An accurate decomposition equation construction and decomposition method for finite signals

Through the construction of generalized eigenvalue equations and matrix calculation, the problem of large phase spectrum error in complex signal processing is solved, and the precise decomposition of complex signals is achieved, which meets the high-precision requirements in engineering applications.

CN113099728BActive Publication Date: 2024-11-26刘保国 +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201980004882.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2019-11-08
Publication Date
2024-11-26
Estimated Expiration
2039-11-08

AI Technical Summary

Technical Problem

Existing signal processing technologies are difficult to accurately decompose complex signals, especially the phase spectrum of signals cannot be accurately obtained, resulting in insufficient accuracy in engineering applications, which cannot meet the frequency and amplitude of fields such as speech processing, voiceprint recognition, biomedicine and other fields. and phase information requirements.

Method used

Using generalized eigenvalue equations, by constructing the combination of matrix A and matrix B, the frequency, amplitude and initial phase of the simple harmonic component of complex signals is provided, and a method of accurate decomposition of finite signals is obtained by using differential and integral operations. , construct an equation about the frequency of simple harmonic components, calculate generalized eigenvalues ​​and eigenvectors, and solve all or part of the components of the signal.

Benefits of technology

The precise decomposition of finite length complex signals is achieved, and the frequency, amplitude and initial phase of all or part of the simple harmonic components of the signal is obtained, which improves the accuracy of signal processing and meets the needs of engineering applications, such as in speech processing and sound. The accuracy and reliability are greatly improved in pattern recognition.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113099728B_ABST
    Figure CN113099728B_ABST
Patent Text Reader

Abstract

The invention relates to a method for constructing and decomposing an accurate decomposition equation of a finite signal, comprising the following steps: step S101, constructing a discrete source signal matrix according to a source signal of a complex signal, and obtaining a differential / integral / differential signal sequence of the discrete source signal from the source signal of the complex signal, thereby constructing a differential / integral / differential signal matrix of the discrete source signal, and using the discrete source signal matrix and / or the differential / integral / differential signal matrix of the discrete source signal to construct a generalized eigenvalue equation about the frequency of a simple harmonic component of the complex signal; step S102, calculating all generalized eigenvalues ​​and generalized eigenvectors according to the generalized eigenvalue equation; or / and calculating part of the generalized eigenvalues ​​of the generalized eigenvalue equation and their corresponding generalized eigenvectors; step S103, solving the frequency, amplitude and initial phase of all simple harmonic components of the complex signal according to all the generalized eigenvalues; and / or solving the frequency, amplitude and initial phase of part of the simple harmonic components of the complex signal according to part of the generalized eigenvalues ​​and their corresponding generalized eigenvectors, so as to solve the problem of accurate decomposition of finite-length and finite-frequency component signals, and having a wide range of engineering demand backgrounds and application values.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to a method for constructing an equation for accurate decomposition of a finite signal and a method for using the equation, and relates to the field of signal processing technology. The invention can be widely used in signal analysis and processing in various fields such as communication, acoustics, imaging, mechanics, electronics, medicine, biology, etc. Background Art

[0002] Signal analysis and processing technology plays an important role in various fields of science and technology, communications, national defense and people's daily life. Especially with the rapid development of 5G communications and modern high-tech, various signal processing technologies are emerging. However, due to the constraints of basic theoretical research, various signal processing technologies cannot accurately decompose complex signals. For example, traditional FFT (fast Fourier transform) technology is subject to the leakage caused by signal truncation and the fence effect caused by digital discretization. The amplitude spectrum and phase spectrum must have errors, especially the phase spectrum. The error is greater, resulting in very little application of the phase spectrum in engineering practice; wavelet analysis technology is a signal processing technology that has been widely studied and applied in recent years, but wavelet analysis technology is heavily dependent on the selectivity of wavelet functions. Choosing different wavelet functions often results in very different analysis results.

[0003] In actual engineering, not only the amplitude spectrum of the signal is important, but the phase spectrum is also very useful. For example: in speech signal processing, the sound of orchestral instruments is pleasant to the ear. In addition to the wide frequency domain, the synchronization of multiple frequency components, that is, the same initial phase, is also a very important factor. However, it is difficult to accurately obtain the initial phase of various frequencies of musical instruments using traditional methods and wavelet analysis; in voiceprint recognition technology, if the phase factor is introduced, the accuracy and reliability of voiceprint recognition can be greatly improved; in the biomedical field, the synchronization of electrocardiogram, heart rate, and pulse signals is also a very important basis for disease diagnosis, but because it is difficult to obtain accurate phase information using traditional methods, doctors often need to use time domain waveforms combined with experience to make a diagnosis. In addition, in communication signals, military radars, ultrasonic and laser measurements, and image processing technologies, more accurate signal frequency, amplitude, and phase information are required.

[0004] On the other hand, signals of infinite length and infinite frequency components are mostly physically unrealizable signals, that is, most signals in actual engineering have finite length and are composed of finite frequency components. Therefore, solving the problem of accurate decomposition of signals of finite length and finite frequency components has a broad engineering demand background and application value. Summary of the invention

[0005] In order to overcome the shortcomings of the above-mentioned prior art, the present invention is aimed at a complex signal of finite length and composed of a finite number of simple harmonic components, discloses a method and solution for accurately decomposing a complex signal into simple harmonic components, provides a method for constructing an equation for accurately decomposing a finite signal, and a method for applying the equation to accurately decompose and obtain the frequency, amplitude and initial phase of all or part of the simple harmonic components constituting the complex signal.

[0006] To achieve the above object, the present invention adopts the following technical solution: a method for constructing an equation for accurate decomposition of a finite signal, wherein the finite signal refers to a complex signal of finite length and composed of a finite number of simple harmonic components, and the equation is a generalized eigenvalue equation for accurate decomposition of the complex signal, characterized in that the generalized eigenvalue equation is represented by a matrix as follows:

[0007] Bv=-λAv or λ -1 Bv=-Av

[0008] Among them, matrix A is a discrete source signal matrix or a differential / integral / differential signal matrix of a discrete source signal; matrix B is a second-order differential or quadratic differential signal matrix of matrix A and is arranged in the same way. The difference between the differential and the differential is that when performing a differential operation, the rate of change of the increment of the discrete function to the discrete variable must be taken into account, while the differential operation only takes into account the increment of the discrete function; λ is the generalized eigenvalue of the generalized eigenvalue equation; v is the generalized eigenvector of the generalized eigenvalue equation.

[0009] Furthermore, if the matrix A is a discrete source signal matrix, it can be expressed as:

[0010]

[0011] Where: x1,x2,x3,…,x 2n is the 2n discrete values ​​of the complex signal, i.e., the discrete source signal, then the matrix B is a second-order differential or second-order differential signal matrix composed of a discrete sequence of second-order differentials or second-order differentials of the discrete source signal, and the elements of the matrix B have the same arrangement as the elements of the matrix A;

[0012] The matrix A and the matrix B in the generalized eigenvalue equation are selected from one or more of the following matrices: a discrete source signal matrix of the complex signal, an n-order differential signal matrix of the discrete source signal of the complex signal, an n-time integral signal matrix of the discrete source signal of the complex signal, and an n-time differential signal matrix of the discrete source signal of the complex signal, wherein n is a natural number;

[0013] The difference between the nth differential and the nth order differential is that when performing the nth order differential operation, the nth order rate of change of the increment of the discrete function to the discrete variable must be taken into account, while the nth differential operation only takes into account the nth increment of the discrete function;

[0014] There are two principles for combining the matrix A and the matrix B in the generalized eigenvalue equation: the first is that the matrix B is the second-order differential of the matrix A; the second is that the matrix B is the quadratic differential of the matrix A.

[0015] Preferably, the discrete source signal of the complex signal and the n-order differential, n-order integral and n-order differential of the discrete source signal of the complex signal can be realized by hardware of a calculus circuit;

[0016] The source signal of the complex signal is divided into two parallel signal paths for synchronous output, the first path including the steps of performing A / D conversion on the source signal of the complex signal and discretizing it to obtain a discrete source signal of the complex signal;

[0017] The second path includes the following steps: performing differentiation / integration operations on the source signal of the complex signal through a differentiation / integration hardware circuit, performing A / D conversion on the differentiation / integration signal of the complex signal and discretizing it to obtain a differentiation / integration / differential signal sequence of the discrete source signal.

[0018] Preferably, the n-th order differential, n-th integration and n-th differential of the complex signal discrete source signal can be obtained by numerically calculating the n-th order differential, n-th integration and n-th differential of the discrete source signal sequence.

[0019] The present invention also provides a method for accurately decomposing a finite signal using a generalized eigenvalue equation of the frequency of a simple harmonic component of a complex signal, wherein the finite signal refers to a complex signal of finite length and composed of a finite number of simple harmonic components, comprising the following steps:

[0020] Step S101, constructing a discrete source signal matrix according to the source signal of the complex signal, and obtaining a differential / integral / differential signal sequence of the discrete source signal from the source signal of the complex signal, thereby constructing a differential / integral / differential signal matrix of the discrete source signal, and using the discrete source signal matrix and / or the differential / integral / differential signal matrix of the discrete source signal to construct a generalized eigenvalue equation about the frequency of the simple harmonic component of the complex signal;

[0021] Step S102, calculating all generalized eigenvalues ​​and generalized eigenvectors according to the generalized eigenvalue equation of the simple harmonic component frequency of the complex signal;

[0022] and / or calculating partial generalized eigenvalues ​​of the generalized eigenvalue equation and their corresponding generalized eigenvectors;

[0023] Step S103, calculating the frequency, amplitude and initial phase of all simple harmonic components of the complex signal according to all generalized eigenvalues;

[0024] And / or the frequency, amplitude and initial phase of some simple harmonic components of the complex signal are solved according to some generalized eigenvalues ​​and their corresponding generalized eigenvectors.

[0025] Preferably, in step S102, the numerical relationship between the generalized eigenvalue of the simple harmonic component frequency of the complex signal and the circular frequency of the simple harmonic component of the complex signal is divided into two cases. In the first case, matrix B is the second-order differential signal matrix of matrix A. In this case, the circular frequency of the simple harmonic component of the complex signal is:

[0026]

[0027] Where: i is the i-th generalized eigenvalue of the generalized eigenvalue equation;

[0028] The second case: Matrix B is the quadratic differential signal matrix of Matrix A. In this case, the circular frequency of the simple harmonic component of the complex signal is:

[0029]

[0030] Where: i is the i-th generalized eigenvalue of the generalized eigenvalue equation, and Δt is the discrete interval of the discrete source signal.

[0031] Further, in step S103, there are two paths to calculate the frequency, amplitude and initial phase of the simple harmonic component.

[0032] Path 1: First, the circular frequencies of all simple harmonic components of the complex signal are calculated according to all the generalized eigenvalues ​​of the generalized eigenvalue equation; then a linear equation system Dy=z is constructed, where: D=C T C, z = C T x, the vector x is composed of the discrete source signals of the complex signal, and the matrix C is composed as follows:

[0033]

[0034] Where: 2m-1≤k≤2n-1, m is the number of independent generalized eigenvalues ​​calculated, C T is the transposed matrix of matrix C; solve the linear equations to obtain vector y; finally, the amplitude and phase of each component are calculated by the following formula:

[0035]

[0036] Where: y 2i-1 and 2i are the 2i-1th and 2ith elements of the vector respectively;

[0037] Path 2: First, the circular frequencies of some simple harmonic components of the complex signal are calculated according to some generalized eigenvalues ​​of the generalized eigenvalue equation; then, the following two-variable linear equation system is constructed using some generalized eigenvectors corresponding to some generalized eigenvalues ​​of the generalized characteristic equation:

[0038]

[0039] in: v k1 and v k2 are the two generalized eigenvectors corresponding to the kth multiple root generalized eigenvalue of the generalized eigenvalue equation, and They are v k1 and v k2 The transposed vector, H k1 and H k2 The composition is as follows:

[0040]

[0041]

[0042] Solve the two-variable linear equation system; finally, calculate the amplitude and phase of the kth simple harmonic component of the complex signal by the following formula:

[0043]

[0044] After obtaining the frequency, amplitude and initial phase of some of the simple harmonic components of the complex signal through path 2, these known simple harmonic components are eliminated from the discrete source signal, the generalized eigenvalue equation is reconstructed, and the method is repeated to continue to decompose and obtain the frequency, amplitude and initial phase of all or part of the simple harmonic components of the remaining part of the complex signal.

[0045] Furthermore, in step S102, a characteristic spectrum decomposition method is used to calculate all generalized eigenvalues ​​and generalized eigenvectors according to the generalized eigenvalue equation, and the characteristic spectrum decomposition method includes:

[0046] Step S301, select matrix A or matrix B, and calculate its eigenvalues ​​and eigenvectors;

[0047] Step S302: According to the calculated eigenvalues, select the eigenvalues ​​whose absolute values ​​are not zero and their corresponding eigenvectors, or select some eigenvalues ​​with larger absolute values ​​and their corresponding eigenvectors as needed;

[0048] Step S303, using the orthogonality of the eigenvectors, converting the generalized eigenvalue equation into a reduced-order standard eigenvalue equation;

[0049] Step S304, calculating all eigenvalues ​​of the standard eigenvalue equation, that is, all eigenvalues ​​of the generalized eigenvalue equation;

[0050] Step S305: reconstruct all generalized eigenvectors of the generalized eigenvalue equation using the eigenvectors of the matrix A or matrix B and the standard eigenvectors of the standard eigenvalue equation.

[0051] Preferably, in step S102, a matrix iteration method is used to calculate some generalized eigenvalues ​​and their corresponding generalized eigenvectors according to the generalized eigenvalue equation, and the matrix iteration method includes:

[0052] Step S401, select matrix A or matrix B, and calculate its eigenvalues ​​and eigenvectors;

[0053] Step S402: According to the calculated eigenvalues, select the eigenvalues ​​whose absolute values ​​are not zero and their corresponding eigenvectors, or select some eigenvalues ​​with larger absolute values ​​and their corresponding eigenvectors as needed;

[0054] Step S403, using the orthogonality of the eigenvectors, converting the generalized eigenvalue equation into a reduced-order standard eigenvalue equation;

[0055] Step S404, selecting an initial eigenvector, setting iteration precision, performing matrix iteration operation on the standard eigenvalue equation, and calculating a standard eigenvalue and its corresponding eigenvector. Repeating this process can calculate several standard eigenvalues ​​of the standard eigenvalue equation and their corresponding standard eigenvectors, wherein the several standard eigenvalues ​​are part of the generalized eigenvalues ​​of the generalized eigenvalue equation;

[0056] Alternatively, a number of initial eigenvectors are selected, the iteration precision is set, and a subspace iteration operation is performed on the standard eigenvalue equation to obtain a number of standard eigenvalues ​​of the standard eigenvalue equation and their corresponding standard eigenvectors. The number of standard eigenvalues ​​obtained in this step are part of the generalized eigenvalues ​​of the generalized eigenvalue equation;

[0057] Step S405: reconstructing some generalized eigenvectors of the generalized eigenvalue equation using the eigenvectors of the matrix A or matrix B and several standard eigenvectors of the standard eigenvalue equation.

[0058] Furthermore, in step S103, the frequency, amplitude and initial phase of all the simple harmonic components, and the frequency, amplitude and initial phase of some of the simple harmonic components are all accurate values.

[0059] The present invention constructs a generalized eigenvalue equation of the frequency of a simple harmonic component of a complex signal, and provides a calculation process of all generalized eigenvalues ​​and generalized eigenvectors of the generalized eigenvalue equation, a calculation process of some generalized eigenvalues ​​of the generalized eigenvalue equation and their corresponding generalized eigenvectors, and a calculation path for solving the frequency, amplitude and initial phase of all simple harmonic components of the complex signal according to all generalized eigenvalues, and a calculation path for solving the frequency, amplitude and initial phase of some simple harmonic components of the complex signal according to some generalized eigenvalues ​​and their corresponding generalized eigenvectors. A precise decomposition method for a complex signal composed of finite length and finite simple harmonic frequency components is provided, and the method includes the construction of a discrete source signal matrix, a way to obtain a discrete source signal differential / integral / differential signal, the construction of a discrete source signal differential / integral / differential signal matrix, and the construction of a generalized eigenvalue equation of the frequency of a simple harmonic component of a complex signal, and the method can accurately decompose and obtain the frequency, amplitude and initial phase of all simple harmonic components and the frequency, amplitude and initial phase of some simple harmonic components.

[0060] The comparison between the signal curve reconstructed by the frequency, amplitude and initial phase of all the simple harmonic components of the complex signal decomposed by the method of the present invention and the source signal curve shows that the two completely overlap; the amplitude-frequency spectrum characteristic curve of the first 99 simple harmonic components of the complex signal decomposed by the method of the present invention is compared with the theoretical value, and the absolute accuracy of the frequency and amplitude of all the simple harmonic components exceeds 0.0001; the initial phase-frequency spectrum characteristic curve of the first 99 simple harmonic components of the complex signal decomposed by the method of the present invention, the maximum value of the absolute value of the initial phase in all the simple harmonic components is 3.1796e-11, compared with the theoretical value of zero, the absolute accuracy of the initial phase in all the simple harmonic components exceeds 3.2e-11. It is fully proved that this method can accurately decompose the frequency, amplitude and initial phase of all the simple harmonic components of a complex signal of finite length and composed of finite simple harmonic frequency components. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a flowchart of a method for accurately decomposing complex signals according to an embodiment of the present invention.

[0062] Figure 2 It is a schematic diagram of the principle of a method for converting a complex source signal into a discrete source signal and differential / integral / differential signals of the discrete source signal.

[0063] Figure 3 It is a schematic diagram of the calculation process of the eigenspectral decomposition method of all generalized eigenvalues ​​and generalized eigenvectors of the generalized eigenvalue equation.

[0064] Figure 4It is a schematic diagram of the calculation process of the matrix iteration method for some generalized eigenvalues ​​of the generalized eigenvalue equation and their corresponding generalized eigenvectors.

[0065] FIG5 is a comparison diagram of the amplitude spectrum, phase spectrum, source signal time domain curve and reconstructed signal time domain curve of all 99-order simple harmonic components of the complex signal decomposed by the method of the present invention.

[0066] 6 is a comparison diagram of the amplitude spectrum, phase spectrum and source signal time domain curve of the first 10-order simple harmonic components of the complex signal decomposed by the method of the present invention and the signal curve reconstructed from the first 10-order simple harmonic components. DETAILED DESCRIPTION

[0067] The principles and advantages of the present invention are further described below in conjunction with the accompanying drawings and embodiments.

[0068] The present invention constructs a generalized eigenvalue equation for characteristic decomposition of complex signals. The generalized eigenvalue and generalized eigenvector of the generalized eigenvalue equation are closely related to the frequency, amplitude and initial phase of the simple harmonic component of the complex signal, and can be expressed in a matrix as follows:

[0069] Bv=-λAv or λ -1 Bv=-Av

[0070] Among them, matrix B is the second-order differential or second-order differential signal matrix of the discrete sequence in matrix A, and the arrangement method is the same. The difference between the second-order differential and the second-order differential is that when performing the second-order differential operation, the second-order change rate of the increment of the discrete function to the discrete variable must be taken into account, while the second-order differential operation only takes into account the second-order increment of the discrete function.

[0071] The composition and combination of the matrix A and the matrix B in the generalized eigenvalue equation include, but are not limited to, the following:

[0072] 1. Matrix A is a discrete source signal matrix of the complex signal and matrix B is a second-order differential signal matrix of the discrete source signal;

[0073] 2. Matrix A is a discrete source signal matrix of the complex signal and matrix B is a quadratic differential signal matrix of the discrete source signal;

[0074] 3. Matrix A is a first-order integral signal matrix of the discrete source signal of the complex signal and matrix B is a first-order differential signal matrix of the discrete source signal;

[0075] 4. Matrix A is a primary integral signal matrix of the discrete source signal of the complex signal and matrix B is a primary differential signal matrix of the discrete source signal;

[0076] 5. Matrix A is the quadratic integral signal matrix of the discrete source signal of the complex signal and matrix B is composed of the discrete source signal matrix.

[0077] And so on.

[0078] If the matrix A is a discrete source signal matrix, it can be expressed as:

[0079]

[0080] Where: x1,x2,x3,…,x 2n are the 2n discrete values ​​of the complex signal, ie, discrete source signals.

[0081] The arrangement order of the elements in the matrix B corresponds one-to-one to the elements in the matrix A. The arrangement modes of the matrices of the various combinations are similar.

[0082] The first-order differential, second-order differential, first-order integral, second-order integral, first-order differential, and second-order differential of the complex signal discrete source signal can all be realized by the hardware of the calculus circuit. The principle is as shown in the attached Figure 2 As shown; or obtained by performing first-order differential, second-order differential, first-order integral, second-order integral, first-order differential, and second-order differential numerical calculations on the discrete source signal sequence.

[0083] The generalized eigenvalue of the generalized eigenvalue equation is closely related to the circular frequency of the simple harmonic component of the complex signal, and the numerical relationship between them is divided into two cases. The first case: Matrix B is the second-order differential signal matrix of Matrix A, and the circular frequency of the simple harmonic component of the complex signal is:

[0084]

[0085] Where: i is the i-th generalized eigenvalue of the generalized eigenvalue equation.

[0086] The second case: Matrix B is the quadratic differential signal matrix of Matrix A. In this case, the circular frequency of the simple harmonic component of the complex signal is:

[0087]

[0088] Where: i is the i-th generalized eigenvalue of the generalized eigenvalue equation, and Δt is the discrete interval of the discrete source signal.

[0089] After obtaining the circular frequency of the simple harmonic component of the complex signal, there are two paths to obtain the amplitude and phase of the simple harmonic component.

[0090] Path 1: First, the circular frequencies of all simple harmonic components of the complex signal are calculated according to all the generalized eigenvalues ​​of the generalized eigenvalue equation; then a linear equation system Dy=z is constructed, where: D=C T C, z = C T x, the vector x is composed of the discrete source signals of the complex signal, and the matrix C is composed as follows:

[0091]

[0092] Where: 2m-1≤k≤2n-1, m is the number of independent generalized eigenvalues ​​calculated, C T is the transposed matrix of matrix C; solve the linear equations to obtain vector y; finally, the amplitude and phase of each component are calculated by the following formula:

[0093]

[0094] Where: y 2i-1 and 2i are the 2i-1th and 2ith elements of the vector respectively.

[0095] Path 2: First, the circular frequencies of some simple harmonic components of the complex signal are calculated according to some generalized eigenvalues ​​of the generalized eigenvalue equation; then, the following two-variable linear equation system is constructed using some generalized eigenvectors corresponding to some generalized eigenvalues ​​of the generalized characteristic equation:

[0096]

[0097] in: v k1 and v k2 are the two generalized eigenvectors corresponding to the kth multiple root generalized eigenvalue of the generalized eigenvalue equation, and They are v k1 and v k2 The transposed vector, H k1 and H k2 The composition is as follows:

[0098]

[0099]

[0100] Solve the two-variable linear equation system; finally, calculate the amplitude and phase of the kth simple harmonic component of the complex signal by the following formula:

[0101]

[0102] After obtaining the frequency, amplitude and initial phase of some of the simple harmonic components of the complex signal through path 2, these known simple harmonic components are eliminated from the discrete source signal, the generalized eigenvalue equation is reconstructed, and the method is repeated to continue to decompose and obtain the frequency, amplitude and initial phase of all or part of the simple harmonic components of the remaining part of the complex signal.

[0103] There are two methods for solving the generalized eigenvalue problem of the generalized eigenvalue equation, namely, the characteristic spectrum decomposition calculation method and the matrix iterative calculation method, as shown in the attached Figure 3 and attached Figure 4 shown.

[0104] The characteristic spectrum decomposition method includes: step S301, selecting matrix A or matrix B, and calculating its eigenvalues ​​and eigenvectors; step S302, selecting eigenvalues ​​with absolute values ​​not zero and their corresponding eigenvectors according to the calculated eigenvalues, or selecting some eigenvalues ​​with larger absolute values ​​and their corresponding eigenvectors as needed; step S303, using the orthogonality of eigenvectors to convert the generalized eigenvalue equation into a reduced-order standard eigenvalue equation; step S304, calculating all eigenvalues ​​of the standard eigenvalue equation, that is, all eigenvalues ​​of the generalized eigenvalue equation; step S305, reconstructing all generalized eigenvectors of the generalized eigenvalue equation using the eigenvectors of the matrix A or matrix B and the standard eigenvectors of the standard eigenvalue equation.

[0105] The matrix iteration method includes: step S401, selecting matrix A or matrix B, and calculating its eigenvalues ​​and eigenvectors; step S402, selecting eigenvalues ​​with non-zero absolute values ​​and their corresponding eigenvectors according to the calculated eigenvalues, or selecting some eigenvalues ​​with larger absolute values ​​and their corresponding eigenvectors as needed; step S403, using the orthogonality of the eigenvectors to convert the generalized eigenvalue equation into a reduced-order standard eigenvalue equation; step S404, selecting an initial eigenvector, setting the iteration accuracy, performing matrix iteration operation on the standard eigenvalue equation, and then calculating a standard eigenvalue and its corresponding eigenvector. Repeating this process can calculate the standard Several standard eigenvalues ​​of the eigenvalue equation and their corresponding standard eigenvectors, the several standard eigenvalues ​​are just some generalized eigenvalues ​​of the generalized eigenvalue equation; or several initial eigenvectors are selected, the iteration precision is set, and subspace iteration operation is performed on the standard eigenvalue equation to obtain several standard eigenvalues ​​of the standard eigenvalue equation and their corresponding standard eigenvectors, the several standard eigenvalues ​​obtained in this step are just some generalized eigenvalues ​​of the generalized eigenvalue equation; step S405, reconstructing some generalized eigenvectors of the generalized eigenvalue equation using the eigenvectors of the matrix A or the matrix B and the several standard eigenvectors of the standard eigenvalue equation.

[0106] An embodiment of the present invention also provides a method for accurate decomposition of a finite signal, comprising the following steps: step S101, constructing a discrete source signal matrix based on the source signal of the complex signal, and obtaining a differential / integral / differential signal sequence of the discrete source signal from the source signal of the complex signal, thereby constructing a differential / integral / differential signal matrix of the discrete source signal, and using the discrete source signal matrix and / or the differential / integral / differential signal matrix of the discrete source signal to construct a generalized eigenvalue equation of the frequency of the simple harmonic component of the complex signal. The difference between the differential and the special differential is that the rate of change of the increment of the discrete function to the discrete variable must be taken into account when performing the differential operation, while the differential operation only takes into account the increment of the discrete function; step S102, calculate all the generalized eigenvalues ​​and generalized eigenvectors according to the generalized eigenvalue equation; or / and calculate some generalized eigenvalues ​​of the generalized eigenvalue equation and their corresponding generalized eigenvectors; step S103, solve the frequency, amplitude and initial phase of all simple harmonic components of the complex signal according to all the generalized eigenvalues; and / or solve the frequency, amplitude and initial phase of some simple harmonic components of the complex signal according to some generalized eigenvalues ​​and their corresponding generalized eigenvectors.

[0107] Figure 1 The flowchart of the method for accurate decomposition of simple harmonic components of a complex signal according to an embodiment of the present invention is schematically shown.

[0108] like Figure 1As shown, the method includes steps S101 to S103.

[0109] In step S101, a discrete source signal matrix is ​​constructed according to the source signal of the complex signal, and a differential / integral / differential signal sequence of the discrete source signal is obtained from the source signal, and on this basis, a differential / integral / differential signal matrix of the discrete source signal is constructed; then, the matrix is ​​used to construct a generalized eigenvalue equation about the frequency of the simple harmonic component of the complex signal.

[0110] In step S102, all or part of the generalized eigenvalues ​​and their corresponding generalized eigenvectors are solved according to the generalized eigenvalue equation.

[0111] In step S103, the frequency, amplitude and initial phase of all simple harmonic components of the complex signal are solved according to the whole generalized eigenvalues, or the frequency, amplitude and initial phase of some simple harmonic components of the complex signal are solved according to the partial generalized eigenvalues ​​and their corresponding generalized eigenvectors.

[0112] Figure 2 The principle diagram of the method for converting the complex signal source signal into a discrete source signal and a differential / integral / differential signal of the discrete source signal is schematically shown.

[0113] like Figure 2 As shown, the method includes two paths: step S201 to step S213 and step S201 to step S224.

[0114] In step S201, the complex source signal is divided into two parallel signals and output synchronously.

[0115] In step S212, the complex source signal is A / D converted and discretized.

[0116] In step S213, a discrete source signal of the complex signal is obtained.

[0117] In step S222, a differentiation / integration operation is performed on the complex source signal by a differentiation / integration hardware circuit.

[0118] In step S223, the differential / integral signal of the complex signal is A / D converted and discretized.

[0119] In step S224, a differential / integral / differential signal sequence of the discrete source signal is obtained. The difference between differential and differentiation is that the differential operation takes into account the rate of change of the discrete function increment to the discrete variable, while the differential operation only takes into account the discrete function increment.

[0120] The differential / integral / difference signal sequence of discrete source signals can also be obtained through numerical calculation methods.

[0121] After obtaining the discrete source signal and the differential / integral / differential signal sequence of the discrete source signal, the matrix A and the matrix B are constructed according to the structure and combination principle of the matrix A and the matrix B.

[0122] Then, the matrix A and the matrix B are used to construct a generalized eigenvalue equation about the frequency of the simple harmonic component of the complex signal.

[0123] Figure 3 A calculation flow chart of the eigenspectral decomposition method of all generalized eigenvalues ​​and generalized eigenvectors of the generalized eigenvalue equation is schematically shown.

[0124] like Figure 3 As shown, the method includes steps S301 to S305.

[0125] In step S301, matrix A or matrix B is selected, and its eigenvalues ​​and eigenvectors are calculated.

[0126] In step S302, according to the calculated eigenvalues, eigenvalues ​​with non-zero absolute values ​​and their corresponding eigenvectors are selected, or some eigenvalues ​​with larger absolute values ​​and their corresponding eigenvectors are selected as needed.

[0127] In step S303, the generalized eigenvalue equation is converted into a reduced-order standard eigenvalue equation by utilizing the orthogonality of the eigenvectors.

[0128] In step S304, all eigenvalues ​​of the standard eigenvalue equation, that is, all eigenvalues ​​of the generalized eigenvalue equation, are calculated.

[0129] In step S305, all generalized eigenvectors of the generalized eigenvalue equation are reconstructed using the eigenvectors of the matrix A or matrix B and the standard eigenvectors of the standard eigenvalue equation.

[0130] After obtaining all the generalized eigenvalues ​​of the generalized eigenvalue equation, the circular frequencies of all the simple harmonic components are calculated according to the relationship between the generalized eigenvalues ​​and the circular frequencies of the simple harmonic components of the complex signal described in the present invention, and then according to the method described in path 1: construct a matrix C; solve the linear equations; calculate the amplitudes and initial phases of all the simple harmonic components of the complex signal.

[0131] Figure 4 The flowchart schematically shows the calculation flow of the matrix iteration method for some generalized eigenvalues ​​of the generalized eigenvalue equation and their corresponding generalized eigenvectors.

[0132] like Figure 4 As shown, the method includes steps S401 to S405.

[0133] Steps S401 to S403 and Figure 3Steps S301 to S303 are the same.

[0134] In step S404, an initial eigenvector is selected, the iteration precision is set, and a matrix iteration operation is performed on the standard eigenvalue equation to calculate a standard eigenvalue and its corresponding eigenvector. The process can be repeated to calculate several standard eigenvalues ​​of the standard eigenvalue equation and their corresponding standard eigenvectors; or several initial eigenvectors are selected, the iteration precision is set, and a subspace iteration operation is performed on the standard eigenvalue equation to obtain several standard eigenvalues ​​of the standard eigenvalue equation and their corresponding standard eigenvectors. The several standard eigenvalues ​​obtained in this step are part of the generalized eigenvalues ​​of the generalized eigenvalue equation.

[0135] In step S405, the eigenvectors of the matrix A or matrix B and several standard eigenvectors of the standard eigenvalue equation are used to reconstruct some generalized eigenvectors of the generalized eigenvalue equation.

[0136] After obtaining the partial generalized eigenvalues ​​of the generalized eigenvalue equation and the corresponding generalized eigenvalue vectors, according to the relationship between the generalized eigenvalues ​​and the simple harmonic component frequencies of the complex signal described in the present invention, the circular frequencies of the corresponding partial simple harmonic components are calculated, and then the method described in path 2 is followed: construct the matrix H k1 , H k2 , calculate e k11 、e k12 、e k21 、e k22 and g k1 , g k2 ; Solve the set of two-variable linear equations; Calculate the amplitude and initial phase of the simple harmonic component of the corresponding part of the complex signal.

[0137] Figure 5 schematically shows the amplitude spectrum, phase spectrum and comparison of the source signal time domain curve and the reconstructed signal curve of all simple harmonic components of the complex signal decomposed by the method of the present invention. The complex signal is composed of the first 99 simple harmonic components of a periodic square wave with a period of 2 seconds and an amplitude of 1.

[0138] Figure 5A The amplitude-frequency spectrum characteristic curve of the first 99-order simple harmonic components of the complex signal decomposed by the method of the present invention is compared with the theoretical value. The absolute accuracy of the frequency and amplitude of all simple harmonic components exceeds 0.0001.

[0139] Figure 5BThe initial phase-frequency spectrum characteristic curve of the first 99-order simple harmonic components of the complex signal decomposed by the method of the present invention is shown in FIG. 1 . The maximum absolute value of the initial phase of all simple harmonic components is 3.1796e-11, while the theoretical value is zero, that is, the absolute accuracy exceeds 3.2e-11.

[0140] Figure 5C This is a comparison diagram of the time domain curve of the reconstructed signal of the first 99-order simple harmonic components of the complex signal decomposed by the method of the present invention and the time domain curve of the source signal. In the figure, the source signal curve and the reconstructed signal curve completely overlap.

[0141] Fig. 6 schematically shows the amplitude spectrum, phase spectrum and time domain curve of the first 10 harmonic components of the complex signal decomposed by the method of the present invention and the comparison diagram of the signal curve reconstructed by the first 10 harmonic components. The complex signal is composed of the first 99 harmonic components of a periodic square wave with a period of 2 seconds and an amplitude of 1.

[0142] Fig. 6A The amplitude-frequency spectrum characteristic curve of the first 10 harmonic components of the complex signal decomposed by the method of the present invention is compared with the theoretical value. The absolute accuracy of the frequency and amplitude of all harmonic components exceeds 0.0001.

[0143] Figure 6B The initial phase-frequency spectrum characteristic curve of the first 10-order simple harmonic components of the complex signal decomposed by the method of the present invention, the maximum absolute value of the initial phase in all simple harmonic components is 9.9863e-7, while the theoretical value is zero, that is, the absolute accuracy exceeds 1.0e-6.

[0144] Figure 6C This is a comparison diagram of the time domain curve of the reconstructed signal of the first 10-order simple harmonic components of the complex signal decomposed by the method of the present invention and the time domain curve of the source signal. The difference between the source signal curve and the reconstructed signal curve in the figure is caused by the fact that the 89-order simple harmonic component is not included in the calculation.

Claims

1. A method for constructing an equation for accurate decomposition of a finite signal, wherein the finite signal is a complex signal of finite length and composed of a finite number of simple harmonic components, and the equation is a generalized eigenvalue equation for accurate decomposition of the complex signal, characterized in that: The generalized eigenvalue equation is expressed in a matrix as: Bv=-λAv or λ -1 Bv=-Off Among them, matrix A is a discrete source signal matrix or a differential / integral / differential signal matrix of a discrete source signal; matrix B is a second-order differential or quadratic differential signal matrix of matrix A and is arranged in the same way. The difference between the differential and the differential is that when performing a differential operation, the rate of change of the increment of the discrete function to the discrete variable must be taken into account, while the differential operation only takes into account the increment of the discrete function; λ is the generalized eigenvalue of the generalized eigenvalue equation; v is the generalized eigenvector of the generalized eigenvalue equation.

2. The method for constructing an equation according to claim 1, characterized in that: If the matrix A is a discrete source signal matrix, it is expressed as: Where: x1,x2,x3,…,x 2n is the 2n discrete values ​​of the complex signal, i.e., the discrete source signal, then the matrix B is a second-order differential or second-order differential signal matrix composed of a discrete sequence of second-order differentials or second-order differentials of the discrete source signal, and the elements of the matrix B have the same arrangement as the elements of the matrix A; The matrix A and the matrix B in the generalized eigenvalue equation are selected from one or more of the following matrices: a discrete source signal matrix of the complex signal, an n-order differential signal matrix of the discrete source signal of the complex signal, an n-time integral signal matrix of the discrete source signal of the complex signal, and an n-time differential signal matrix of the discrete source signal of the complex signal, wherein n is a natural number; The difference between the nth differential and the nth order differential is that when performing the nth order differential operation, the nth order rate of change of the increment of the discrete function to the discrete variable must be taken into account, while the nth differential operation only takes into account the nth increment of the discrete function; There are two principles for combining the matrix A and the matrix B in the generalized eigenvalue equation: the first is that the matrix B is the second-order differential of the matrix A; the second is that the matrix B is the quadratic differential of the matrix A.

3. The method for constructing an equation according to claim 1 or 2, characterized in that: The discrete source signal of the complex signal and the n-order differential, n-order integral and n-order differential of the discrete source signal of the complex signal can be realized by hardware of the calculus circuit; The source signal of the complex signal is divided into two parallel signal paths for synchronous output, the first path including the steps of performing A / D conversion on the source signal of the complex signal and discretizing it to obtain a discrete source signal of the complex signal; The second path includes the following steps: performing differentiation / integration operations on the source signal of the complex signal through a differentiation / integration hardware circuit, performing A / D conversion on the differentiation / integration signal of the complex signal and discretizing it to obtain a differentiation / integration / differential signal sequence of the discrete source signal.

4. The method for constructing an equation according to claim 1 or 2, characterized in that: The n-order differential, n-times integral and n-times differential of the complex signal discrete source signal can be obtained by numerically calculating the n-order differential, n-times integral and n-times differential of the discrete source signal sequence.

5. A method for accurately decomposing a finite signal, wherein the finite signal is a complex signal of finite length and composed of a finite number of simple harmonic components, characterized in that: The following steps are included: Step S101, constructing a discrete source signal matrix according to the source signal of the complex signal, and obtaining a differential / integral / differential signal sequence of the discrete source signal from the source signal of the complex signal, thereby constructing a differential / integral / differential signal matrix of the discrete source signal, and using the discrete source signal matrix and / or the differential / integral / differential signal matrix of the discrete source signal, using any one of the methods of claims 1 to 4 to construct a generalized eigenvalue equation about the frequency of a simple harmonic component of the complex signal; Step S102, calculating all generalized eigenvalues ​​and generalized eigenvectors according to the generalized eigenvalue equation of the simple harmonic component frequency of the complex signal; and / or calculating partial generalized eigenvalues ​​of the generalized eigenvalue equation and their corresponding generalized eigenvectors; Step S103, calculating the frequency, amplitude and initial phase of all simple harmonic components of the complex signal according to all generalized eigenvalues; And / or the frequency, amplitude and initial phase of some simple harmonic components of the complex signal are solved according to some generalized eigenvalues ​​and their corresponding generalized eigenvectors.

6. The accurate decomposition method of finite signal according to claim 5, characterized in that: In step S102, the numerical relationship between the generalized eigenvalue of the simple harmonic component frequency of the complex signal and the circular frequency of the simple harmonic component of the complex signal is divided into two cases. In the first case, matrix B is the second-order differential signal matrix of matrix A. In this case, the circular frequency of the simple harmonic component of the complex signal is: Where: i is the i-th generalized eigenvalue of the generalized eigenvalue equation; The second case: Matrix B is the quadratic differential signal matrix of Matrix A. In this case, the circular frequency of the simple harmonic component of the complex signal is: Where: i is the i-th generalized eigenvalue of the generalized eigenvalue equation, and Δt is the discrete interval of the discrete source signal.

7. The accurate decomposition method of a finite signal according to claim 5, characterized in that: In step S103, there are two paths to calculate the frequency, amplitude and initial phase of the simple harmonic component. Path 1: First, the circular frequencies of all simple harmonic components of the complex signal are calculated according to all the generalized eigenvalues ​​of the generalized eigenvalue equation; then a linear equation system Dy=z is constructed, where: D=C T C, z = C T x, the vector x is composed of the discrete source signals of the complex signal, and the matrix C is composed as follows: Where: 2m-1≤k≤2n-1, m is the number of independent generalized eigenvalues ​​calculated; n is the dimension of the generalized eigenvalue equation; C T is the transposed matrix of matrix C; solving the linear equations obtains vector y; is the phase of the mth component signal; finally, the amplitude and phase of each component are calculated by the following formula: Where: y 2i-1 and 2i are the 2i-1th and 2ith elements of the vector respectively; tg -1 is the inverse tangent function: Path 2: First, the circular frequencies of some simple harmonic components of the complex signal are calculated according to some generalized eigenvalues ​​of the generalized eigenvalue equation; then, the following two-variable linear equation system is constructed using some generalized eigenvectors corresponding to some generalized eigenvalues ​​of the generalized eigenvalue equation: in: v k1 and v k2 are the two generalized eigenvectors corresponding to the kth multiple root generalized eigenvalue of the generalized eigenvalue equation, and They are v k1 and v k2 The transposed vector, f k1 and f k2 is the variable to be determined, H k1 and H k2 The composition is as follows: A is the discrete source signal matrix, and the two-variable linear equation system is solved; finally, the amplitude and phase of the kth simple harmonic component of the complex signal are calculated by the following formula: After obtaining the frequency, amplitude and initial phase of some of the simple harmonic components of the complex signal through path 2, these known simple harmonic components are eliminated from the discrete source signal, the generalized eigenvalue equation is reconstructed, and the method is repeated to continue to decompose and obtain the frequency, amplitude and initial phase of all or part of the simple harmonic components of the remaining part of the complex signal.

8. The accurate decomposition method of finite signal according to claim 5, characterized in that: In step S102, a characteristic spectrum decomposition method is used to calculate all generalized eigenvalues ​​and generalized eigenvectors according to the generalized eigenvalue equation, and the characteristic spectrum decomposition method includes: Step S301, select matrix A or matrix B, and calculate its eigenvalues ​​and eigenvectors; Step S302: According to the calculated eigenvalues, select eigenvalues ​​whose absolute values ​​are not zero and their corresponding eigenvectors; Step S303, using the orthogonality of the eigenvectors, converting the generalized eigenvalue equation into a reduced-order standard eigenvalue equation; Step S304, calculating all eigenvalues ​​of the standard eigenvalue equation, that is, all generalized eigenvalues ​​of the generalized eigenvalue equation; Step S305: reconstruct all generalized eigenvectors of the generalized eigenvalue equation using the eigenvectors of the matrix A or matrix B and the standard eigenvectors of the standard eigenvalue equation.

9. The accurate decomposition method of finite signal according to claim 5, characterized in that: In step S102, a matrix iteration method is used to calculate some generalized eigenvalues ​​and their corresponding generalized eigenvectors according to the generalized eigenvalue equation, and the matrix iteration method includes: Step S401, select matrix A or matrix B, and calculate its eigenvalues ​​and eigenvectors; Step S402: According to the calculated eigenvalues, select eigenvalues ​​whose absolute values ​​are not zero and their corresponding eigenvectors; Step S403, using the orthogonality of the eigenvectors, converting the generalized eigenvalue equation into a reduced-order standard eigenvalue equation; Step S404, selecting an initial eigenvector, setting iteration precision, performing matrix iteration operation on the standard eigenvalue equation, and calculating a standard eigenvalue and its corresponding eigenvector. Repeating this process can calculate several standard eigenvalues ​​of the standard eigenvalue equation and their corresponding standard eigenvectors, wherein the several standard eigenvalues ​​are part of the generalized eigenvalues ​​of the generalized eigenvalue equation; Alternatively, a number of initial eigenvectors are selected, the iteration precision is set, and a subspace iteration operation is performed on the standard eigenvalue equation to obtain a number of standard eigenvalues ​​of the standard eigenvalue equation and their corresponding standard eigenvectors. The number of standard eigenvalues ​​obtained in this step are part of the generalized eigenvalues ​​of the generalized eigenvalue equation; Step S405: reconstructing some generalized eigenvectors of the generalized eigenvalue equation using the eigenvectors of the matrix A or matrix B and several standard eigenvectors of the standard eigenvalue equation.

10. The accurate decomposition method of finite signal according to claim 5, characterized in that: In step S103, the frequency, amplitude and initial phase of all the simple harmonic components, and the frequency, amplitude and initial phase of some simple harmonic components are all accurate values.