Life body detection model generation method, life body detection method and related equipment

Abstract

A method for generating a life body detection model, a life body detection method and a related device, the method for generating the life body detection model comprising: acquiring channel state information of a plurality of paths of receiving signals of a Wi-Fi receiving device according to a preset acquisition period, each path of receiving signal comprising a plurality of subcarriers; performing conjugate multiplication processing on channel state information of any two of the multiple paths of receiving signals obtained in each acquisition period to obtain a plurality of corresponding conjugate multiplication result matrixes; respectively extracting amplitude information and phase information of the plurality of subcarriers from the plurality of conjugate multiplication result matrixes; on the basis of the amplitude information and the phase information of the multiple subcarriers, multiple pieces of corresponding life body detection training data are obtained, and a life body detection training data set is formed; and carrying out learning training by adopting the life body detection training data in the life body detection training data set to obtain a corresponding life body detection model. According to the scheme in the embodiment of the invention, the accuracy of life body detection can be improved.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of living body detection, and in particular, to a method for generating a living body detection model, a living body detection method, and related devices. Background Art

[0002] With the development of wireless communication technology, living body detection has been widely applied in fields such as positioning in an environment to be detected, personnel intrusion detection, home medical monitoring, and new human-computer interaction interfaces. Common living body detection methods include those based on vision, sensors, or Wi-Fi, etc.

[0003] The Wi-Fi-based living body detection method has been widely applied due to its non-invasive nature. The Wi-Fi-based living body detection method mainly uses signal features including Received Signal Strength Indicator (RSSI) and Channel State Information (CSI) for living body detection. Compared with the received signal strength indicator, the channel state information contains richer multipath information, which can provide higher accuracy and resolution for living body detection.

[0004] However, the accuracy of the current Wi-Fi-based living body detection method still needs to be improved. Summary of the Invention

[0005] The problem solved by the embodiments of the present invention is to provide a method for generating a living body detection model, a living body detection method, and related devices, which is beneficial to improving the performance of the living body detection model, thereby further improving the accuracy of living body detection.

[0006] To solve the above problems, an embodiment of the present invention provides a method for generating a living body detection model, including:

[0007] Obtain the channel state information of multiple received signals of a Wi-Fi receiving device according to a preset acquisition period, and each received signal includes multiple subcarriers;

[0008] Perform conjugate multiplication processing on the channel state information of any two of the multiple received signals obtained in each acquisition period to obtain corresponding multiple conjugate multiplication result matrices;

[0009] Extract the amplitude information and phase information of the multiple subcarriers from the multiple conjugate multiplication result matrices respectively;

[0010] Based on the amplitude information and phase information of the multiple subcarriers, obtain corresponding multiple pieces of living body detection training data, and form a living body detection training data set;

[0011] Use the life form detection training data in the life form detection training dataset for learning and training to obtain the corresponding life form detection model.

[0012] Optionally, the life form detection training data includes:

[0013] The life form breathing frequency value calculated based on the correlation matrix of the phase difference information of the multiple subcarriers;

[0014] The maximum eigenvalue, the second largest eigenvalue, and the entropy of the eigenvalues after normalization of the correlation coefficient matrix of the amplitude information of the multiple subcarriers;

[0015] The energy value of the first principal component, the energy value of the second principal component, and the energy value of the third principal component obtained by performing principal component analysis on the amplitude information of the multiple subcarriers, the ratio between the variance of the first principal component and the difference value of the eigenvector corresponding to the variance of the first principal component, the ratio between the variance of the second principal component and the difference value of the eigenvector corresponding to the variance of the second principal component, and the ratio between the difference value of the eigenvector corresponding to the variance of the third principal component, and the kurtosis value of the second principal component;

[0016] The maximum eigenvalue and the second largest eigenvalue after normalization of the correlation coefficient matrix of the phase difference information of the multiple subcarriers;

[0017] The standard deviation of the maximum eigenvalue after normalization of the correlation coefficient matrix of the amplitude information of the multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods;

[0018] The standard deviation of the second largest eigenvalue after normalization of the correlation coefficient matrix of the phase difference information of the multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods.

[0019] Optionally, the value range of N is from 5 times to 10 times.

[0020] Optionally, the time length of the acquisition period is from 1 s to 3 s.

[0021] Optionally, the life form detection model includes a support vector machine classifier.

[0022] Optionally, the life form includes at least one of a human body and an animal body with a breathing frequency close to that of a human body.

[0023] Correspondingly, an embodiment of the present invention further provides a generation module for a life form detection model, including:

[0024] A first acquisition sub-module, adapted to acquire the channel state information of the multiple received signals of the Wi-Fi receiving device according to a preset acquisition period, and each received signal includes multiple subcarriers;

[0025] The conjugate multiplication sub-module is adapted to perform conjugate multiplication processing on the channel state information of any two of the multiple received signals obtained in each acquisition period, and obtain a corresponding plurality of conjugate multiplication result matrices;

[0026] The information extraction sub-module is adapted to extract the amplitude information and phase information of the plurality of sub-carriers from the plurality of conjugate multiplication result matrices respectively;

[0027] The data acquisition sub-module is adapted to obtain corresponding multiple pieces of vital sign detection training data based on the amplitude information and phase information of the plurality of sub-carriers, and form a vital sign detection training data set;

[0028] The model training sub-module is adapted to perform learning and training using the vital sign detection training data in the vital sign detection training data set, and obtain a corresponding vital sign detection model.

[0029] Optionally, the vital sign detection training data obtained by the data acquisition sub-module includes:

[0030] The vital sign respiration frequency value calculated based on the phase difference information correlation matrix of the plurality of sub-carriers;

[0031] The maximum eigenvalue, the second largest eigenvalue, and the eigenvalue entropy after normalization of the amplitude information correlation coefficient matrix of the plurality of sub-carriers;

[0032] The energy value of the first principal component, the energy value of the second principal component, and the energy value of the third principal component obtained after performing principal component analysis on the amplitude information of the plurality of sub-carriers, the ratio between the variance of the first principal component and the difference value of the eigenvector corresponding to the variance of the first principal component, the ratio between the variance of the second principal component and the difference value of the eigenvector corresponding to the variance of the second principal component, and the ratio between the difference value of the eigenvector corresponding to the variance of the third principal component, the kurtosis value of the second principal component;

[0033] The maximum eigenvalue and the second largest eigenvalue after normalization of the phase difference information correlation coefficient matrix of the plurality of sub-carriers;

[0034] The standard deviation of the maximum eigenvalue after normalization of the amplitude information correlation coefficient matrix of the plurality of sub-carriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods;

[0035] The standard deviation of the second largest eigenvalue after normalization of the phase difference information correlation coefficient matrix of the plurality of sub-carriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods.

[0036] Optionally, the value range of N is from 5 to 10 times.

[0037] Optionally, the time length of the acquisition period is from 1 s to 3 s.

[0038] Optionally, the living body detection model includes a support vector machine classifier.

[0039] Optionally, the living body includes at least one of a human body and an animal body with a breathing frequency close to that of a human body.

[0040] Correspondingly, an embodiment of the present invention further provides a living body detection method, including:

[0041] Obtaining channel state information of multiple received signals of a Wi-Fi receiving device in a current detection period;

[0042] Inputting the obtained channel state information of multiple received signals of the Wi-Fi receiving device in the current detection period into a living body detection model generated by the living body detection model generation method described in any one of the above, and obtaining a corresponding living body detection result.

[0043] Optionally, the living body detection method further includes:

[0044] Using the living body detection model to obtain the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple subcarriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple subcarriers in the previous (M - 1) detection periods;

[0045] Using the living body detection model to output a detection result that there is no living body in the environment to be detected when it is determined that the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple subcarriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple subcarriers in the previous (M - 1) detection periods are both greater than a preset threshold.

[0046] Optionally, the value range of M is from 5 to 10 times.

[0047] Correspondingly, an embodiment of the present invention further provides a living body detection module, including:

[0048] An acquisition sub-module, adapted to acquire channel state information of multiple received signals of a Wi-Fi receiving device in a current detection period;

[0049] A detection sub-module, adapted to input the obtained channel state information of multiple received signals of the Wi-Fi receiving device in the current detection period into a living body detection model generated by the living body detection model generation method described in any one of the above, and obtain a corresponding living body detection result.

[0050] Optionally, each received signal includes multiple subcarriers;

[0051] The detection sub-module is further adapted to use the living body detection model to obtain the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple sub-carriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple sub-carriers in the previous (M - 1) detection periods; when the living body detection model determines that the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple sub-carriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple sub-carriers in the previous (M - 1) detection periods are both greater than a preset threshold, it outputs a detection result indicating that there is no living body in the environment to be detected.

[0052] Optionally, the value range of M is from 5 times to 10 times.

[0053] Correspondingly, an embodiment of the present invention further provides a chip, on which a generation module of the living body detection model as described in any one of the above or a living body detection module as described in any one of the above is integrated.

[0054] Correspondingly, an embodiment of the present invention further provides an electronic device, including at least one memory and at least one processor, where the memory stores one or more computer instructions, and wherein the one or more computer instructions are executed by the processor to implement the generation method of the living body detection model as described in any one of the above or the living body detection method as described in any one of the above.

[0055] Correspondingly, an embodiment of the present invention further provides a storage medium, which stores one or more computer instructions, and the one or more computer instructions are used to implement the generation method of the living body detection model as described in any one of the above or the living body detection method as described in any one of the above.

[0056] Compared with the prior art, the technical solution of the embodiment of the present invention has the following advantages:

[0057] An embodiment of the present invention provides a method for generating a living body detection model, including: obtaining the channel state information of multiple received signals of a Wi-Fi receiving device according to a preset acquisition period, and each received signal includes multiple sub-carriers; respectively performing conjugate multiplication processing on the channel state information of any two of the multiple received signals obtained in each acquisition period to obtain corresponding multiple conjugate multiplication result matrices; respectively extracting the amplitude information and phase information of the multiple sub-carriers from the multiple conjugate multiplication result matrices; based on the amplitude information and phase information of the multiple sub-carriers, obtaining corresponding multiple pieces of living body detection training data to form a living body detection training data set; using the living body detection training data in the living body detection training data set for learning and training to obtain a corresponding living body detection model.

[0058] In the method for generating a life form detection model according to an embodiment of the present invention, first, conjugate multiplication processing is performed on the channel state information of any two of the multiple received signals of a Wi-Fi receiving device obtained in each acquisition period, which can eliminate the same phase error generated by the channel state information of different received signals. Then, the amplitude information and phase information of the multiple subcarriers are respectively extracted from the multiple conjugate multiplication result matrices, and based on the amplitude information and phase information of the multiple subcarriers, corresponding multiple pieces of life form detection training data are obtained, which can eliminate the change blind areas of the amplitude information or phase information of the movement at certain positions in the environment to be detected, and correspondingly help to improve the accuracy of the life form detection training data, thereby helping to improve the performance of the formed life form detection model, and further helping to improve the accuracy of life form detection. Description of the Drawings

[0059] Figure 1 It is a schematic flowchart of an embodiment of the method for generating a life form detection model provided by the technical solution of the present invention;

[0060] Figure 2 It is a schematic structural diagram of an embodiment of the generation module of the life form detection model provided by the technical solution of the present invention;

[0061] Figure 3 It is a schematic flowchart of an embodiment of the life form detection method provided by the technical solution of the present invention;

[0062] Figure 4 It is a schematic structural diagram of an embodiment of the life form detection module provided by the technical solution of the present invention;

[0063] Figure 5 It is a schematic diagram of an optional hardware structure of an embodiment of an electronic device provided by the technical solution of the present invention. Detailed Embodiments

[0064] Currently, the accuracy of the current Wi-Fi-based life form detection method still needs to be improved.

[0065] To solve the above technical problems, an embodiment of the present invention provides a method for generating a life form detection model, including: obtaining channel state information of multiple received signals of a Wi-Fi receiving device according to a preset acquisition period, where each received signal includes multiple subcarriers; respectively performing conjugate multiplication processing on the channel state information of any two of the multiple received signals obtained in each acquisition period to obtain corresponding multiple conjugate multiplication result matrices; respectively extracting amplitude information and phase information of multiple subcarriers from the multiple conjugate multiplication result matrices; obtaining corresponding multiple life form detection training data based on the amplitude information and phase information of the multiple subcarriers to form a life form detection training data set; and performing learning and training using the life form detection training data in the life form detection training data set to obtain a corresponding life form detection model.

[0066] In the embodiment of the present invention, by performing conjugate multiplication processing on the channel state information of any two of the multiple received signals of the Wi-Fi receiving device obtained in each acquisition period, the same phase error generated by the channel state information of different received signals can be eliminated. Then, amplitude information and phase information of multiple subcarriers are respectively extracted from the multiple conjugate multiplication result matrices, and corresponding multiple life form detection training data are obtained based on the amplitude information and phase information of the multiple subcarriers, which can eliminate the change blind areas of the amplitude information or phase information of the movement at certain positions in the environment to be detected, thereby helping to improve the accuracy of the life form detection training data, and thus helping to improve the performance of the formed life form detection model, and further helping to improve the accuracy of life form detection.

[0067] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of specific embodiments of the present invention will be given with reference to the accompanying drawings.

[0068] Figure 1 The flowchart in an embodiment of the method for generating a life form detection model provided by the technical solution of the present invention is shown. Refer to Figure 1 , a method for generating a life form detection model may specifically include the following steps:

[0069] Step S110: Obtain channel state information of multiple received signals of a Wi-Fi receiving device according to a preset acquisition period, where each received signal includes multiple subcarriers;

[0070] Step S120: Respectively perform conjugate multiplication processing on the channel state information of any two of the multiple received signals obtained in each acquisition period to obtain corresponding multiple conjugate multiplication result matrices;

[0071] Step S130: Respectively extract amplitude information and phase information of multiple subcarriers from the multiple conjugate multiplication result matrices;

[0072] Step S140: Based on the amplitude information and phase information of multiple subcarriers, obtain corresponding multiple pieces of vital sign detection training data, and form a vital sign detection training data set.

[0073] Step S150: Use the vital sign detection training data in the vital sign detection training data set for learning and training to obtain a corresponding vital sign detection model.

[0074] Please continue to refer to Figure 1 , execute Step S110, and obtain the channel state information of multiple received signals of the Wi-Fi receiving device according to a preset acquisition period. Each received signal includes multiple subcarriers.

[0075] Obtain the channel state information of multiple received signals of the Wi-Fi receiving device according to a preset acquisition period. Each received signal includes multiple subcarriers, which provides a basis for subsequently performing conjugate multiplication processing on the channel state information of any two of the multiple received signals obtained in each acquisition period to obtain corresponding multiple conjugate multiplication result matrices.

[0076] In some embodiments, the Wi-Fi transmitting device uses one transmitting antenna to transmit Wi-Fi signals, and the Wi-Fi receiving device uses multiple receiving antennas to simultaneously receive the Wi-Fi signals transmitted by the Wi-Fi transmitting device.

[0077] In this embodiment, the Wi-Fi transmitting device uses one transmitting antenna to transmit Wi-Fi signals, and the Wi-Fi receiving device uses two receiving antennas to simultaneously receive the Wi-Fi signals transmitted by the Wi-Fi transmitting device.

[0078] Among them, the Wi-Fi signal is a signal frequency response (Channel Frequency Response, CFR) data packet. Correspondingly, in this embodiment, the Wi-Fi transmitting device uses one transmitting antenna to transmit a frequency response data packet, and the Wi-Fi receiving device uses two receiving antennas to simultaneously receive the frequency response data packet transmitted by the Wi-Fi transmitting device.

[0079] In this embodiment, each received signal includes multiple subcarriers. Correspondingly, the channel state information of each received signal, that is, the channel state information of multiple subcarriers of each receiving antenna of the Wi-Fi receiving device.

[0080] The channel state information of the multi-path received signals of a Wi-Fi receiving device reflects the amplitude and phase information of multiple sub-carriers of each receiving antenna of the Wi-Fi receiving device. Among them, the position where a living body is located, the moving direction of the living body, etc. will cause changes in the amplitude and phase of multiple sub-carriers of each receiving antenna of the Wi-Fi receiving device. Therefore, obtaining the channel state information of the multi-path received signals of the Wi-Fi receiving device according to a preset acquisition period can achieve the detection of a living body.

[0081] The acquisition period of the channel state information of the multi-path received signals of the Wi-Fi receiving device can be set by those skilled in the art according to the acquisition requirements of the channel state information of the multi-path received signals of the Wi-Fi receiving device.

[0082] It can be understood that the acquisition period of the channel state information of the multi-path received signals of the Wi-Fi receiving device should not be too long and should not be too short either. If the acquisition period of the channel state information of the multi-path received signals of the Wi-Fi receiving device is too long, the real-time performance of the channel state information of the multi-path received signals of the Wi-Fi receiving device in the time dimension obtained accordingly will be reduced, thereby reducing the real-time performance of subsequent data processing, and the subsequent computing amount will also be increased; if the acquisition time of the channel state information of the multi-path received signals of the Wi-Fi receiving device is too short, the complete channel state information of the multi-path received signals of the Wi-Fi receiving device in the time dimension cannot be obtained, and the accuracy of the data obtained subsequently will be reduced accordingly, and further the accuracy of the generated living body detection model will be affected. For this reason, in this embodiment, the time length of the acquisition period is 1 s to 3 s.

[0083] Please continue to refer to Figure 1 , perform step S120, and perform conjugate multiplication processing on the channel state information of any two of the multi-path received signals obtained in each acquisition period to obtain corresponding multiple conjugate multiplication result matrices.

[0084] Performing conjugate multiplication processing on the channel state information of any two of the multi-path received signals obtained in each acquisition period to obtain corresponding multiple conjugate multiplication result matrices provides a basis for subsequently extracting the amplitude information and phase information of multiple sub-carriers from the multiple conjugate multiplication result matrices respectively.

[0085] Performing conjugate multiplication on the channel state information of any two of the multiple received signals of the Wi-Fi receiving device obtained in each acquisition period can eliminate the same random phase error generated by the channel state information of multiple subcarriers of different receiving antennas, thereby improving the accuracy of the channel state information of multiple subcarriers of the receiving antennas of the Wi-Fi receiving device. Subsequently, extracting the amplitude information and phase information of multiple subcarriers from multiple conjugate multiplication result matrices respectively can improve the accuracy of the obtained amplitude information and phase information of multiple subcarriers.

[0086] When the Wi-Fi transmitting device uses P transmitting antennas to transmit channel frequency response data packets, the Wi-Fi receiving device uses Q receiving antennas to simultaneously receive the channel frequency response data packets transmitted by the Wi-Fi transmitting device, and the received signal of each receiving antenna includes S subcarriers, the channel frequency response matrix composed of L channel frequency response data packets transmitted within a preset acquisition period can be expressed as:

[0087]

[0088] where H represents the channel frequency response matrix, and H k represents the kth channel frequency response data packet, and k is an integer greater than or equal to 1 and less than or equal to L.

[0089] Correspondingly, each channel frequency response data packet sent by each transmitting antenna of the Wi-Fi transmitting device to each receiving antenna of the Wi-Fi receiving device can be expressed as:

[0090]

[0091] where, represents the tth channel frequency response data packet sent by the ith transmitting antenna of the Wi-Fi transmitting device to the jth receiving antenna of the Wi-Fi receiving device, represents the channel frequency response data on the sth subcarrier in the jth receiving antenna of the Wi-Fi receiving device in the tth channel frequency response data packet sent by the ith transmitting antenna of the Wi-Fi transmitting device to the jth receiving antenna of the Wi-Fi receiving device, and s is an integer greater than or equal to 1 and less than or equal to S.

[0092] Correspondingly, the tth channel frequency response data packet sent by P transmitting antennas of the Wi-Fi transmitting device to Q receiving antennas of the Wi-Fi receiving device can be expressed as:

[0093]

[0094] where H tIt represents the t-th channel frequency response data matrix sent by the P root transmitting antennas of the Wi-Fi transmitting device to the Q root receiving antennas of the Wi-Fi receiving device.

[0095] Correspondingly, the corresponding multiple conjugate multiplication result matrices obtained by performing conjugate multiplication processing on the channel state information of any two of the multiple received signals obtained in each acquisition period can be expressed as:

[0096]

[0097] where conj(.) represents the conjugate operation.

[0098] In this embodiment, the Wi-Fi transmitting device uses one transmitting antenna to transmit Wi-Fi signals, and the Wi-Fi receiving device uses two receiving antennas to simultaneously receive the Wi-Fi signals transmitted by the Wi-Fi transmitting device. Correspondingly, the conjugate multiplication result matrix obtained by performing conjugate multiplication processing on the channel state information of the two received signals obtained in each acquisition period can be expressed as:

[0099]

[0100] where H cm represents the conjugate multiplication result matrix obtained by performing conjugate multiplication processing on the channel state information of the two received signals obtained in each acquisition period.

[0101] Please continue to refer to Figure 1 , and execute step S130 to extract the amplitude information and phase information of multiple subcarriers from multiple conjugate multiplication result matrices respectively.

[0102] Extracting the amplitude information and phase information of multiple subcarriers from multiple conjugate multiplication result matrices respectively provides a basis for subsequently obtaining corresponding multiple sets of vital sign detection training data based on the amplitude information and phase information of multiple subcarriers and forming a vital sign detection training data set.

[0103] The step of extracting the amplitude information of multiple subcarriers from multiple conjugate multiplication result matrices respectively includes: obtaining the amplitude information correlation coefficient matrices of multiple subcarriers respectively based on the conjugate multiplication result matrices.

[0104] In this embodiment, the Wi-Fi transmitting device uses one transmitting antenna to transmit Wi-Fi signals, and the Wi-Fi receiving device uses two receiving antennas to simultaneously receive the Wi-Fi signals transmitted by the Wi-Fi transmitting device. Correspondingly, the step of extracting the amplitude information of multiple subcarriers from multiple conjugate multiplication result matrices respectively includes: obtaining the amplitude information correlation coefficient matrices of multiple subcarriers respectively based on the conjugate multiplication result matrices.

[0105] Specifically, the correlation coefficient matrix of the amplitude information of multiple subcarriers can be calculated using the following formula:

[0106] R am =corr(||H cm ||,||H cm T ||) (6)

[0107] where R am represents the correlation coefficient matrix of the amplitude information of multiple subcarriers, corr(.) represents the operation of calculating the correlation coefficient, ||H cm || represents the operation of calculating the amplitude of the conjugate multiplication result matrix H cm || represents the operation of calculating the amplitude of the transposed matrix H cm T of the conjugate multiplication result matrix H cm T .

[0108] Correspondingly, the steps of extracting the phase information of multiple subcarriers from multiple conjugate multiplication result matrices respectively include: obtaining the correlation coefficient matrix and the correlation matrix of the phase difference information of multiple subcarriers respectively based on the conjugate multiplication result matrices.

[0109] Specifically, the correlation coefficient matrix of the phase difference information of multiple subcarriers is calculated using the following formula:

[0110] R ph =corr(∠H cm ,∠H cm T ) (7)

[0111] where R ph represents the correlation coefficient matrix of the phase difference information of multiple subcarriers, ∠H cm represents the operation of taking the phase of the conjugate multiplication result matrix H cm ∠H cm T represents the operation of taking the phase of the transposed matrix H cm of the conjugate multiplication result matrix H cm T .

[0112] Please continue to refer to Figure 1 , perform step S140, and obtain corresponding multiple pieces of life detection training data based on the amplitude information and phase information of multiple subcarriers to form a life detection training data set.

[0113] Based on the amplitude information and phase information of multiple subcarriers, multiple corresponding pieces of vital sign detection training data are obtained to form a vital sign detection training data set, providing a basis for subsequent learning and training using the vital sign detection training data in the vital sign detection training data set to obtain the corresponding vital sign detection model.

[0114] In this embodiment, the vital sign detection training data includes 15 features in the following 6 aspects:

[0115] (1) The vital sign breathing frequency value calculated based on the correlation matrix of the phase difference information of multiple subcarriers;

[0116] (2) The maximum eigenvalue, the second largest eigenvalue, and the entropy of the eigenvalues after normalization of the correlation coefficient matrix of the amplitude information of multiple subcarriers.

[0117] (3) The maximum eigenvalue and the second largest eigenvalue after normalization of the correlation coefficient matrix of the phase difference information of multiple subcarriers;

[0118] (4) The energy value of the first principal component, the energy value of the second principal component, and the energy value of the third principal component obtained by performing principal component analysis on the amplitude information of multiple subcarriers, the ratio between the variance of the first principal component and the difference value of the eigenvector corresponding to the variance of the first principal component, the ratio between the variance of the second principal component and the difference value of the eigenvector corresponding to the variance of the second principal component, the ratio between the variance of the third principal component and the difference value of the eigenvector corresponding to the variance of the third principal component, and the kurtosis value of the second principal component;

[0119] (5) The standard deviation of the maximum eigenvalue after normalization of the correlation coefficient matrix of the amplitude information of multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods;

[0120] (6) The standard deviation of the second largest eigenvalue after normalization of the correlation coefficient matrix of the phase difference information of multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods.

[0121] In specific implementation, vital signs with different levels of movement will cause different changes between different subcarriers of each receiving antenna of the Wi-Fi receiving device, thereby affecting the maximum eigenvalue, the second largest eigenvalue, and the entropy of the eigenvalues of the correlation coefficient matrix of the amplitude information of multiple subcarriers of each receiving antenna of the Wi-Fi receiving device, and simultaneously affecting the maximum eigenvalue and the second largest eigenvalue of the correlation coefficient matrix of the phase difference information of multiple subcarriers of each receiving antenna of the Wi-Fi receiving device, as well as the standard deviation of the maximum eigenvalue after normalization of the correlation coefficient matrix of the amplitude information of multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods, and the standard deviation of the second largest eigenvalue after normalization of the correlation coefficient matrix of the amplitude information of multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods.

[0122] Therefore, in this embodiment, the maximum eigenvalue, the second largest eigenvalue, and the eigenvalue entropy of the normalized correlation coefficient matrix of the amplitude information of multiple subcarriers, the maximum eigenvalue and the second largest eigenvalue of the normalized correlation coefficient matrix of the phase difference information of multiple subcarriers, and the standard deviation of the maximum eigenvalue of the normalized correlation coefficient matrix of the amplitude information of multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods, and the standard deviation of the second largest eigenvalue of the normalized correlation coefficient matrix of the amplitude information of multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods are used as the representations of the dynamic and static characteristics of the living body.

[0123] In this embodiment, the maximum eigenvalue, the second largest eigenvalue, and the eigenvalue entropy of the normalized correlation coefficient matrix of the amplitude information of multiple subcarriers can be respectively expressed as:

[0124] e amm = max(eigen(R am ) / L) (8)

[0125] e ams = smax(eigen(R am ) / L) (9)

[0126] E entropy = -∑P i log(P i ) (10)

[0127] Wherein, e amm represents the maximum eigenvalue of the normalized correlation coefficient matrix of the amplitude information of multiple subcarriers, e ams represents the second largest eigenvalue of the normalized correlation coefficient matrix of the amplitude information of multiple subcarriers, eigen(.) represents the operation of obtaining eigenvalues, max(.) represents the operation of taking the maximum value, smax(.) represents the operation of taking the second largest value, E entropy represents the eigenvalue entropy of the normalized correlation coefficient matrix of the amplitude information of multiple subcarriers, P i represents the probability of the i-th eigenvalue of the normalized correlation coefficient matrix of the amplitude information of multiple subcarriers, and log(.) represents the operation of taking the logarithm.

[0128] In this embodiment, the maximum eigenvalue and the second largest eigenvalue of the normalized correlation coefficient matrix of the phase difference information of multiple subcarriers can be respectively expressed as:

[0129] e phm = max(eigen(R ph ) / L) (11)

[0130] e phs= smax(eigen(R ph ) / L) (12)

[0131] where e phm represents the maximum eigenvalue after normalization of the correlation coefficient matrix of the phase difference information of multiple subcarriers, and e phs represents the second largest eigenvalue after normalization of the correlation coefficient matrix of the phase difference information of multiple subcarriers.

[0132] According to actual needs, the value of N can be obtained by those skilled in the art according to the training requirements of the life detection model. As an example, the value range of N is from 5 times to 10 times.

[0133] In a specific implementation, different degrees of movement and stillness of the living beings located around the detection environment, or the different degrees of movement and stillness of the living beings passing through the wall, will cause changes in the first principal component, the second principal component, and the third principal component obtained by performing principal component analysis on the amplitude information of multiple subcarriers.

[0134] Correspondingly, the energy value of the first principal component, the energy value of the second principal component, and the energy value of the third principal component, the ratio between the variance of the first principal component and the difference value of the eigenvector corresponding to the variance of the first principal component, the ratio between the variance of the second principal component and the difference value of the eigenvector corresponding to the variance of the second principal component, and the ratio between the variance of the third principal component and the difference value of the eigenvector corresponding to the variance of the third principal component, and the kurtosis value of the second principal component are used as the characterization of the dynamic and static characteristics of the living beings passing through the wall.

[0135] At the same time, performing principal component analysis on the amplitude information of multiple subcarriers can reduce the dimension of the amplitude information of multiple subcarriers and can filter out low-frequency noise to a certain extent.

[0136] In this embodiment, the energy value of the i-th principal component obtained by performing principal component analysis on the amplitude information of multiple subcarriers can be calculated respectively by using the following formula:

[0137]

[0138] And:

[0139]

[0140] c i = He i (15)

[0141] E = eigen(C) = (e 0 … e S ) (16)

[0142]

[0143] Among them, g i represents the energy value of the i-th principal component obtained by performing principal component analysis on the amplitude information of multiple subcarriers, and FFT(k) represents the value of the k-th frequency component after performing Fourier transform on the i-th principal component, c i represents the i-th principal component obtained by performing principal component analysis on the amplitude information of multiple subcarriers, represents the conjugate multiplication result matrix H cm The amplitude matrix after removing the DC component.

[0144] In this embodiment, the following formula is used to calculate the ratio between the variance of the i-th principal component and the difference value of the eigenvector corresponding to the variance of the i-th principal component:

[0145]

[0146] And:

[0147]

[0148] Among them, represents the ratio between the variance of the i-th principal component and the difference value of the eigenvector corresponding to the variance of the i-th principal component, var(c i ) represents the variance operation performed on the i-th principal component, represents the difference value of the eigenvector corresponding to the variance of the i-th principal component.

[0149] In this embodiment, the kurtosis value of the second principal component is calculated using the following formula:

[0150]

[0151] Among them, z represents the kurtosis value of the second principal component, c 2 represents the second principal component, represents the mean value of the second principal component.

[0152] It should be noted that when there is a stationary human body in the detection environment, the dynamic and static characteristics in the detection environment are very similar to the dynamic and static characteristics of passing through the wall. At this time, the vital body respiration frequency value calculated based on the correlation matrix of the phase difference information of multiple subcarriers can be used to distinguish the environment with a stationary vital body from the environment without people.

[0153] In this embodiment, based on the correlation matrix of the phase difference information of multiple subcarriers, the Multiple Signal Classification (MUSIC) algorithm is used to calculate the vital body respiration frequency value.

[0154] In this embodiment, the following formula is used to calculate the correlation matrix of the phase difference information of multiple subcarriers:

[0155]

[0156] where R represents the correlation matrix of the phase difference information of multiple subcarriers, represents the conjugate multiplication result matrix H cm phase matrix after removing the DC component, represents the conjugate multiplication result matrix H cm phase matrix after removing the DC component transpose matrix of.

[0157] In other embodiments, based on the correlation matrix of the phase difference information of multiple subcarriers, other super-resolution algorithms can also be used to calculate the respiration frequency value of a living body. Among them, other super-resolution algorithms include the root multiple classification (root_MUSIC) algorithm, the minimum variance distortionless response (MVDR) algorithm, and the estimation of signal parameters using rotational invariance techniques (ESPRIT) algorithm, etc.

[0158] When calculating the respiration frequency value of a living body, the sliding step length is the acquisition period. In other words, the respiration frequency value of a living body is calculated once for each acquisition period, and the time length of the sliding time window when calculating the respiration frequency value of a living body is related to the respiration frequency of the living body. Specifically, the time length of the sliding time window should be greater than or equal to the time length of one breath of the living body and be an integer multiple of the time length of one breath of the living body.

[0159] In this embodiment, in order to simultaneously meet the requirements of Wi-Fi device communication and living body detection, the living body detection model is used for human detection. The respiration frequency of a human body is generally 0.1 Hz to 0.6 Hz. Therefore, in this embodiment, the time length of the sliding time window is 1.7 s to 10 s.

[0160] In other embodiments, the living body detection model can also be used for human detection of other animal bodies with a respiration frequency similar to that of a human body. Among them, other animal bodies with a respiration frequency similar to that of a human body include cats, dogs, rabbits, pigs, horses, cows, and sheep, etc.

[0161] Please continue to refer to Figure 1, step S150 is executed to perform learning and training using the life form detection training data in the life form detection training dataset, and obtain the corresponding life form detection model.

[0162] Perform learning and training using the life form detection training data in the life form detection training dataset, and obtain the corresponding life form detection model, so that the life form detection model can be used for life form detection.

[0163] In this embodiment, perform learning and training using the human detection training data in the human detection training dataset, and obtain the corresponding human detection model.

[0164] In some embodiments, the life form detection model includes a support vector machine (SVM) classifier. Correspondingly, use the support vector machine algorithm to perform learning and training on the life form detection training data in the life form detection training dataset, and obtain the corresponding life form detection model.

[0165] In other embodiments, the life form detection model can also be other classifiers, such as an Adaptive Boosting (Adaboost) classifier, a Bayes classifier, a BackPropagation (BP) neural network classifier, etc., or can also be other feature classification models, which are not limited here.

[0166] Correspondingly, an embodiment of the present invention also provides a generation module for a life form detection model.

[0167] Figure 2 The structural schematic diagram of an embodiment of the generation module for the life form detection model provided by the technical solution of the present invention is shown. Refer to Figure 2 , a generation module 200 for a life form detection model, includes: a first acquisition sub-module 201, adapted to acquire the channel state information of multiple received signals of a Wi-Fi receiving device according to a preset acquisition period, and each received signal includes multiple subcarriers; a conjugate multiplication sub-module 202, adapted to perform conjugate multiplication processing on the channel state information of any two of the multiple received signals acquired in each acquisition period, and obtain the corresponding multiple conjugate multiplication result matrices; an information extraction sub-module 203, adapted to extract the amplitude information and phase information of multiple subcarriers from the multiple conjugate multiplication result matrices respectively; a data acquisition sub-module 204, adapted to obtain the corresponding multiple life form detection training data based on the amplitude information and phase information of the multiple subcarriers, and form a life form detection training dataset; a model training sub-module 205, adapted to perform learning and training using the life form detection training data in the life form detection training dataset, and obtain the corresponding life form detection model.

[0168] The generation module of the life form detection model in the embodiments of the present invention can be used to execute the aforementioned method for generating the life form detection model, or other functional modules can also be used to execute the aforementioned method for generating the life form detection model. For the method for generating the life form detection model, please refer to the detailed description in the foregoing part and will not be elaborated herein.

[0169] Correspondingly, the embodiments of the present invention further provide a life form detection method.

[0170] Figure 3 The flowchart of an embodiment of the life form detection method provided by the technical solution of the present invention is shown. Refer to Figure 3 A life form detection method may specifically include the following steps:

[0171] Step S310: Obtain the channel state information of multiple received signals of the Wi-Fi receiving device in the current detection period;

[0172] Step S320: Input the obtained channel state information of multiple received signals of the Wi-Fi receiving device in the current detection period into the life form detection model generated by the method for generating the life form detection model as described above, and obtain the corresponding life form detection result.

[0173] In some embodiments, input the obtained channel state information of multiple received signals of the Wi-Fi receiving device in the current detection period into the life form detection model generated by the method for generating the life form detection model as described above, so that the life form detection model can extract 15 features in 6 aspects mentioned in step S140, and determine whether there is a life form in the environment to be detected according to the extracted features.

[0174] In this embodiment, the life form detection model is a human body detection model. Correspondingly, input the obtained channel state information of multiple received signals of the Wi-Fi receiving device in the current detection period into the human body detection model, so that the human body detection model can extract 15 features in 6 aspects mentioned in step S140, and determine whether there is a human body in the environment to be detected according to the extracted features.

[0175] In other embodiments, the life form detection model can also be used to detect other animal bodies with a breathing frequency similar to that of a human body.

[0176] In this embodiment, the life form detection method further includes:

[0177] Step S330: Use the life form detection model to obtain the entropy of the eigenvalues after normalization of the correlation coefficient matrix of the amplitude information of multiple subcarriers in the current detection period and the entropy of the eigenvalues after normalization of the correlation coefficient matrix of the amplitude information of multiple subcarriers in the previous (M - 1) detection periods;

[0178] Step S340: When it is determined by the living body detection model that the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple subcarriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple subcarriers in the previous (M - 1) detection periods are both greater than a preset threshold, an inspection result that there is no living body in the environment to be detected is output.

[0179] When it is determined by the living body detection model that the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple subcarriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of multiple subcarriers in the previous (M - 1) detection periods are both greater than a preset threshold, and an inspection result that there is no living body in the environment to be detected is output, the influence of the living bodies of neighbors on the living body detection of the environment to be detected can be avoided, false alarms can be prevented, and correspondingly, it is beneficial to further improve the accuracy of living body detection.

[0180] The value of M can be set according to the needs of living body detection. As an example, the value range of M is from 5 times to 10 times.

[0181] Correspondingly, an embodiment of the present invention further provides a living body detection module.

[0182] Figure 4 The structural schematic diagram of an embodiment of the living body detection module provided by the technical solution of the present invention is shown. Refer to Figure 4 , a living body detection module 400 includes: an acquisition sub-module 401, adapted to acquire the channel state information of multiple received signals of a Wi-Fi receiving device in the current detection period; a detection sub-module 402, adapted to input the acquired channel state information of multiple received signals of the Wi-Fi receiving device in the current detection period into a living body detection model generated by the generation method of the living body detection model as described above, and obtain a corresponding living body detection result.

[0183] The living body detection module in the embodiment of the present invention can be used to execute the foregoing living body detection method, or other functional modules can also be used to execute the foregoing living body detection method. For the living body detection method, please refer to the detailed description in the foregoing part, and details will not be repeated here.

[0184] Correspondingly, an embodiment of the present invention further provides a chip, on which a generation module or a living body detection module of the living body detection model as described in the embodiment of the present invention is integrated. Among them, for the generation module of the living body detection model or the living body detection module, please refer to the detailed description in the foregoing part, and details will not be repeated here.

[0185] Correspondingly, an embodiment of the present invention further provides a storage medium storing one or more computer instructions for implementing the method for generating a living body detection model or the living body detection method as described in the embodiment of the present invention. For the method for generating the living body detection model or the living body detection method, please refer to the detailed description in the foregoing part, which will not be elaborated herein.

[0186] Correspondingly, an embodiment of the present invention further provides an electronic device including at least one memory and at least one processor, where the memory stores one or more computer instructions, and the one or more computer instructions are executed by the processor to implement the method for generating a living body detection model or the living body detection method as described in the embodiment of the present invention. For the method for generating the living body detection model or the living body detection method, please refer to the detailed description in the foregoing part, which will not be elaborated herein.

[0187] An optional hardware structure of the electronic device provided by the embodiment of the present invention may be as Figure 5 shown, including: at least one processor 01, at least one communication interface 02, at least one memory 03, and at least one communication bus 04.

[0188] In the embodiment of the present invention, the number of the processor 01, the communication interface 02, the memory 03, and the communication bus 04 is at least one, and the processor 01, the communication interface 02, and the memory 03 complete communication with each other through the communication bus 04.

[0189] The communication interface 02 may be an interface of a communication module for network communication, such as an interface of a GSM module.

[0190] The processor 01 may be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiment of the present invention.

[0191] The memory 03 may include a high-speed RAM memory, and may also include a non-volatile memory, for example, at least one disk memory.

[0192] Wherein, the memory 03 stores one or more computer instructions, and the one or more computer instructions are executed by the processor 01 to implement the method for generating a living body detection model or the living body detection method as described in the embodiment of the present invention.

[0193] It should be noted that the above-mentioned implementation terminal device may further include other devices (not shown) that may not be essential to the disclosed content of the embodiments of the present invention; since these other devices may not be essential for understanding the disclosed content of the embodiments of the present invention, the embodiments of the present invention will not introduce them one by one.

[0194] The embodiments of the present invention further provide a storage medium, which stores one or more computer instructions for implementing the method for generating a living body detection model or the living body detection method according to the embodiments of the present invention.

[0195] The above embodiments of the present invention are combinations of elements and features of the present invention. Unless otherwise mentioned, the elements or features can be regarded as selective. Each element or feature can be practiced without being combined with other elements or features. In addition, the embodiments of the present invention can be constructed by combining some elements and / or features. The operation sequences described in the embodiments of the present invention can be rearranged. Some configurations of any embodiment can be included in another embodiment and replaced by the corresponding configuration of another embodiment. It is obvious to those skilled in the art that the claims that do not have an explicit citation relationship with each other in the appended claims can be combined into the embodiments of the present invention, or can be included as new claims in the amendments after the submission of this application.

[0196] The embodiments of the present invention can be implemented by various means such as, by way of example, hardware, firmware, software, or a combination thereof. In the hardware configuration mode, the method according to the exemplary embodiments of the present invention can be implemented by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, etc.

[0197] In the firmware or software configuration mode, the embodiments of the present invention can be implemented in the form of modules, processes, functions, etc. The software code can be stored in the memory unit and executed by the processor. The memory unit is located inside or outside the processor and can send data to the processor and receive data from the processor via various known means.

[0198] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present invention. Various modifications to these embodiments will be obvious to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to these embodiments shown herein, but will be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0199] Although the present invention is disclosed as above, the present invention is not limited thereto. Any person skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the scope defined by the claims.

Claims

1. A method for generating a living body detection model, characterized in that, comprising: Obtaining the channel state information of multiple received signals of a Wi-Fi receiving device according to a preset acquisition period, and each received signal includes multiple subcarriers; Performing conjugate multiplication processing on the channel state information of any two of the multiple received signals obtained in each acquisition period respectively to obtain corresponding multiple conjugate multiplication result matrices; Extracting the amplitude information and phase information of the multiple subcarriers from the multiple conjugate multiplication result matrices respectively; Based on the amplitude information and phase information of the multiple subcarriers, obtaining corresponding multiple pieces of living body detection training data to form a living body detection training data set; Using the living body detection training data in the living body detection training data set for learning and training to obtain a corresponding living body detection model.

2. The method for generating a living body detection model according to claim 1, characterized in that, the living body detection training data includes: The living body breathing frequency value calculated based on the phase difference information correlation matrix of the multiple subcarriers; The maximum eigenvalue, the second largest eigenvalue and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of the multiple subcarriers; The energy value of the first principal component, the energy value of the second principal component and the energy value of the third principal component obtained by performing principal component analysis on the amplitude information of the multiple subcarriers, the ratio between the variance of the first principal component and the difference value of the eigenvector corresponding to the variance of the first principal component, the ratio between the variance of the second principal component and the difference value of the eigenvector corresponding to the variance of the second principal component, the ratio between the difference value of the eigenvector corresponding to the variance of the third principal component, and the kurtosis value of the second principal component; The maximum eigenvalue and the second largest eigenvalue after normalization of the phase difference information correlation coefficient matrix of the multiple subcarriers; The standard deviation of the maximum eigenvalue after normalization of the amplitude information correlation coefficient matrix of the multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods; The standard deviation of the second largest eigenvalue after normalization of the phase difference information correlation coefficient matrix of the multiple subcarriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods.

3. The method for generating a living body detection model according to claim 2, characterized in that, the value range of N is from 5 times to 10 times.

4. The method for generating a living body detection model according to claim 1, characterized in that, the time length of the acquisition period is from 1 s to 3 s.

5. The method for generating a living body detection model according to claim 1, characterized in that, the living body detection model includes a support vector machine classifier.

6. The method for generating a living body detection model according to any one of claims 1 to 5, characterized in that, the living body includes at least one of a human body and an animal body with a breathing frequency close to that of a human body.

7. A generating module for a living body detection model, characterized in that, comprising: A first acquisition sub-module adapted to obtain the channel state information of multiple received signals of a Wi-Fi receiving device according to a preset acquisition period, and each received signal includes multiple subcarriers; The conjugate multiplication sub-module is adapted to perform conjugate multiplication processing on the channel state information of any two of the multiple received signals obtained in each acquisition period respectively, and obtain corresponding multiple conjugate multiplication result matrices; The information extraction sub-module is adapted to extract the amplitude information and phase information of the multiple sub-carriers from the multiple conjugate multiplication result matrices respectively; The data acquisition sub-module is adapted to obtain corresponding multiple pieces of vital sign detection training data based on the amplitude information and phase information of the multiple sub-carriers, and form a vital sign detection training data set; The model training sub-module is adapted to perform learning and training using the vital sign detection training data in the vital sign detection training data set, and obtain a corresponding vital sign detection model.

8. The vital sign detection model generation module according to claim 7, wherein, The vital sign detection training data obtained by the data acquisition sub-module includes: The vital sign respiration frequency value calculated based on the phase difference information correlation matrix of the multiple sub-carriers; The maximum eigenvalue, the second largest eigenvalue and the eigenvalue entropy after normalization of the amplitude information correlation coefficient matrix of the multiple sub-carriers; The energy value of the first principal component, the energy value of the second principal component and the energy value of the third principal component obtained by performing principal component analysis on the amplitude information of the multiple sub-carriers, the ratio between the variance of the first principal component and the difference value of the eigenvector corresponding to the variance of the first principal component, the ratio between the variance of the second principal component and the difference value of the eigenvector corresponding to the variance of the second principal component, the ratio between the variance of the third principal component and the difference value of the eigenvector corresponding to the variance of the third principal component, and the kurtosis value of the second principal component; The maximum eigenvalue and the second largest eigenvalue after normalization of the phase difference information correlation coefficient matrix of the multiple sub-carriers; The standard deviation of the maximum eigenvalue of the amplitude information correlation coefficient matrix of the multiple sub-carriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods; The standard deviation of the second largest eigenvalue of the phase difference information correlation coefficient matrix of the multiple sub-carriers corresponding to the current acquisition period and the previous (N - 1) acquisition periods.

9. The vital sign detection model generation module according to claim 7, wherein, The value range of N is from 5 times to 10 times.

10. The vital sign detection model generation module according to claim 6, wherein, The time length of the acquisition period is from 1 s to 3 s.

11. The vital sign detection model generation module according to claim 6, wherein, The vital sign detection model includes a support vector machine classifier.

12. The vital sign detection model generation module according to any one of claims 7 to 11, wherein, The vital sign includes at least one of a human body and an animal body with a respiration frequency close to that of a human body.

13. A vital sign detection method, wherein, it includes: Obtain the channel state information of the multiple received signals of the Wi-Fi receiving device in the current detection period; Input the channel state information of the multi-path received signals of the Wi-Fi receiving device in the current detection period into the life form detection model generated by the method for generating a life form detection model according to any one of claims 1 to 6, and obtain the corresponding life form detection result.

14. The life form detection method according to claim 13, wherein, each of the received signals includes a plurality of subcarriers; the life form detection method further includes: using the life form detection model to obtain the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of a plurality of subcarriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of a plurality of subcarriers in the previous (M - 1) detection periods; using the life form detection model to output a detection result that there is no life form in the environment to be detected when it is determined that the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of a plurality of subcarriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of a plurality of subcarriers in the previous (M - 1) detection periods are both greater than a preset threshold.

15. The life form detection method according to claim 14, wherein, the value range of M is from 5 to 10 times.

16. A life form detection module, wherein, comprising: an acquisition sub-module, adapted to acquire the channel state information of the multi-path received signals of the Wi-Fi receiving device in the current detection period; a detection sub-module, adapted to input the channel state information of the multi-path received signals of the Wi-Fi receiving device in the current detection period into the life form detection model generated by the method for generating a life form detection model according to any one of claims 1 to 6, and obtain the corresponding life form detection result.

17. The life form detection module according to claim 16, wherein, each of the received signals includes a plurality of subcarriers; the detection sub-module is further adapted to use the life form detection model to obtain the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of a plurality of subcarriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of a plurality of subcarriers in the previous (M - 1) detection periods; use the life form detection model to output a detection result that there is no life form in the environment to be detected when it is determined that the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of a plurality of subcarriers in the current detection period and the entropy of the eigenvalues after normalization of the amplitude information correlation coefficient matrix of a plurality of subcarriers in the previous (M - 1) detection periods are both greater than a preset threshold.

18. The life form detection module according to claim 17, wherein, the value range of M is from 5 to 10 times.

19. A chip, wherein, the chip is integrated with a generation module of a life form detection model according to any one of claims 7 - 12 or a life form detection module according to any one of claims 16 to 18.

20. An electronic device, wherein, Comprising at least one memory and at least one processor, the memory storing one or more computer instructions, wherein the one or more computer instructions are executed by the processor to implement the method for generating a living body detection model according to any one of claims 1 to 6 or the living body detection method according to any one of claims 13 to 15.

21. A storage medium, characterized in that the storage medium stores one or more computer instructions, and the one or more computer instructions are used to implement the method for generating a living body detection model according to any one of claims 1 to 6 or the living body detection method according to any one of claims 13 to 15.

Images