35 results about "Symmetric matrix" patented technology

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 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.

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 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.

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 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 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.

The invention discloses a regression model establishment method suitable for a symmetric positive definite matrix data set, and belongs to the technical field of machine learning and data science. The method comprises the following steps: firstly, converting an SPD matrix in a training set into a symmetric matrix by adopting a spectral analysis method, and reserving geometric characteristics of the matrix; secondly, performing symmetric matrix regression relation modeling by applying a kernel ridge regression method based on a Log-Euclidean kernel function, and effectively capturing intrinsic statistical characteristics of the matrix; and finally, a regression prediction result is converted into an SPD space through exponential operation of the matrix, so that the physical significance and the practical application value of the result are ensured. The series of technical means not only overcome the limitation of the traditional method, but also significantly improve the prediction precision and robustness of the model.

A beamformee receives a channel coefficient matrix H of a wireless communication channel between a beamformer and the beamformee. The channel coefficient matrix is a 2×N complex matrix having two rows corresponding to two antenna of the beamformee and N columns corresponding to N antenna of the beamformer. A real symmetric matrix M is determined based on the matrix H followed by determining eigenvalues and eigenvectors of matrix M. Singular vectors of the matrix H based on the eigenvectors are determined where the singular vectors define a steering matrix. The steering matrix is transmitted to a beamformer, wherein a beam is steered by the beamformer to the beamformee based on the steering matrix.

The application discloses a remote sensing small sample target detection method for feature library migration. Firstly, multi-scale feature extraction is performed on the input remote sensing image to be detected based on a deep convolutional neural network to generate a detection model of common categories; a small sample feature library is set, and the small sample feature library is added to the detection model for training, so that the detection model can maintain the detection ability for common targets while detecting small samples; in view of the high symmetry of the remote sensing image to be detected, a symmetric matrix parameter coefficient is introduced when the detection result is predicted after the training of the detection model is completed, the positioning branch parameter is corrected, and the small sample target detection precision is improved. The above method solves the problem that the modeling of the target category features is poor in the remote sensing image, and the target cannot be accurately recognized, and improves the effectiveness of the remote sensing image detection in actual application.

The present application relates to the technical field of data processing and artificial intelligence algorithm, and particularly relates to a multi-agent collaborative data accurate query method and system and a medium, the method comprising: given a user query and a group of agents, constructing a knowledge graph based on the results returned by the agents to the user query; evaluating the matching degree of the agents and the user query by comprehensively considering the user query and the knowledge graph; generating the difference characteristics of the corresponding agents by analyzing the feedback difference of the knowledge graph; constructing a conflict index matrix by combining the matching degree and the difference characteristics, obtaining a symmetric matrix after correction, and establishing a conflict sequence; judging the credibility of the results of the agents according to the conflict sequence, determining the termination condition of the iteration of the corresponding agents, and outputting the current results of the agents when the termination condition is reached; and obtaining the query results by fusing the current results of all the agents and returning them to the user, so as to improve the accuracy of the query results.

The invention discloses a monocular vision-based vehicle grounding point completion method, device and equipment. The method comprises the following steps: obtaining homogeneous coordinates of three visible vehicle grounding points of at least five vehicles shot by a monocular camera in an image coordinate system; determining a linear equation of two connecting straight lines of the three visible vehicle grounding points on the image plane; according to the conjugate relation of the two straight lines of each vehicle with respect to the homogeneous symmetric matrix, constructing a linear equation of a linear equation of the two connecting straight lines of the vehicle in the image plane, and solving to obtain the homogeneous symmetric matrix; obtaining homogeneous coordinates of three visible vehicle grounding points of the target vehicle in the image coordinate system; determining a linear equation of four connecting straight lines of the four vehicle grounding points of the target vehicle on the image plane; according to the conjugate relation of the four connecting straight lines about the homogeneous symmetric matrix, a linear equation set is constructed, and homogeneous coordinates of the fourth to-be-complemented vehicle grounding point in the image coordinate system are obtained through solving.

The invention discloses a radar-oriented order constraint Jacobi eigenvalue self-sorting method and system, solves the problem of resource consumption caused by an additional sorting module in a Jacobi iteration eigenvalue decomposition post-processing flow in the prior art, realizes direct output of ordered eigenvalues and eigenvectors, and omits an independent sorting module. The method comprises the steps that a real symmetric matrix A obtained through radar baseband signal sampling and covariance estimation is processed to obtain a compressed array, and a feature vector matrix V is initialized; in the parallel iteration process, index pairs are generated according to a scheduling sequence, sub-matrixes are extracted, two types of asymmetric rotation operators are generated through a condition driving mechanism, and a feature vector matrix is synchronously updated to track feature vectors; global data replacement is carried out, and non-diagonal element energy judgment convergence is calculated; after iteration is ended, feature values arranged in a descending order can be directly extracted from the main diagonal of the final compressed array, and corresponding feature vectors are output and used for direction of arrival estimation.

The embodiment of the application provides a linear regression method based on privacy protection, which comprises the following steps: an ith participant determines a regularized symmetric matrix of a column-encrypted data transformation encrypted matrix corresponding to an original data matrix of the participant or a transformed form of the regularized symmetric matrix, and sends the regularized symmetric matrix or the transformed form to a first computing node; a jth participant cooperates with the ith participant to determine a product matrix of column-encrypted data transformation encrypted matrices of the jth participant and the ith participant or a transformed form of the product matrix, and the first computing node saves the product matrix or the transformed form; i and j are integers greater than or equal to 1 and less than or equal to m, j is not equal to i and j is less than i, and m is the total number of all participants; the first computing node cooperates with each participant and a label data owner to determine a multiplication encrypted linear regression coefficient vector by using the regularized symmetric matrix and the product matrix; the method provided by the application is simple, and encryption does not need to use complex mixing circuit technology.

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 application relates to a privacy protection-based linear regression method, which comprises the following steps: a first participant constructs a column-encrypted data matrix of an original data matrix formed by private original data as a submatrix according to the original data matrix, and after a decomposition factor matrix of an inverse matrix of a regular symmetric matrix of the column-encrypted data matrix is calculated, an encrypted influence matrix corresponding to the decomposition factor matrix, which comprises a column-encrypted data and factor product matrix and a multiplication-encryption factor matrix, is sent to a label data owner or forwarded to the label data owner by other participants according to a preset encryption transmission sequence; and the label data owner calculates a multiplication-encryption linear regression coefficient vector according to a private label column vector and the received encrypted influence matrix. Compared with the prior art, the application does not need to introduce a trusted initializer and an encryption service provider, and does not need to realize a very high complexity confusion circuit, so that the implementation complexity can be reduced, and the actual application requirements are met.

The invention discloses a target trace point association method based on a graph neural network, and relates to the field of aviation multi-sensor data fusion. The method comprises the following steps of: performing similarity evaluation on two point traces by using parameters such as sensor detection errors and the like in a similarity characterization method; according to the graph data structure construction method, similarity evaluation is carried out on trace points of multiple data sources and multiple targets, and then a graph adjacency matrix is formed. According to the graph net spirit meridian model training method, a target association graph network is constructed based on an enhanced graph attention network model, and the network model is trained by using historical data. According to the graph neural network model reasoning method, the trained graph network model is used for accessing test data for reasoning. A similarity matrix reconstruction method converts a reasoning result into a symmetric matrix. The spectral clustering method is used for carrying out spectral clustering on the reconstructed similarity matrix, so that the problem of incomplete association of multiple points is solved. Compared with a traditional clustering algorithm, the accuracy of the method is improved by 7%, and data input with higher precision is provided for multi-sensor data fusion.

The invention discloses a time sequence delay modeling and causality interpretable method for industrial soft measurement, and provides a space-time delay correlation causality discovery (STDCD) model for the modeling problem of time delay and causality between variables in an industrial process. Carrying out space-time delay correlation causal discovery and fusion modeling, introducing a space-time delay attention mechanism, and constructing a two-dimensional causal graph containing a time delay attribute; dynamically discovering a causal relationship, calculating attention path intensity of time feature coupling, establishing a symmetric matrix and performing threshold determination to obtain an undirected graph, and determining a causal direction through a multi-dimensional dynamic complexity index; and designing an STLD-Net network, extracting high-order feature interaction information by using a graph convolution layer, and dynamically adjusting the attention degree of the model on historical data in combination with a delay sensing mechanism. Experimental results show that the hidden space-time causal relationship in the industrial process is effectively captured, the prediction precision and the multi-dimensional causal interpretability are remarkably improved, and the method is suitable for soft measurement modeling of a complex industrial system.

The invention discloses a circuit breaker phase selection control method and device based on multi-parameter dynamic compensation and adaptive learning, a medium and equipment, and relates to the technical field of substation phase selection opening and closing control. The method comprises the following steps: fusing multi-parameter data acquired by a plurality of pre-installed sensors, and constructing a symmetric matrix of the multi-parameter data; calculating a nonlinear interaction item according to the symmetric matrix to obtain a normalized compensation amount; and compensating nominal opening and closing time of the circuit breaker according to the normalized compensation quantity to obtain compensation opening and closing time of phase selection of the circuit breaker. And phase selection control of multi-parameter coupling compensation and dynamic self-learning is realized.