Underdetermined blind source separation method, device and electronic device based on continuous wavelet transform

Abstract

The present invention provides an underdetermined blind source separation method, device and electronic device based on continuous wavelet transform, which can be applied to the field of pan-information communication technology. The method comprises: performing continuous wavelet transform on multi-channel signals to be separated to obtain complex signals; wherein the multi-channel signals to be separated are obtained by linearly combining multiple signal sources, and the signal sources include target signals and noise signals; determining multiple single source points from all time-frequency points corresponding to the complex signals, and processing the multiple single source points to determine the coefficients of the underdetermined equation group; performing short-time Fourier transform on the multi-channel signals to be separated to obtain the constant terms of the underdetermined equation group; solving the underdetermined equation group according to the coefficients of the underdetermined equation group and the constant terms of the underdetermined equation group to obtain a prediction signal corresponding to the target signal.

Description

Technical Field

[0001] The present invention relates to the field of pan-information communication technology, and more specifically, to an underdetermined blind source separation method, device and electronic equipment based on continuous wavelet transform. Background Art

[0002] Underdetermined Blind Source Separation (UBSS) is an algorithm framework for separating mixed signals. It can be widely used in speech recognition and wireless communication fields, and is suitable for situations where the number of sensors is less than the number of signal sources.

[0003] One of the technical paths of UBSS can be achieved by sparse component analysis (SCA) and compressed sensing. This technical path only requires that the input signal and the output signal have sparseness, so it is necessary to use appropriate time-frequency analysis methods to introduce sparsity.

[0004] In the process of realizing the concept of the present invention, it was found through research that there are at least technical problems in the related technology that the time-frequency analysis method has low robustness, complex calculation process and is difficult to apply to high-noise application scenarios. Summary of the invention

[0005] In view of this, the present invention provides an underdetermined blind source separation method, device and electronic device based on continuous wavelet transform.

[0006] One aspect of the present invention provides an underdetermined blind source separation method based on continuous wavelet transform, comprising:

[0007] A multi-channel signal to be separated is subjected to continuous wavelet transform to obtain a complex signal; wherein the multi-channel signal to be separated is obtained by linearly combining multiple signal sources, and the signal sources include a target signal and a noise signal; multiple single source points are determined from all time-frequency points corresponding to the complex signal, and the multiple single source points are processed to determine the coefficients of the underdetermined equation group; a multi-channel signal to be separated is subjected to short-time Fourier transform to obtain the constant term of the underdetermined equation group; the underdetermined equation group is solved according to the coefficients of the underdetermined equation group and the constant term of the underdetermined equation group to obtain a predicted signal corresponding to the target signal.

[0008] According to an embodiment of the present invention, a plurality of single source points are determined from all time-frequency points corresponding to a complex signal, including: obtaining characteristic parameters corresponding to all time-frequency points according to the real part and the imaginary part of the complex signal values ​​corresponding to all time-frequency points; and screening out a plurality of single source points from all time-frequency points corresponding to the complex signal according to the characteristic parameters corresponding to all time-frequency points and a predetermined threshold.

[0009] According to an embodiment of the present invention, multiple single source points are processed to determine the coefficients of the underdetermined equation group, including: normalizing the multiple single source points according to the corresponding relationship between the multiple single source points and the multiple signal sources to obtain the influence parameters of the multiple signal sources on the multi-channel signals to be separated; and obtaining the coefficients of the underdetermined equation group according to the influence parameters.

[0010] According to an embodiment of the present invention, coefficients of an underdetermined system of equations are obtained based on influencing parameters, including: determining a matrix vector corresponding to the coefficients of the underdetermined system of equations based on the influencing parameters; determining a compensation vector corresponding to the coefficients of the underdetermined system of equations based on a noise signal; and obtaining the coefficients of the underdetermined system of equations based on the matrix vector corresponding to the coefficients of the underdetermined system of equations and the compensation vector.

[0011] According to an embodiment of the present invention, the underdetermined equation group includes multiple underdetermined equations corresponding to time-frequency points, and a short-time Fourier transform is performed on the multi-channel signals to be separated to obtain constant terms of the underdetermined equation group, including: performing a short-time Fourier transform on the multi-channel signals to be separated to obtain time-frequency domain signals; determining the constant term corresponding to each underdetermined equation based on the time-frequency domain signal value corresponding to each time-frequency point; and obtaining the constant term of the underdetermined equation group based on the constant term corresponding to each underdetermined equation.

[0012] According to an embodiment of the present invention, the underdetermined system of equations is solved according to the coefficients of the underdetermined system of equations and the constant terms of the underdetermined system of equations to obtain a prediction signal corresponding to the target signal, including: constructing an underdetermined system of equations according to the coefficients of the underdetermined system of equations and the constant terms of the underdetermined system of equations; solving the underdetermined system of equations through compressed sensing processing to achieve the equivalence of the solution of the underdetermined system of equations to linear programming; obtaining complex data of the prediction signal corresponding to the target signal according to the solution of the underdetermined system of equations; and performing short-time inverse Fourier transform on the complex data to obtain a prediction signal corresponding to the target signal.

[0013] According to an embodiment of the present invention, the method further comprises: performing filtering processing on the prediction signal to obtain a separation signal.

[0014] Another aspect of the present invention provides an underdetermined blind source separation device based on continuous wavelet transform, comprising:

[0015] A first transformation module is used to perform continuous wavelet transformation on the multi-channel signal to be separated to obtain a complex signal; wherein the multi-channel signal to be separated is obtained by linearly combining multiple signal sources, and the signal sources include target signals and noise signals;

[0016] A coefficient determination module is used to determine a plurality of single source points from all time-frequency points corresponding to the complex signal, and process the plurality of single source points to determine the coefficients of the underdetermined equation group;

[0017] The second transformation module is used to perform short-time Fourier transformation on the multi-channel signals to be separated to obtain constant terms of the underdetermined equations;

[0018] The signal solving module is used to solve the underdetermined equation group according to the coefficients of the underdetermined equation group and the constant term of the underdetermined equation group to obtain a prediction signal corresponding to the target signal.

[0019] Another aspect of the present invention provides an electronic device, comprising:

[0020] one or more processors;

[0021] a memory for storing one or more programs,

[0022] When the one or more programs are executed by the one or more processors, the one or more processors implement the method as described above.

[0023] Another aspect of the present invention provides a computer program product, the computer program product comprising computer executable instructions, the instructions being used to implement the method as described above when being executed.

[0024] According to an embodiment of the present invention, a continuous wavelet transform is performed on a multi-channel signal to be separated obtained by linearly combining multiple signal sources, a single source point is screened for the obtained complex signal, the coefficients of the underdetermined system of equations are determined, and then a short-time Fourier transform is performed on the multi-channel signal to be separated to obtain the constant term of the underdetermined system of equations. Since the two time-frequency analysis methods of continuous wavelet transform and short-time Fourier transform are combined to construct an underdetermined linear system of equations and solve the predicted signal, the inverse transform of the continuous wavelet transform is avoided, thereby avoiding the distortion of the predicted signal, and the robustness is high; and since the coefficients are obtained by processing the single source point, there is no need to pre-process the data, so the algorithm cost is low, and the required input data is less, so it can be applied to application scenarios including high noise. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] The above and other objects, features and advantages of the present invention will become more apparent through the following description of the embodiments of the present invention with reference to the accompanying drawings, in which:

[0026] Figure 1A flow chart of an underdetermined blind source separation method based on continuous wavelet transform according to an embodiment of the present invention is shown.

[0027] Figure 2 A schematic diagram of an algorithm flow framework according to an embodiment of the present invention is shown.

[0028] Figure 3 A schematic diagram of a target signal according to an embodiment of the present invention is shown.

[0029] Figure 4a A schematic diagram of a sinusoidal interference signal with a frequency of 5 Hz according to an embodiment of the present invention is shown.

[0030] Figure 4b A schematic diagram of a triangular wave interference signal with a frequency of 2 Hz according to an embodiment of the present invention is shown.

[0031] Figure 4c A schematic diagram of a sinusoidal interference signal with a frequency of 0.8 Hz according to an embodiment of the present invention is shown.

[0032] Figure 4d A schematic diagram of a damped oscillation interference signal with a frequency of 3 Hz and a damping coefficient of 0.01 according to an embodiment of the present invention is shown.

[0033] Figure 5a A schematic diagram showing an algorithm input signal according to an embodiment of the present invention is shown.

[0034] Figure 5b A schematic diagram showing an algorithm output signal according to an embodiment of the present invention is shown.

[0035] Figure 6a A schematic diagram of an effect corresponding to a target signal in a related technology according to an embodiment of the present invention is shown.

[0036] Figure 6b A schematic diagram of the effect corresponding to the reverse target signal in the related art according to an embodiment of the present invention is shown.

[0037] Figure 6c A schematic diagram showing the effect of the underdetermined blind source separation method based on continuous wavelet transform according to an embodiment of the present invention corresponding to the target signal is shown.

[0038] Figure 6d A schematic diagram showing the effect of the underdetermined blind source separation method based on continuous wavelet transform according to an embodiment of the present invention corresponding to the target signal is shown.

[0039] Figure 7 A block diagram of an underdetermined blind source separation device based on continuous wavelet transform according to an embodiment of the present invention is shown.

[0040] Figure 8A block diagram of an electronic device suitable for implementing the above-described underdetermined blind source separation method based on continuous wavelet transform according to an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0041] Below, embodiments of the present invention will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present invention. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of embodiments of the present invention. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of concepts of the present invention.

[0042] The terms used herein are only for describing specific embodiments and are not intended to limit the present invention. The terms "comprise", "include", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0043] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0044] When using expressions such as "at least one of A, B, and C, etc.", they should generally be interpreted according to the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B, and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0045] In the embodiments of the present invention, the collection, updating, analysis, processing, use, transmission, provision, disclosure, storage and other aspects of the data involved (for example, including but not limited to user personal information) are in compliance with the provisions of relevant laws and regulations, are used for legitimate purposes, and do not violate public order and good morals. In particular, necessary measures are taken for user personal information to prevent illegal access to user personal information data and maintain the security of user personal information and network security.

[0046] In the embodiment of the present invention, before obtaining or collecting the user's personal information, the user's authorization or consent is obtained.

[0047] UBSS is a technology used to separate mixed signals. For the above-mentioned technical path consisting of SCA and compressed sensing, the input signal and the output signal need to be sparse. Therefore, related technologies often use time-frequency signal analysis to convert time-varying observation signals into the time-frequency domain to obtain additional information and introduce sparsity.

[0048] Short-time Fourier transform (STFT) is a basic method of time-frequency analysis, which is evolved from the Fourier transform frequency domain analysis method. The essence of STFT is to obtain time information by windowing at time points, and then perform frequency domain analysis on signal segments. This also leads to the selection of window functions, time points and frequency points affecting the analysis results, resulting in poor robustness. Usually complex preprocessing is required for UBSS, which also increases the computational cost.

[0049] Constant Q non-stationary Gabor transform (CQ-NSGT) is a self-windowing time-frequency analysis method based on the improved STFT. It can reduce the input by automatically identifying the characteristics of the input signal. Although it can improve a certain degree of robustness, its time and frequency points still need to be manually selected, and the adaptive selection often requires cyclic calculations, and the algorithm is complex.

[0050] Wavelet analysis mainly includes discrete wavelet analysis, wavelet packet analysis, and continuous wavelet analysis (CWT), all of which can achieve the requirements of time segmentation at high frequencies, frequency segmentation at low frequencies, and automatically adapt to time-frequency analysis. However, the principles of discrete wavelet analysis and wavelet packet analysis are different from STFT, so they are not suitable for this technical path. Continuous wavelet analysis lacks a reliable inverse operation process, and the signal restoration process obtained by separation is prone to distortion.

[0051] In view of this, in order to at least partially solve the above technical problems, an embodiment of the present invention provides an underdetermined blind source separation method based on continuous wavelet transform, comprising: performing continuous wavelet transform on a multi-channel signal to be separated to obtain a complex signal; wherein the multi-channel signal to be separated is obtained by linearly combining multiple signal sources, and the signal source includes a target signal and a noise signal; determining multiple single source points from all time-frequency points corresponding to the complex signal, and processing the multiple single source points to determine the coefficients of the underdetermined equation group; performing short-time Fourier transform on the multi-channel signal to be separated to obtain the constant term of the underdetermined equation group; solving the underdetermined equation group according to the coefficients of the underdetermined equation group and the constant term of the underdetermined equation group to obtain a predicted signal corresponding to the target signal.

[0052] Figure 1 A flow chart of an underdetermined blind source separation method based on continuous wavelet transform according to an embodiment of the present invention is shown.

[0053] like Figure 1 As shown, the method 100 includes operations S110 to S140.

[0054] In operation S110, continuous wavelet transform is performed on the multi-channel signals to be separated to obtain complex signals.

[0055] According to an embodiment of the present invention, the multi-channel signal to be separated is obtained by linearly combining multiple signal sources, and multiple sensors are used to respond to the multi-channel signal to be separated. The signal sources include target signals and noise signals.

[0056] According to an embodiment of the present invention, a complex wavelet should be used to perform continuous wavelet transform on a multi-channel signal to be separated so as to retain the phase information of the signal, such as cgau4 or cgau8 in a complex Gaussian function (cgau).

[0057] According to an embodiment of the present invention, the complex signal is a time-frequency domain signal corresponding to the multi-channel signal to be separated, and the multi-channel signal to be separated can be expanded from the time domain to the time-frequency domain to simultaneously display the time characteristics and frequency characteristics of the signal.

[0058] In operation S120, a plurality of single source points are determined from all time-frequency points corresponding to the complex signal, and the plurality of single source points are processed to determine coefficients of the underdetermined equation group.

[0059] According to an embodiment of the present invention, the time-frequency point represents a combination of a certain time point and a certain frequency point in the time-frequency domain. For example, the time-frequency point A corresponds to the complex signal at 0.35s on the time axis and 5Hz on the frequency axis.

[0060] According to an embodiment of the present invention, a single source point is a time-frequency point obtained from a single signal source, which is distributed in a straight line in a corresponding linear space, wherein each point on the straight line represents a collection of all single source points obtained from a specific signal source.

[0061] According to an embodiment of the present invention, an underdetermined set of equations can be used to separate a target signal from a multi-channel signal to be separated. The coefficients of the underdetermined set of equations are a mixing matrix, and each row in the mixing matrix represents a process of separating a corresponding target signal from a multi-channel signal to be separated, and also represents a process of obtaining a signal to be separated by combining signal sources.

[0062] In operation S130, short-time Fourier transform is performed on the multi-channel signals to be separated to obtain constant terms of the underdetermined equations.

[0063] According to the embodiment of the present invention, the constant term of the underdetermined equation group is accurately estimated through short-time Fourier transform, and the underdetermined equation group can be constructed according to the coefficients of the underdetermined equation group and the constant term of the underdetermined equation group.

[0064] In operation S140, the underdetermined system of equations is solved according to the coefficients of the underdetermined system of equations and the constant term of the underdetermined system of equations to obtain a prediction signal corresponding to the target signal.

[0065] According to an embodiment of the present invention, the root mean square error may be used as an evaluation parameter to evaluate the quality of the prediction signal.

[0066] According to an embodiment of the present invention, a continuous wavelet transform is performed on a multi-channel signal to be separated obtained by linearly combining multiple signal sources, a single source point is screened for the obtained complex signal, the coefficients of the underdetermined system of equations are determined, and then a short-time Fourier transform is performed on the multi-channel signal to be separated to obtain the constant term of the underdetermined system of equations. Since the two time-frequency analysis methods of continuous wavelet transform and short-time Fourier transform are combined to construct an underdetermined linear system of equations and solve the predicted signal, the inverse transform of the continuous wavelet transform is avoided, thereby avoiding the distortion of the predicted signal, and the robustness is high; and since the coefficients are obtained by processing the single source point, there is no need to pre-process the data, so the algorithm cost is low, and the required input data is less, so it can be applied to application scenarios including high noise.

[0067] According to an embodiment of the present invention, a plurality of single source points are determined from all time-frequency points corresponding to a complex signal, including: obtaining characteristic parameters corresponding to all time-frequency points according to the real part and the imaginary part of the complex signal values ​​corresponding to all time-frequency points; and screening out a plurality of single source points from all time-frequency points corresponding to the complex signal according to the characteristic parameters corresponding to all time-frequency points and a predetermined threshold.

[0068] According to an embodiment of the present invention, a multi-channel signal to be separated Perform continuous wavelet transform to obtain complex signal in time-frequency domain , the characteristic parameter Q can be obtained according to the following formula (1).

[0069] (1)

[0070] in, Represents the real part of the complex signal at a certain time-frequency point, Represents the imaginary part of a complex signal at a certain time-frequency point.

[0071] According to an embodiment of the present invention, in order to screen out a single source point, it is necessary to satisfy that for each time-frequency point, the ratio of the real part to the imaginary part of the complex signal is constant, indicating that the phase of the signal received by the sensor is consistent, that is, the following formula (2).

[0072] (2)

[0073] Among them, const is the abbreviation of constant.

[0074] According to an embodiment of the present invention, in actual screening, in order to improve robustness, the characteristic parameters can be screened according to a predetermined threshold, and the predetermined threshold represents a preset maximum error tolerance value, which is used to limit the tolerance to errors. The following formula (3) can be used to screen multiple single source points from all time-frequency points corresponding to the complex signal according to the characteristic parameters and the predetermined threshold corresponding to all time-frequency points.

[0075] (3)

[0076] Where T is the matrix transpose symbol, is a predetermined threshold value, which can be in the range of 0.05 to 0.15.

[0077] According to an embodiment of the present invention, all time-frequency points are screened by using a predetermined threshold and characteristic parameters to determine a single source point, thereby reducing the amount of subsequent data processing and improving robustness to a certain extent.

[0078] According to an embodiment of the present invention, multiple single source points are processed to determine the coefficients of the underdetermined equation group, including: normalizing the multiple single source points according to the corresponding relationship between the multiple single source points and the multiple signal sources to obtain the influence parameters of the multiple signal sources on the multi-channel signals to be separated; and obtaining the coefficients of the underdetermined equation group according to the influence parameters.

[0079] According to an embodiment of the present invention, the normalization process includes a variety of different clustering algorithms, such as: a hierarchical density-based clustering algorithm (HDBSCAN), a K-means clustering algorithm (K-means), and an artificial bee colony algorithm (Artificial Bee Colony algorithm, ABC).

[0080] For example, the influence parameter can be expressed as a vector, corresponding to the influence of a certain signal source on the multi-channel signals to be separated.

[0081] According to an embodiment of the present invention, coefficients of an underdetermined system of equations are obtained based on influencing parameters, including: determining a matrix vector corresponding to the coefficients of the underdetermined system of equations based on the influencing parameters; determining a compensation vector corresponding to the coefficients of the underdetermined system of equations based on a noise signal; and obtaining the coefficients of the underdetermined system of equations based on the matrix vector corresponding to the coefficients of the underdetermined system of equations and the compensation vector.

[0082] According to an embodiment of the present invention, the influencing parameters are arranged to obtain vectors in the mixing matrix. Since the target signal needs to be separated, a column of compensation vectors, such as a column of 1, can be added to the vector formed by the influencing parameters to simplify the calculation.

[0083] For example, the coefficient K of the underdetermined equation system can be expressed as follows (4).

[0084] (4)

[0085] Among them, k 12 ~k mn represents the influencing parameter, m is the number of sensors, and n is the number of signal sources.

[0086] According to an embodiment of the present invention, the underdetermined equation group includes multiple underdetermined equations corresponding to time-frequency points, and a short-time Fourier transform is performed on the multi-channel signals to be separated to obtain constant terms of the underdetermined equation group, including: performing a short-time Fourier transform on the multi-channel signals to be separated to obtain time-frequency domain signals; determining the constant term corresponding to each underdetermined equation based on the time-frequency domain signal value corresponding to each time-frequency point; and obtaining the constant term of the underdetermined equation group based on the constant term corresponding to each underdetermined equation.

[0087] According to an embodiment of the present invention, a multi-channel signal to be separated Perform short-time Fourier transform to obtain the time-frequency domain signal The signal value in the time-frequency domain at each time-frequency point corresponds to a constant term of an underdetermined linear equation system in the underdetermined equation system, so a series of linear equation systems can be obtained based on the constant term and the coefficient K.

[0088] According to an embodiment of the present invention, the underdetermined equation group can be expressed as the following formula (5).

[0089] (5)

[0090] in, is the solution to the underdetermined system of equations.

[0091] According to an embodiment of the present invention, the underdetermined system of equations is solved according to the coefficients of the underdetermined system of equations and the constant terms of the underdetermined system of equations to obtain a prediction signal corresponding to the target signal, including: constructing an underdetermined system of equations according to the coefficients of the underdetermined system of equations and the constant terms of the underdetermined system of equations; solving the underdetermined system of equations through compressed sensing processing to achieve the equivalence of the solution of the underdetermined system of equations to linear programming; obtaining complex data of the prediction signal corresponding to the target signal according to the solution of the underdetermined system of equations; and performing short-time inverse Fourier transform on the complex data to obtain a prediction signal corresponding to the target signal.

[0092] According to an embodiment of the present invention, the solution of the underdetermined system of equations is converted into a linear programming optimization problem through compressed sensing technology. For example, compressed sensing technology including noise processing can be used, specifically including: Total Variation Denoising, Dantzig Selector and Matching Pursuit.

[0093] According to an embodiment of the present invention, a short-time inverse Fourier transform is performed on the complex data. , and obtain the predicted signal corresponding to the target signal .

[0094] According to an embodiment of the present invention, an underdetermined linear equation group is constructed based on two time-frequency analysis methods, namely continuous wavelet transform and short-time Fourier transform, and the complex form of the prediction signal is obtained by solving it. Then, the prediction signal is obtained by inverse short-time Fourier transform, thereby avoiding the inverse transform of the continuous wavelet transform and thus avoiding distortion of the prediction signal, and having higher robustness.

[0095] According to an embodiment of the present invention, the method further comprises: filtering the prediction signal to obtain a separation signal.

[0096] According to the embodiment of the present invention, a simple filtering process can be performed on the prediction signal to obtain a higher quality separation signal.

[0097] To better understand the above The overall process is as follows. Figure 2 A relatively clear explanation is given to the algorithm flow of the above underdetermined blind source separation method.

[0098] Figure 2 A schematic diagram of an algorithm flow framework according to an embodiment of the present invention is shown.

[0099] like Figure 2 As shown, the algorithm input , respectively Continuous wavelet transform and short-time Fourier transform are performed. After continuous wavelet transform, the single source points are screened and normalized to obtain the mixing matrix as the coefficients of the underdetermined equations. Short-time Fourier transform can confirm the constant terms of the underdetermined equations. Compressed sensing technology is used to solve the equations, and the obtained solution is output by short-time Fourier transform. .

[0100] According to an embodiment of the present invention, the root mean square error of the above-mentioned underdetermined blind source separation method based on continuous wavelet transform can be verified by simulating a signal source.

[0101] For example, a simulation data set with a duration of 100 seconds and a sampling rate of 50 Hz can be prepared. A sine signal with a frequency of 0.01 Hz and an amplitude of 100 is used as the target signal base, and an anti-phase sine signal with a frequency of 0.5 Hz and an amplitude of 30 that lasts for one cycle is superimposed at 25 seconds as a mutation signal to obtain the target signal.

[0102] Figure 3 A schematic diagram of a target signal according to an embodiment of the present invention is shown.

[0103] like Figure 3 As shown, the horizontal axis is time and the vertical axis is amplitude.

[0104] According to an embodiment of the present invention, a sine wave signal with an amplitude of 80 and a frequency of 5 Hz, a triangle wave signal with a frequency of 2 Hz, a sine wave signal with a frequency of 0.8 Hz, and a damped oscillation signal with a damping coefficient of 0.01 at 3 Hz can be used as interference signals to randomly generate a noise matrix W as shown in the following formula (6).

[0105] (6)

[0106] Figure 4a A schematic diagram of a sinusoidal interference signal with a frequency of 5 Hz according to an embodiment of the present invention is shown.

[0107] Figure 4b A schematic diagram of a triangular wave interference signal with a frequency of 2 Hz according to an embodiment of the present invention is shown.

[0108] Figure 4c A schematic diagram of a sinusoidal interference signal with a frequency of 0.8 Hz according to an embodiment of the present invention is shown.

[0109] Figure 4d A schematic diagram of a damped oscillation interference signal with a frequency of 3 Hz and a damping coefficient of 0.01 according to an embodiment of the present invention is shown.

[0110] like Figure 4a~Figure 4d As shown, the horizontal axis is amplitude and the vertical axis is time.

[0111] According to an embodiment of the present invention, Gaussian noise with a variance of 20 may be added to each channel to simulate a high noise application scenario. In order to simulate a noisy multi-channel signal to be separated, the target signal may be completely superimposed on the noise matrix W as an algorithm input.

[0112] Figure 5a A schematic diagram showing an algorithm input signal according to an embodiment of the present invention is shown.

[0113] Figure 5b A schematic diagram showing an algorithm output signal according to an embodiment of the present invention is shown.

[0114] like Figure 5a As shown in the figure, the horizontal axis represents time and the vertical axis represents amplitude. From top to bottom, the noisy signals to be separated of the four channels of the algorithm input signal are shown in sequence. According to the image shape, it can be seen that the mutation signal is almost completely submerged in the noise signal. To quantify the noise level, the root mean square error of the algorithm input signal and the target signal amplitude are 31.9679, 27.8871, 24.6261, and 52.0239 respectively.

[0115] like Figure 5b As shown, the horizontal axis represents time and the vertical axis represents amplitude. The algorithm output signals of the four channels are shown from top to bottom.

[0116] According to an embodiment of the present invention, in order to verify the algorithm effect of the combination of continuous wavelet transform and short-time Fourier transform of the embodiment of the present invention, a new data set consisting of the same parameter data set as above, the same noise data and the reverse target signal corresponding to the above target signal can be used for processing to achieve a comparison of the algorithm implementation effects of the related technology and the embodiment of the present invention.

[0117] Figure 6a A schematic diagram of an effect corresponding to a target signal in a related technology according to an embodiment of the present invention is shown.

[0118] Figure 6b A schematic diagram of the effect corresponding to the reverse target signal in the related art according to an embodiment of the present invention is shown.

[0119] like Figure 6a and Figure 6b As shown, the horizontal axis represents time and the vertical axis represents amplitude. The root mean square error of the signal to be separated formed by the target signal is 18.4930, and the root mean square error of the signal to be separated formed by the reverse target signal is 24.5479. It can be seen that the algorithm in the related art is unstable, and the parameters need to be reselected for each slightly changed data set to achieve better results.

[0120] Figure 6c A schematic diagram showing the effect of the underdetermined blind source separation method based on continuous wavelet transform according to an embodiment of the present invention corresponding to the target signal is shown.

[0121] Figure 6d A schematic diagram showing the effect of the underdetermined blind source separation method based on continuous wavelet transform according to an embodiment of the present invention corresponding to the target signal is shown.

[0122] like Figure 6c and 6dAs shown, the horizontal axis represents time and the vertical axis represents amplitude. The root mean square error of the signal to be separated formed by the target signal is 2.1146, and the root mean square error of the signal to be separated formed by the reverse target signal is 1.9299. Compared with the effect of the related art, the error is smaller and there is no need to reselect parameters for each slightly changed data set.

[0123] Figure 7 A block diagram of an underdetermined blind source separation device based on continuous wavelet transform according to an embodiment of the present invention is shown.

[0124] like Figure 7 As shown, the device 700 includes a first transform module 710 , a coefficient determination module 720 , a second transform module 730 and a signal solution module 740 .

[0125] The first transformation module 710 is used to perform continuous wavelet transformation on the multi-channel signal to be separated to obtain a complex signal; wherein the multi-channel signal to be separated is obtained by linearly combining multiple signal sources, and the signal sources include target signals and noise signals.

[0126] The coefficient determination module 720 is used to determine a plurality of single source points from all time-frequency points corresponding to the complex signal, and process the plurality of single source points to determine the coefficients of the underdetermined equation group.

[0127] The second transformation module 730 is used to perform short-time Fourier transform on the multi-channel signals to be separated to obtain constant terms of the underdetermined equations.

[0128] The signal solving module 740 is used to solve the underdetermined equation group according to the coefficients of the underdetermined equation group and the constant term of the underdetermined equation group to obtain a prediction signal corresponding to the target signal.

[0129] According to an embodiment of the present invention, the coefficient determination module 720 includes an obtaining submodule and a screening submodule.

[0130] The obtaining submodule is used to obtain the characteristic parameters corresponding to all the time-frequency points according to the real part and the imaginary part of the complex signal value corresponding to all the time-frequency points.

[0131] The screening submodule is used to screen out a plurality of single source points from all the time-frequency points corresponding to the first time-frequency domain signal according to the characteristic parameters corresponding to all the time-frequency points and a predetermined threshold.

[0132] According to an embodiment of the present invention, the second coefficient determination module 730 includes a processing submodule and an obtaining submodule.

[0133] The processing submodule is used to perform normalization processing on the multiple single source points according to the corresponding relationship between the multiple single source points and the multiple signal sources, so as to obtain the influence parameters of the multiple signal sources on the multi-channel signals to be separated.

[0134] The obtaining submodule is used to obtain the coefficients of the underdetermined equations according to the influencing parameters.

[0135] According to an embodiment of the present invention, the obtaining submodule includes a first determining unit, a second determining unit and a obtaining unit.

[0136] The first determination unit is used to determine the matrix vector corresponding to the coefficients of the underdetermined equation group according to the influencing parameters.

[0137] The second determining unit is used to determine a compensation vector corresponding to the coefficients of the underdetermined equation group based on the noise signal.

[0138] The obtaining unit is used to obtain the coefficients of the underdetermined equation group according to the matrix vector corresponding to the coefficients of the underdetermined equation group and the compensation vector.

[0139] According to an embodiment of the present invention, the first transform module 710 includes a first obtaining submodule, a determining submodule and a second obtaining submodule.

[0140] The first submodule is used to perform short-time Fourier transform on the multi-channel signals to be separated to obtain time-frequency domain signals.

[0141] The determination submodule is used to determine the constant term corresponding to each underdetermined equation according to the time-frequency domain signal value corresponding to each time-frequency point.

[0142] The second obtaining submodule is used to obtain the constant term of the underdetermined equation group according to the constant term corresponding to each underdetermined equation.

[0143] According to an embodiment of the present invention, the signal solving module 740 includes a construction submodule, a processing submodule, an acquisition submodule and a transformation submodule.

[0144] The construction submodule is used to construct the underdetermined equation group according to the coefficients of the underdetermined equation group and the constant term of the underdetermined equation group.

[0145] The processing submodule is used to solve the underdetermined equation group through compressed sensing processing, so as to achieve the equivalent of solving the underdetermined equation group to linear programming.

[0146] The submodule is used to obtain complex data of a prediction signal corresponding to a target signal according to a solution of the underdetermined equation group.

[0147] The transformation submodule is used to perform short-time inverse Fourier transform on the complex data to obtain a prediction signal corresponding to the target signal.

[0148] According to an embodiment of the present invention, the apparatus 700 further includes a filtering module.

[0149] The filtering module is used to filter the prediction signal to obtain a separated signal.

[0150] According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits, or at least part of the functions of any one of them can be implemented in one module. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be split into multiple modules for implementation. According to the embodiments of the present invention, any one or more of the modules, submodules, units, and subunits can be at least partially implemented as hardware circuits, such as field programmable gate arrays (FPGAs), programmable logic arrays (PLAs), systems on chips, systems on substrates, systems on packages, application specific integrated circuits (ASICs), or can be implemented by hardware or firmware in any other reasonable way of integrating or packaging the circuit, or by any one of the three implementation methods of software, hardware, and firmware, or by a proper combination of any of them. Alternatively, according to the embodiments of the present invention, one or more of the modules, submodules, units, and subunits can be at least partially implemented as computer program modules, and when the computer program modules are run, the corresponding functions can be executed.

[0151] For example, any of the first transform module 710, the coefficient determination module 720, the second transform module 730 and the signal solution module 740 can be combined in one module / unit / sub-unit for implementation, or any of the modules / units / sub-units can be split into multiple modules / units / sub-units. Alternatively, at least part of the functions of one or more of these modules / units / sub-units can be combined with at least part of the functions of other modules / units / sub-units and implemented in one module / unit / sub-unit. According to an embodiment of the present invention, at least one of the first transform module 710, the coefficient determination module 720, the second transform module 730 and the signal solution module 740 can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation methods of software, hardware and firmware or in a suitable combination of any of them. Alternatively, at least one of the first transform module 710, the coefficient determination module 720, the second transform module 730 and the signal solution module 740 may be at least partially implemented as a computer program module, which may perform corresponding functions when executed.

[0152] It should be noted that the data processing system part in the embodiment of the present invention corresponds to the data processing method part in the embodiment of the present invention. The description of the data processing system part specifically refers to the data processing method part, which will not be repeated here.

[0153] Figure 8 A block diagram of an electronic device suitable for implementing the above-described underdetermined blind source separation method based on continuous wavelet transform according to an embodiment of the present invention is shown. Figure 8 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0154] like Figure 8 As shown, the electronic device 800 according to an embodiment of the present invention includes a processor 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage part 808 to a random access memory (RAM) 803. The processor 801 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 801 may also include an onboard memory for caching purposes. The processor 801 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present invention.

[0155] In RAM 803, various programs and data required for the operation of electronic device 800 are stored. Processor 801, ROM 802 and RAM 803 are connected to each other via bus 804. Processor 801 performs various operations of the method flow according to the embodiment of the present invention by executing the programs in ROM 802 and / or RAM 803. It should be noted that the program can also be stored in one or more memories other than ROM 802 and RAM 803. Processor 801 can also perform various operations of the method flow according to the embodiment of the present invention by executing the programs stored in the one or more memories.

[0156] According to an embodiment of the present invention, the electronic device 800 may further include an input / output (I / O) interface 808, and the input / output (I / O) interface 805 is also connected to the bus 804. The electronic device 800 may also include one or more of the following components connected to the input / output (I / O) interface 805: an input part 806 including a keyboard, a mouse, etc.; an output part 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage part 808 including a hard disk, etc.; and a communication part 809 including a network interface card such as a LAN card, a modem, etc. The communication part 809 performs communication processing via a network such as the Internet. The drive 810 is also connected to the input / output (I / O) interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed, so that the computer program read therefrom is installed into the storage part 808 as needed.

[0157] According to an embodiment of the present invention, the method flow according to an embodiment of the present invention can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable storage medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 809, and / or installed from the removable medium 811. When the computer program is executed by the processor 801, the above-mentioned functions defined in the system of the embodiment of the present invention are executed. According to an embodiment of the present invention, the system, equipment, device, module, unit, etc. described above can be implemented by a computer program module.

[0158] The present invention also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiment; or may exist independently without being assembled into the device / apparatus / system. The above computer-readable storage medium carries one or more programs, and when the above one or more programs are executed, the method according to the embodiment of the present invention is implemented.

[0159] According to an embodiment of the present invention, the computer-readable storage medium may be a non-volatile computer-readable storage medium. For example, it may include but is not limited to: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present invention, the computer-readable storage medium may be any tangible medium containing or storing a program, which may be used by or in combination with an instruction execution system, an apparatus or a device.

[0160] For example, according to an embodiment of the present invention, the computer-readable storage medium may include the ROM 802 and / or the RAM 803 described above and / or one or more memories other than the ROM 802 and the RAM 803 .

[0161] An embodiment of the present invention also includes a computer program product, which includes a computer program, which contains program code for executing the method provided by the embodiment of the present invention. When the computer program product runs on an electronic device, the program code is used to enable the electronic device to implement the underdetermined blind source separation method based on continuous wavelet transform provided by the embodiment of the present invention.

[0162] When the computer program is executed by the processor 801, the above functions defined in the system / device of the embodiment of the present invention are executed. According to the embodiment of the present invention, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0163] In one embodiment, the computer program may rely on tangible storage media such as optical storage devices, magnetic storage devices, etc. In another embodiment, the computer program may also be transmitted and distributed in the form of signals on a network medium, and downloaded and installed through the communication part 809, and / or installed from a removable medium 811. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to: wireless, wired, etc., or any suitable combination of the above.

[0164] According to an embodiment of the present invention, the program code for executing the computer program provided by the embodiment of the present invention can be written in any combination of one or more programming languages. Specifically, these computing programs can be implemented using high-level process and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, Java, C++, python, "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on the remote computing device, or entirely on the remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., using an Internet service provider to connect through the Internet).

[0165] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. In this regard, each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions. It can be understood by those skilled in the art that the features recorded in the various embodiments of the present invention can be combined and / or combined in various ways, even if such a combination or combination is not explicitly recorded in the present invention. In particular, without departing from the spirit and teachings of the present invention, the features recorded in the various embodiments of the present invention can be combined and / or combined in various ways. All these combinations and / or combinations fall within the scope of the present invention.

[0166] The embodiments of the present invention are described above. However, these embodiments are only for the purpose of illustration, and are not intended to limit the scope of the present invention. Although each embodiment is described above, it does not mean that the measures in each embodiment cannot be used in combination advantageously. Without departing from the scope of the present invention, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present invention.

Claims

1. An underdetermined blind source separation method based on continuous wavelet transform, comprising: Performing continuous wavelet transform on multi-channel signals to be separated to obtain complex signals; wherein the multi-channel signals to be separated are obtained by linearly combining multiple signal sources, and the signal sources include target signals and noise signals; Determining a plurality of single source points from all time-frequency points corresponding to the complex signal, and processing the plurality of single source points to determine coefficients of the underdetermined equation group, including: According to the correspondence between the multiple single source points and the multiple signal sources, the multiple single source points are normalized to obtain the influence parameters of the multiple signal sources on the multi-channel signals to be separated; According to the influencing parameters, the coefficients of the underdetermined equations are obtained, including: Determining a matrix vector corresponding to coefficients of the underdetermined system of equations according to the influencing parameters; Determining, based on the noise signal, a compensation vector corresponding to coefficients of the underdetermined system of equations; Obtaining coefficients of the underdetermined system of equations according to the matrix vector and the compensation vector corresponding to the coefficients of the underdetermined system of equations; Performing short-time Fourier transform on the multi-channel signals to be separated to obtain constant terms of the underdetermined equations; The underdetermined system of equations is solved according to the coefficients of the underdetermined system of equations and the constant term of the underdetermined system of equations to obtain a prediction signal corresponding to the target signal.

2. The method according to claim 1, wherein: The step of determining a plurality of single source points from all time-frequency points corresponding to the complex signal comprises: According to the real part and the imaginary part of the complex signal value corresponding to each of the time-frequency points, the characteristic parameters corresponding to each of the time-frequency points are obtained; According to the characteristic parameters corresponding to all the time-frequency points and the predetermined threshold, a plurality of single source points are screened out from all the time-frequency points corresponding to the complex signal.

3. The method according to claim 1, wherein: The underdetermined equation group includes a plurality of underdetermined equations corresponding to the time-frequency points, and the short-time Fourier transform is performed on the multi-channel signal to be separated to obtain the constant term of the underdetermined equation group, including: Performing short-time Fourier transform on the multi-channel signals to be separated to obtain time-frequency domain signals; Determining a constant term corresponding to each of the underdetermined equations according to a time-frequency domain signal value corresponding to each of the time-frequency points; The constant term of the underdetermined equation group is obtained according to the constant term corresponding to each of the underdetermined equations.

4. The method according to claim 1, wherein: The step of solving the underdetermined set of equations according to the coefficients of the underdetermined set of equations and the constant term of the underdetermined set of equations to obtain a prediction signal corresponding to the target signal includes: constructing the underdetermined system of equations according to the coefficients of the underdetermined system of equations and the constant terms of the underdetermined system of equations; Solving the underdetermined system of equations through compressed sensing processing to achieve the equivalent of solving the underdetermined system of equations to linear programming; Obtaining complex data of the prediction signal corresponding to the target signal according to the solution of the underdetermined equation group; Performing short-time inverse Fourier transform on the complex data to obtain the predicted signal corresponding to the target signal.

5. The method according to claim 1, wherein: Also includes: The prediction signal is filtered to obtain a separation signal.

6. An underdetermined blind source separation device based on continuous wavelet transform, comprising: A first transformation module is used to perform continuous wavelet transformation on the multi-channel signal to be separated to obtain a complex signal; wherein the multi-channel signal to be separated is obtained by linearly combining multiple signal sources, and the signal sources include target signals and noise signals; A coefficient determination module is used to determine multiple single source points from all time-frequency points corresponding to the complex signal, and process the multiple single source points to determine the coefficients of the underdetermined equation group, including: normalizing the multiple single source points according to the corresponding relationship between the multiple single source points and the multiple signal sources to obtain the influence parameters of the multiple signal sources on the multi-channel signals to be separated; obtaining the coefficients of the underdetermined equation group according to the influence parameters, including: determining the matrix vector corresponding to the coefficients of the underdetermined equation group according to the influence parameters; determining the compensation vector corresponding to the coefficients of the underdetermined equation group based on the noise signal; obtaining the coefficients of the underdetermined equation group according to the matrix vector and the compensation vector corresponding to the coefficients of the underdetermined equation group; performing short-time Fourier transform on the multi-channel signals to be separated to obtain the constant terms of the underdetermined equation group; A second transformation module is used to perform short-time Fourier transform on the multi-channel signal to be separated to obtain a constant term of the underdetermined equation group; The signal solving module is used to solve the underdetermined equation group according to the coefficients of the underdetermined equation group and the constant term of the underdetermined equation group to obtain a prediction signal corresponding to the target signal.

7. An electronic device comprising: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 5.

8. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 5.

Images