protection. The method comprises: a current participant receives an inverse decomposition factor matrix of a regularized symmetric matrix, or a transformed form thereof, wherein the regularized symmetric matrix comprises, as a submatrix, a column-encryption data transformation encryption matrix corresponding to private raw data of a preceding participant; the current participant cooperates with the preceding participant to determine respective product matrixes or transformed forms thereof of the current participant and the preceding participant; the current participant sends the inverse decomposition factor matrix or the transformed form thereof to a subsequent participant; or, the current participant obtains and utilizes the respective product matrixes or the transformed forms thereof of the preceding participant and the current participant, and cooperates with the preceding participant and a label data owner to determine a multiplicatively encrypted linear regression coefficient vector, wherein the multiplicatively encrypted linear regression coefficient vector is equal to the product of the respective inverse decomposition factor matrix of the preceding participant and the current participant and a first product vector. The method provided in the present application can reduce implementation complexity, satisfying practical application requirements.

This application provides a linear regression method for horizontally partitioning data and based on privacy protection, comprising: each of at least two data owners using their own private data samples to obtain ciphertext related to their original symmetric matrix and ciphertext of their original product vector, wherein the original symmetric matrix and the original product vector are respectively the symmetric matrix and product vector corresponding to several data samples; the data owners send the ciphertext related to the original symmetric matrix and the ciphertext of the original product vector to a first server; the first server receives the ciphertext related to the original symmetric matrix and the ciphertext of the original product vector sent by each of the at least two data owners, and cooperates with a second server to obtain the target linear regression coefficients; according to the method proposed in this application, apart from the data owners holding their own private data, neither the first server nor the second server can steal the private data based on their own conditions, thereby ensuring the security of the private data.

The invention discloses a dynamic updating method and system for urban and rural planning data based on big data, and relates to the technical field of smart cities, through standardized processing and sequence table construction of multi-source heterogeneous urban and rural data in the step S1, the consistency of a data structure and the precision of time stamps are realized, and the dynamic updating efficiency of urban and rural planning data is improved. The problems of non-uniform data formats, timestamp missing and the like in a traditional system are solved. And S2, introducing construction of a difference symmetry matrix E and a fusion index Xts, and quantitatively identifying time update differences among various types of data, thereby effectively judging the data synchronism, providing a basic judgment basis for subsequent update or not, and improving the data perception capability and misjudgment avoidance capability of the system. And S3, judging whether a self-adaptive weight mechanism needs to be started or not through the fusion index Xts, and dynamically adjusting the participation degree of lagged data and structural stable data in combination with a Sigmoid adjustment function and an attention weighting strategy, thereby realizing a weighted fusion process driven by data credibility.

The present invention provides a method, system and device for eigenvalue decomposition of heterogeneous symmetric matrices based on GPU parallel acceleration, which belongs to the field of high-performance computing technology. The present invention breaks the design that the five sub-processes in the traditional symmetric matrix eigenvalue decomposition process can only be strictly executed in serial, solves the problem of idle CPU and GPU computing resources in serial algorithms, and greatly improves computing efficiency. At the same time, the present invention proposes a novel BC-Back method based on GPU implementation for the BC-Back stage, which transfers the BC-Back process from the CPU with limited parallel capability to the GPU that can be large-scale parallel, greatly improving the parallel efficiency of the BC-Back stage. The present invention makes full use of the characteristics of heterogeneous architecture and GPU parallelism, realizes the parallelism between the eigenvalue solution sub-processes of the symmetric matrix and the internal parallelism of BC-Back, makes full use of CPU and GPU resources, and greatly improves the computing performance of the computer for the eigenvalue solution process of the symmetric matrix.

The present application relates to a kind of frequency response calculation method, device and electronic equipment, it is related to engineering technical field, it is intended to reduce the amount of calculation, improve the calculation efficiency of frequency response, comprising: in the process of frequency scanning to the detected structure, for each frequency point scanned, based on frequency point, modal solution parameter, degree of freedom information and the physical attribute of detected structure, construct the frequency response equation with damping, wherein, frequency point is used to reflect the excitation frequency applied to the detected structure, degree of freedom information is used to indicate at least one response dimension in frequency response analysis Concerned with;The frequency response equation with damping is reduced order, and first frequency response equation is obtained;In the case that there is damping in the physical attribute of detected structure, based on complex symmetric matrix, first frequency response equation is twice reduced order, and second frequency response equation is obtained;Based on second frequency response equation, the frequency response result corresponding to frequency point is obtained.

The invention relates to the technical field of electromagnetic reconnaissance, and particularly discloses a Hermitian matrix parallel characteristic decomposition method for an electromagnetic reconnaissance system, and the method comprises the steps: constructing a 2N-order real symmetric matrix S based on a real part and an imaginary part of an N-order Hermitian matrix of the electromagnetic reconnaissance system; performing M times of iterative computation on the 2N-order real symmetric matrix S and the feature vector matrix R based on the scheduling rule matrix T to obtain a corresponding real symmetric matrix S 'and a feature vector matrix R'; and finally, on the basis of elements of the matrix S'and the matrix R ', obtaining eigenvalues of the N-order Hermitian matrix and eigenvectors corresponding to the eigenvalues. According to the method, the complex matrix is converted into the real matrix for processing, the scheduling rule matrix T is designed for iterative calculation, and the method has high decomposition precision. The same processing method can be adopted to perform characteristic decomposition on the Hermitian matrix of odd or even dimensions, and the method can be suitable for any processing system supporting parallel computing and is wide in application range.

The invention discloses a channel map assisted large-scale MIMO low-orbit satellite communication pilot frequency scheduling method and system, and the method comprises the steps: constructing a Fisher information matrix based on a channel parameter vector based on a receiving signal model of a large-scale MIMO low-orbit satellite communication system under a pilot frequency multiplexing condition; spatial features are extracted based on the Fisher information matrix, a feature matrix is constructed, the feature matrix is a symmetric matrix and used for quantifying parameter estimation correlation between different user terminals, and element values in the feature matrix are determined according to an autocorrelation matrix and a cross-correlation matrix in the Fisher information matrix; performing dimension reduction on the feature matrix to generate a channel map; and based on the channel map, pilot frequency scheduling is carried out by using relative spatial positions among users in the channel map so as to suppress pilot frequency interference caused by pilot frequency multiplexing. According to the method, the channel estimation precision and the spectrum efficiency of the low-orbit satellite large-scale MIMO system in the pilot frequency multiplexing scene can be effectively improved.