Signal detection method and device, equipment and storage medium

Abstract

The invention discloses a signal detection method, device and equipment and a storage medium, which are used for improving the accuracy of respiratory signal and heartbeat signal detection. The method comprises the following steps: preprocessing an echo signal to obtain a target phase signal; on the basis of the target phase signal and an obtained pre-constructed overall dictionary matrix, a respiration signal and a heartbeat signal are obtained; the overall dictionary matrix is a discontinuous matrix constructed according to a dictionary matrix corresponding to the respiration signals and a dictionary matrix corresponding to the heartbeat signals. According to the method and the device, the overall dictionary moment constructed by adopting the dictionary matrixes corresponding to the respiration signals and the heartbeat signals is a discontinuous matrix, so that the accuracy of separation of the respiration signals and the heartbeat signals is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of signal detection, and in particular, to a signal detection method, device, equipment, and storage medium. Background Art

[0002] In the fields of personal health management and sleep quality monitoring, non-invasive and portable monitoring devices have gradually gained attention. Especially in the home environment, such devices not only need to continuously monitor physiological parameters such as breathing and heart rate without disturbing the user's daily life, but also require high precision and reliability. As an important long-range life detection tool, radar provides a contactless method for monitoring life signals. By processing the echoes reflected from the human body, radar can capture the tiny displacements generated by the heartbeat and breathing.

[0003] However, the displacement generated by the heartbeat is usually too weak to be easily masked by the larger-amplitude breathing displacement. In this case, it becomes extremely challenging to accurately detect the heartbeat signal. At the same time, since the frequencies of life signals are very close, usually around 1 Hz, it is difficult to effectively separate them using ordinary frequency filters. This problem has led to the situation that existing life signal detection methods can usually only detect stronger breathing signals, resulting in inaccurate detection of the heartbeat signal. This technical difficulty limits the ability of existing radar products for health monitoring to provide accurate health data. Summary of the Invention

[0004] The purpose of the present invention is to provide a signal detection method, device, equipment, and storage medium to improve the accuracy of breathing signal and heartbeat signal detection.

[0005] In a first aspect, an embodiment of the present application provides a signal detection method, which includes:

[0006] Obtain an echo signal; perform preprocessing on the echo signal to obtain a target phase signal; based on the target phase signal and the pre-constructed overall dictionary matrix obtained, obtain a breathing signal and a heartbeat signal.

[0007] In the present application, a pre-constructed overall dictionary matrix is used to separate radar signals. Since the overall dictionary matrix is a discontinuous matrix, the accuracy of separating breathing signals and heartbeat signals is ensured.

[0008] In some possible embodiments, the echo signal includes a fast time dimension and a slow time dimension. Performing preprocessing on the echo signal to obtain a target phase signal includes: performing coherent accumulation on the fast time dimension in the echo signal to obtain a coherently accumulated signal; performing a phase extraction operation based on the coherently accumulated signal to obtain phase information; performing calibration processing on the phase information to obtain a target phase signal.

[0009] In this application, by performing coherent accumulation on the echo signal, the signal-to-noise ratio is improved, ensuring the accuracy of the extracted phase information.

[0010] In some possible embodiments, correcting the phase information to obtain a target phase signal includes: performing linear regression processing on the phase information to obtain the slope and intercept corresponding to the phase information; performing amplitude-phase processing on the phase information based on the slope and intercept to obtain a processed phase signal; and performing normalization processing on the processed phase signal to obtain the target phase signal.

[0011] In this application, by correcting the phase information, the phase linear drift caused by the environment can be eliminated.

[0012] In some possible embodiments, obtaining the respiration signal and the heartbeat signal based on the target phase signal and the pre-constructed overall dictionary matrix includes: obtaining the pre-constructed overall dictionary matrix; obtaining the sparse representation of the target phase signal on the overall dictionary matrix based on the overall dictionary matrix and the target phase signal; obtaining the respiration signal based on the overall dictionary matrix, and obtaining the heartbeat signal based on the sparse representation.

[0013] In this application, by constructing a discontinuous overall dictionary matrix, the respiration signal and the heartbeat signal obtained in this application are more accurate.

[0014] In some possible embodiments, obtaining the respiration signal based on the overall dictionary matrix includes: performing a fast Fourier transform on each column in the overall dictionary matrix to obtain the frequency of each column; for the frequency of each column, determining whether the frequency belongs to a preset respiration frequency range; if it belongs, determining the frequency as the first frequency; constructing a first support set based on the first frequency; and obtaining the respiration signal based on the first support set.

[0015] In this application, constructing the respiration signal in the time domain makes the constructed respiration signal more accurate.

[0016] In some possible embodiments, obtaining the heartbeat signal based on the sparse representation includes: performing a fast Fourier transform on each row in the sparse representation to obtain the frequency of each row; for the frequency of each row, determining whether the frequency belongs to a preset heartbeat frequency range; if it belongs, determining the frequency as the second frequency; constructing a second support set based on the second frequency; and obtaining the heartbeat signal based on the second support set.

[0017] In this application, constructing the heartbeat signal in the time domain makes the constructed respiration signal more accurate.

[0018] In some possible embodiments, before obtaining the respiration signal and the heartbeat signal by comparing the target phase signal with the pre-constructed overall dictionary matrix, the method further includes: updating the pre-constructed overall dictionary matrix to obtain an updated overall dictionary matrix; obtaining the respiration signal and the heartbeat signal based on the target phase signal and the pre-constructed overall dictionary matrix includes: obtaining the respiration signal and the heartbeat signal based on the target phase signal and the updated overall dictionary matrix.

[0019] In this application, by updating the overall dictionary matrix, the overall dictionary matrix becomes more accurate, thereby further improving the accuracy of the respiration signal and the heartbeat signal.

[0020] In some possible embodiments, updating the pre-constructed overall dictionary matrix includes: obtaining the sparse representation of the target phase signal on the overall dictionary matrix based on the overall dictionary matrix and the target phase signal; updating the sparse representation using the orthogonal matching pursuit method to obtain an updated sparse representation; updating the overall dictionary matrix using the singular value decomposition method based on the updated sparse representation to obtain an updated overall dictionary matrix.

[0021] In this application, by updating the overall dictionary matrix, the overall dictionary matrix becomes more accurate, thereby further improving the accuracy of the respiration signal and the heartbeat signal.

[0022] In some possible embodiments, after obtaining the respiration signal and the heartbeat signal, the method further includes: obtaining the respiration rate based on the respiration signal and a preset respiration rate range, and obtaining the heart rate based on the heartbeat signal and a preset heart rate range.

[0023] In this application, the heart rate can be obtained based on the heartbeat signal and the respiration rate can be obtained according to the respiration signal, which is convenient for more intuitively understanding the user's physical condition.

[0024] In some possible embodiments, obtaining the respiration rate based on the respiration signal and a preset respiration rate range includes: performing direct current filtering on the respiration signal to obtain a filtered respiration signal; performing a fast Fourier transform on the filtered respiration signal to obtain a first respiration rate; determining the first respiration rate within the preset respiration rate range as a second respiration rate; and taking the maximum value of the second respiration rate as the respiration rate.

[0025] In this application, by processing the respiration signal, the influence of the environment on the respiration signal is further avoided, thereby ensuring that the obtained respiration rate is more accurate.

[0026] In some possible embodiments, obtaining a heart rate based on a heartbeat signal and a preset heart rate range includes: performing DC filtering on the heartbeat signal to obtain a filtered heartbeat signal; performing a fast Fourier transform on the filtered heartbeat signal to obtain a first heart rate; determining a first heart rate within the preset heart rate range as a second heart rate; and taking the maximum value among the second heart rates as the heart rate.

[0027] In the present application, by processing the heartbeat signal, the influence of the environment on the heartbeat signal is further avoided, thereby ensuring that the obtained heart rate is more accurate.

[0028] In some possible embodiments, a preset overall dictionary matrix is constructed according to the following method: obtaining a dictionary matrix corresponding to a respiration signal based on a preset respiration frequency range and a preset number of samples; obtaining a dictionary matrix corresponding to a heartbeat signal based on a preset heart rate range and a preset number of samples; obtaining a dictionary matrix corresponding to an interference signal based on a preset interference frequency range and a preset number of samples; and obtaining an overall dictionary matrix based on the dictionary matrix corresponding to the respiration signal, the dictionary matrix corresponding to the heartbeat signal, and the dictionary matrix corresponding to the interference signal.

[0029] In the present application, constructing a discontinuous overall dictionary matrix based on the respiration signal, the heartbeat signal, and the interference signal ensures the accuracy of separating the echo signal.

[0030] In some possible embodiments, after obtaining the overall dictionary matrix, the method further includes: determining a respiration phase and a heartbeat phase through an accumulator; obtaining a total phase based on the respiration phase, the heartbeat phase, and a preset interference phase; constructing a training sample set based on the total phase; and updating the overall dictionary matrix using the training sample set, and taking the updated overall dictionary matrix as the overall dictionary matrix.

[0031] In the present application, by updating the overall dictionary matrix, the overall dictionary matrix becomes more accurate.

[0032] In some possible embodiments, before obtaining the echo signal, the method further includes: transmitting a radar signal; the radar signal is used to form an echo signal after being reflected by the human body.

[0033] In a second aspect, an embodiment of the present application further provides a signal detection device, the device includes:

[0034] a receiving module, obtaining an echo signal; a preprocessing module, configured to preprocess the echo signal to obtain a target phase signal;

[0035] A signal separation module, configured to obtain a respiration signal and a heartbeat signal based on a target phase signal and a pre-constructed overall dictionary matrix; the overall dictionary matrix is a discontinuous matrix constructed according to a dictionary matrix corresponding to the respiration signal and a dictionary matrix corresponding to the heartbeat signal.

[0036] In some possible embodiments, the preprocessing module is specifically configured to: perform coherent accumulation on the fast time dimension in the echo signal to obtain a coherently accumulated signal; perform a phase extraction operation based on the coherently accumulated signal to obtain phase information; perform a correction process on the phase information to obtain a target phase signal.

[0037] In some possible embodiments, the preprocessing module is specifically configured to: perform linear regression processing on the phase information to obtain a slope and an intercept corresponding to the phase information; perform amplitude-phase processing on the phase information based on the slope and the intercept to obtain a processed phase signal; perform normalization processing on the processed phase signal to obtain a target phase signal.

[0038] In some possible embodiments, the signal separation module is specifically configured to: obtain a pre-constructed overall dictionary matrix; obtain a sparse representation of the target phase signal on the overall dictionary matrix based on the overall dictionary matrix and the target phase signal; obtain a respiration signal based on the overall dictionary matrix and obtain a heartbeat signal based on the sparse representation.

[0039] In some possible embodiments, the signal separation module is specifically configured to: perform a fast Fourier transform on each column in the overall dictionary matrix to obtain the frequency of each column; for the frequency of each column, determine whether the frequency belongs to a preset respiration frequency range; if so, determine the frequency as a first frequency; construct a first support set based on the first frequency; obtain a respiration signal based on the first support set.

[0040] In some possible embodiments, the signal separation module is specifically configured to: perform a fast Fourier transform on each row in the sparse representation to obtain the frequency of each row; for the frequency of each row, determine whether the frequency belongs to a preset heartbeat frequency range; if so, determine the frequency as a second frequency; construct a second support set based on the second frequency; obtain a heartbeat signal based on the second support set.

[0041] In some possible embodiments, the signal separation module is further configured to: perform an update process on the pre-constructed overall dictionary matrix to obtain an updated overall dictionary matrix; obtain a respiration signal and a heartbeat signal based on the target phase signal and the pre-constructed overall dictionary matrix obtained, including: obtaining a respiration signal and a heartbeat signal based on the target phase signal and the updated overall dictionary matrix.

[0042] In some possible embodiments, the signal separation module is specifically configured to: obtain the sparse representation of the target phase signal on the overall dictionary matrix based on the overall dictionary matrix and the target phase signal; update the sparse representation by using the orthogonal matching pursuit method to obtain the updated sparse representation; and update the overall dictionary matrix by using the singular value decomposition method based on the updated sparse representation to obtain the updated overall dictionary matrix.

[0043] In some possible embodiments, the signal separation module is further configured to: obtain the breathing frequency based on the breathing signal and the preset breathing frequency range, and obtain the heartbeat frequency based on the heartbeat signal and the preset heartbeat frequency range.

[0044] In some possible embodiments, the signal separation module is specifically configured to: perform DC filtering on the breathing signal to obtain the filtered breathing signal; perform fast Fourier transform on the filtered breathing signal to obtain the first breathing frequency; determine the first breathing frequency within the preset breathing frequency range as the second breathing frequency; and use the maximum value in the second breathing frequency as the breathing frequency.

[0045] In some possible embodiments, the signal separation module is specifically configured to: perform DC filtering on the heartbeat signal to obtain the filtered heartbeat signal; perform fast Fourier transform on the filtered heartbeat signal to obtain the first heartbeat frequency; determine the first heartbeat frequency within the preset heartbeat frequency range as the second heartbeat frequency; and use the maximum value in the second heartbeat frequency as the heartbeat frequency.

[0046] In some possible embodiments, the signal separation module is specifically configured to: obtain the dictionary matrix corresponding to the breathing signal based on the preset breathing frequency range and the preset number of samples; obtain the dictionary matrix corresponding to the heartbeat signal based on the preset heartbeat frequency range and the preset number of samples; obtain the dictionary matrix corresponding to the interference signal based on the preset interference frequency range and the preset number of samples; and obtain the overall dictionary matrix based on the dictionary matrix corresponding to the breathing signal, the dictionary matrix corresponding to the heartbeat signal, and the dictionary matrix corresponding to the interference signal.

[0047] In some possible embodiments, the signal separation module is further configured to: determine the breathing phase and the heartbeat phase through an accumulator; obtain the total phase based on the breathing phase, the heartbeat phase, and the preset interference phase;

[0048] construct a training sample set based on the total phase; update the overall dictionary matrix by using the training sample set, and use the updated overall dictionary matrix as the overall dictionary matrix.

[0049] In some possible embodiments, the receiving module is further configured to: send a radar signal; the radar signal is used to form an echo signal after being reflected by a human body.

[0050] In a third aspect, another embodiment of the present application further provides an electronic device, including at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute any method provided in the embodiment of the first aspect of the present application.

[0051] In a fourth aspect, another embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is used to cause a computer to execute any method provided in the embodiment of the first aspect of the present application.

[0052] Other features and advantages of the present application will be described in the following description, and in part will be obvious from the description, or will be understood by implementing the present application. The objectives and other advantages of the present application can be realized and obtained by the structures specifically pointed out in the written description, claims, and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a page schematic diagram of a non-invasive monitoring device for a signal detection method provided by an embodiment of the present application;

[0054] Figure 2 It is an echo schematic diagram of a respiration signal and a heartbeat signal of a signal detection method provided by an embodiment of the present application;

[0055] Figure 3 It is an overall process schematic diagram of a signal detection method provided by an embodiment of the present application;

[0056] Figure 4 It is a process schematic diagram of preprocessing an echo signal of a signal detection method provided by an embodiment of the present application;

[0057] Figure 5 It is a process schematic diagram of correcting phase information of a signal detection method provided by an embodiment of the present application;

[0058] Figure 6A It is a phase schematic diagram of an echo signal containing respiration and heartbeat information of a signal detection method provided by an embodiment of the present application;

[0059] Figure 6B It is a signal schematic diagram after coherent accumulation in the fast time dimension of a signal detection method provided by an embodiment of the present application;

[0060] Figure 6C It is a schematic diagram of a target phase signal after amplitude-phase error correction and normalization of a signal detection method provided by an embodiment of the present application;

[0061] Figure 7 It is a schematic flowchart of obtaining a respiratory signal and a heartbeat signal for a signal detection method provided by an embodiment of the present application;

[0062] Figure 8 It is a schematic flowchart of constructing an overall dictionary matrix for a signal detection method provided by an embodiment of the present application;

[0063] Figure 9A It is a schematic flowchart of training an overall dictionary matrix for a signal detection method provided by an embodiment of the present application;

[0064] Figure 9B It is a schematic diagram of an actually measured respiratory signal, heartbeat signal, and radar phase signal for a signal detection method provided by an embodiment of the present application;

[0065] Figure 9C It is a schematic diagram of three groups of simulation signals for a signal detection method provided by an embodiment of the present application;

[0066] Figure 10 It is a schematic flowchart of updating a pre-constructed overall dictionary matrix by using a training sample set for a signal detection method provided by an embodiment of the present application;

[0067] Figure 11 It is a schematic flowchart of obtaining a respiratory signal based on an overall dictionary matrix for a signal detection method provided by an embodiment of the present application;

[0068] Figure 12 It is a schematic flowchart of obtaining a heartbeat signal based on sparse representation for a signal detection method provided by an embodiment of the present application;

[0069] Figure 13 It is a schematic flowchart of obtaining a respiratory rate based on a respiratory signal and a preset respiratory rate range for a signal detection method provided by an embodiment of the present application;

[0070] Figure 14 It is a schematic flowchart of obtaining a heartbeat rate based on a heartbeat signal and a preset heartbeat rate range for a signal detection method provided by an embodiment of the present application;

[0071] Figure 15A It is a schematic diagram of a target phase signal for a signal detection method provided by an embodiment of the present application;

[0072] Figure 15B For a signal detection method provided by an embodiment of the present application, using the present application Figure 3 The obtained respiratory signal schematic diagram shown in;

[0073] Figure 15CThe heartbeat signal schematic diagram obtained by using the steps shown in the signal detection method provided in the embodiments of the present application and the present application Figure 3 as shown;

[0074] Figure 16 The flowchart of a signal detection method provided in the embodiments of the present application;

[0075] Figure 17 The flowchart of updating the pre-constructed overall dictionary matrix in the signal detection method provided in the embodiments of the present application;

[0076] Figure 18 The overall flowchart of the signal detection method provided in the embodiments of the present application;

[0077] Figure 19 The device schematic diagram of the signal detection method provided in the embodiments of the present application;

[0078] Figure 20 The schematic diagram of an electronic device for the signal detection method provided in the embodiments of the present application. Detailed implementation manners

[0079] The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0080] For a better understanding of the technical solutions of the present application, the embodiments of the present application will be described in detail below with reference to the accompanying drawings.

[0081] It should be clear that the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts belong to the scope of protection of the present application.

[0082] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. The singular forms of "a", "the" and "said" used in the embodiments of the present application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0083] It should be understood that the term " / and / " used herein is only a description of the associated relationship of the associated objects, indicating that there can be three relationships, for example, A and / or B, which can mean: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after.

[0084] The inventors' research found that in the field of personal health management and sleep quality monitoring, non-invasive and portable monitoring devices have gradually gained attention. Especially in a home environment, such as Figure 1 shown, such devices not only need to continuously monitor physiological parameters such as respiratory and heart rates without disturbing the user's daily life, but also require high precision and reliability. Radar, as an important long-range life detection tool, provides a contactless method for monitoring life signals. By processing the echoes reflected from the human body, radar can capture the minute displacements generated by the heartbeat and respiration.

[0085] However, the displacement generated by the heartbeat is usually too weak to be easily masked by the larger-amplitude respiratory displacement, as Figure 2 shown. In this case, it becomes extremely challenging to accurately detect the heartbeat signal. At the same time, since the frequencies of life signals are very close, usually around 1 Hz, it is difficult to effectively separate them using ordinary frequency filters. This problem has led to the situation that existing life signal detection methods can usually only detect stronger respiratory signals, resulting in inaccurate detection of the heartbeat signal, which limits the ability of existing radar products for health monitoring to provide accurate health data.

[0086] To address the above problems, the embodiments of the present application provide a signal detection method, apparatus, device, and storage medium to solve the above problems. The inventive concept of the present application can be summarized as: obtaining an echo signal; preprocessing the echo signal to obtain a target phase signal; obtaining a respiratory signal and a heartbeat signal based on the target phase signal and a pre-constructed overall dictionary matrix obtained; the overall dictionary matrix is a discontinuous matrix constructed according to the dictionary matrix corresponding to the respiratory signal and the dictionary matrix corresponding to the heartbeat signal. In the present application, the overall dictionary matrix constructed using the dictionary matrices corresponding to the respiratory signal and the heartbeat signal respectively has discontinuous frequencies. Since the frequency gap between the respiratory signal and the heartbeat signal is relatively large, the respiratory signal and the heartbeat signal can be better separated according to the frequencies corresponding to the respiratory signal and the heartbeat signal respectively, thereby ensuring the accuracy of separating the respiratory signal and the heartbeat signal.

[0087] A signal detection method provided by the present application is applicable to various terminal devices, such as devices with radar signal detection capabilities including but not limited to computers, laptops, smartphones, tablets, smartwatches, smart bracelets, wireless earphones, vehicle-mounted terminals, etc.

[0088] To facilitate a further understanding of a signal detection method provided by the embodiments of the present application, the following will provide a detailed description of a signal detection method provided by the embodiments of the present application with reference to the accompanying drawings:

[0089] As Figure 3As shown in the figure, it is a schematic diagram of the overall process of a signal detection method provided by an embodiment of the present application, where:

[0090] In step 301: Receive the echo signal.

[0091] In the present application, when the execution entity of a signal detection method has the ability to send radar signals, the echo signal is the signal formed after the radar signal is reflected by the human body after the execution entity sends the radar signal; when the execution entity of the signal detection method does not have the ability to send radar signals, the echo signal is the signal formed after the radar signal is transmitted by the human body after being sent by a device with the ability to send radar signals.

[0092] That is, in the present application, the radar signal can be emitted by the device that executes Figure 3 the steps, or can be emitted by a separate radar device. The present application does not make any limitations in this regard.

[0093] To ensure the accuracy of signal detection, when the user uses a signal detection method provided by the present application, the device that sends the radar signal can be placed facing the chest. For example: If the user lies flat on the bed and the device that sends the radar signal is a millimeter-wave radar device, the millimeter-wave radar device can be set at the position on the ceiling facing the chest to implement Figure 3 the steps. If the device that executes the steps is the user's mobile phone, the millimeter-wave radar sends the radar signal, and the mobile phone receives the generated echo signal; if a radar emission device is installed in the user's mobile phone and the user sleeps on the side, the mobile phone can be hung on the wall facing the chest. It should be noted that the present application does not limit the installation position of the device that sends the radar signal, and the above are only two examples.

[0094] It should be noted that the above is only an example with the radar being a millimeter-wave radar, and it does not limit the type of radar. The radar device in the present application can be an ultra-wideband radar (UWB radar), a millimeter-wave radar, etc.

[0095] To save computing resources on the basis of ensuring the accuracy of signal detection, the steps shown in Figure 3 can be executed periodically. Each time the steps shown in Figure 3 are executed, the echo signal used is the echo signal received in this period.

[0096] For example: If the preset period is 10 minutes, the steps shown in Figure 3 are executed every ten minutes.

[0097] In step 302: Preprocess the echo signal to obtain the target phase signal.

[0098] In this application, in order to ensure the accuracy and stability of signal detection, after obtaining the echo signal, it is necessary to preprocess the echo signal to improve the signal-to-noise ratio of the signal.

[0099] In some possible embodiments, the echo signal includes a fast time dimension and a slow time dimension. When preprocessing the echo signal, it can be specifically implemented as Figure 4 the steps shown below, where:

[0100] In step 401: Coherently accumulate the fast time dimension in the echo signal to obtain the coherently accumulated signal.

[0101] In this application, the echo signal can be expressed as: S(r, t), where r represents the fast time dimension and t represents the slow time dimension.

[0102] In some possible embodiments, coherently accumulating the echo signal in the fast time dimension can improve the signal-to-noise ratio. When coherently accumulating the fast time dimension in the echo signal, formula 1 can be used, where:

[0103]

[0104] where R is the number of fast time dimensions in the received echo signal, S(r, t) is the received echo signal, S acc (t) is the coherently accumulated signal, r is the fast time dimension, and t is the slow time dimension.

[0105] For example: The period is 10 minutes. A signal detection was performed at 12:10. Now the time is 12:20. Then a signal detection needs to be performed here. The echo signal participating in this signal detection is the echo signal received between 12:10 and 12:20.

[0106] In step 402: Based on the coherently accumulated signal, perform a phase extraction operation to obtain phase information.

[0107] In this application, after coherently accumulating the echo signal, a phase extraction operation needs to be performed to obtain the phase information corresponding to the echo signal.

[0108] In some possible embodiments, when performing a phase extraction operation on the coherently accumulated signal, formula 2 can be used, where:

[0109]

[0110] where, is the phase information, S acc (t) is the coherently accumulated signal, and the arg(·) function is used to extract the phase angle of the complex signal.

[0111] In step 403: perform a correction process on the phase information to obtain a target phase signal.

[0112] In this application, in order to eliminate the linear drift caused by the system or the environment, after obtaining the phase information, it is necessary to perform a correction process on the phase information.

[0113] In some possible embodiments, the correction process on the phase information can be specifically implemented as the steps shown in Figure 5 as follows, where:

[0114] In step 501: perform a linear regression process on the phase information to obtain the slope and intercept corresponding to the phase information.

[0115] In this application, when performing a linear regression process on the phase information, formula 3 can be used to obtain the slope corresponding to the phase information, and formula 5 can be used to obtain the intercept corresponding to the phase information, where:

[0116]

[0117] where: a is the slope, N is the number of slow time dimensions in the received echo signal, is the mean of the slow time dimensions in the received echo signal, is the mean of.

[0118]

[0119] where b is the intercept, a is the slope, is the mean of the slow time dimensions in the received echo signal, is the mean of the phase information.

[0120] In step 502: perform an amplitude-phase process on the phase information based on the slope and intercept to obtain a processed phase signal.

[0121] After obtaining the slope and intercept, the amplitude-phase process can be performed on the phase information according to the linear trend. In specific implementation, formula 5 can be used, where:

[0122]

[0123] where, is the processed phase signal, is the phase information, b is the intercept, a is the slope, and t is the slow time dimension.

[0124] In step 503: perform a normalization process on the processed phase signal to obtain a target phase signal.

[0125] In this application, in order to further ensure the accuracy of signal detection, after obtaining the processed phase signal, it is necessary to normalize the processed phase signal.

[0126] In some possible embodiments, when normalizing the processed phase signal, formula 6 can be used, where:

[0127]

[0128] where, is the target phase signal, is the processed phase signal, is the maximum value in the processed phase signal, is the minimum value in the processed phase signal.

[0129] For example: The phase of the echo signal containing respiration and heartbeat information obtained is as Figure 6A shown, then the signal after coherent accumulation in the fast time dimension is as Figure 6B shown, and it can be seen from Figure 6B that the signal-to-noise ratio has been significantly improved. The target phase signal after amplitude-phase error correction and normalization is as Figure 6C shown, and compared with Figure 6B , the trend of linear drift is well eliminated.

[0130] In step 303: Based on the target phase signal and the pre-constructed overall dictionary matrix obtained, the respiration signal and the heartbeat signal are obtained.

[0131] In this application, the overall dictionary matrix is divided into three parts, which respectively represent the respiration signal, the heartbeat signal, and the interference signal. The three parts in the overall dictionary matrix in this application are discontinuous and have different frequencies. The discontinuity of the three parts of the overall dictionary matrix can better separate the respiration signal, the heartbeat signal, and the interference signal parts.

[0132] In some possible embodiments, based on the target phase signal and the pre-constructed overall dictionary matrix obtained, the respiration signal and the heartbeat signal are obtained, which can be specifically implemented as the steps as Figure 7 shown, where:

[0133] Step 701: Obtain the pre-constructed overall dictionary matrix.

[0134] In some possible embodiments, constructing the overall dictionary matrix can be specifically implemented as the steps as Figure 8 shown, where:

[0135] In step 801: Based on a preset respiration frequency range and a preset number of samplings, a dictionary matrix corresponding to the respiration signal is obtained. In this application, a preset respiration frequency range can be set in advance. Set the lower limit of the preset respiration frequency range to Set the upper limit to Set the preset number of samplings corresponding to respiration to k B , then the dictionary matrix corresponding to the respiration signal is as shown in Equation 7:

[0136]

[0137] where D B is the dictionary matrix corresponding to the respiration signal, and the basis vectors therein where are sampled from the preset respiration frequency range based on the frequency interval, and the frequency interval is obtained based on Equation 8 and the preset number of samplings corresponding to respiration, where:

[0138]

[0139] where Δf B is the frequency interval, is the upper limit of the preset respiration frequency range, is the lower limit of the preset respiration frequency range, and k B is the preset number of samplings corresponding to respiration.

[0140] For example: The preset number of samplings is 11, the upper limit of the preset respiration frequency range is 2 Hz, and the lower limit is 0.5 Hz, then the frequency interval is 0.15 Hz.

[0141] In step 802: Based on a preset heart rate frequency range and a preset pair of sampling numbers, a dictionary matrix corresponding to the heart rate signal is obtained.

[0142] In this application, a preset heart rate frequency range can be set in advance. Set the lower limit of the preset heart rate frequency range to Set the upper limit to Set the preset number of samplings corresponding to heart rate to k H , then the dictionary matrix corresponding to the heart rate signal is as shown in Equation 9:

[0143]

[0144] where D H is the dictionary matrix corresponding to the heart rate signal, and the basis vectors therein where are sampled from the preset heart rate frequency range based on the frequency interval, and the frequency interval is obtained based on Equation 10 and the preset number of samplings, where:

[0145]

[0146] Among them, Δf H is the frequency interval, is the upper limit of the preset heart rate frequency range, is the lower limit of the preset heart rate frequency range, and k H is the preset number of samples corresponding to the heart beat.

[0147] For example: if the preset number of samples is 11, the upper limit of the preset heart rate frequency range is 2 Hz, and the lower limit is 0.5 Hz, then the frequency interval is 0.15 Hz.

[0148] In step 803: Based on the preset interference frequency range and the preset number of samples, a dictionary matrix corresponding to the interference signal is obtained.

[0149] In this application, the preset interference frequency range can be set in advance. The lower limit of the preset interference frequency range is set to The upper limit is set to The preset number of samples corresponding to the interference is set to k N , then the dictionary matrix corresponding to the interference signal is as shown in Formula 11:

[0150]

[0151] Among them, d N is the dictionary matrix corresponding to the interference signal, and the basis vectors therein wherein are sampled from the preset interference frequency range based on the frequency interval, and the frequency interval is obtained based on Formula 12 and the preset number of samples, where:

[0152]

[0153] Among them, Δf N is the frequency interval, is the upper limit of the preset interference frequency range, is the lower limit of the preset interference frequency range, and k N is the preset number of samples corresponding to the interference.

[0154] For example: if the preset number of samples is 11, the upper limit of the preset interference frequency range is 2 Hz, and the lower limit is 0.5 Hz, then the frequency interval is 0.15 Hz.

[0155] In step 804: Based on the dictionary matrix corresponding to the respiratory signal, the dictionary matrix corresponding to the heart beat signal, and the dictionary matrix corresponding to the interference signal, an overall dictionary matrix is obtained.

[0156] In this application, the overall dictionary matrix can be regarded as a combination of the dictionary matrices corresponding to the respiration signal, the heartbeat signal, and the interference signal respectively. Therefore, the overall dictionary matrix in this application can be represented by Formula 13, where:

[0157] D = D B + D H + D H , (Formula 13)

[0158] Among them, D is the overall dictionary matrix, D B is the dictionary matrix corresponding to the respiration signal, D H is the dictionary matrix corresponding to the heartbeat signal, D N is the dictionary matrix corresponding to the interference signal.

[0159] In some possible embodiments, in order to make the overall dictionary matrix more in line with the actual situation, after obtaining the overall dictionary matrix, the overall dictionary matrix can be trained, which can be specifically implemented as the steps shown in Figure 9A as follows, where:

[0160] In step 901: Determine the respiration phase and the heartbeat phase through an accumulator, and obtain the total phase based on the respiration phase, the heartbeat phase, and a preset interference phase.

[0161] Since the frequencies of respiration and heartbeat are determined by the number of breaths and heartbeats per minute of an individual respectively, the respiration frequency is as shown in Formula 14, and the heartbeat frequency is as shown in Formula 15:

[0162] where:

[0163]

[0164] Among them, f B is the respiration frequency, bpm B is the total number of breaths of an individual in one minute.

[0165]

[0166] Among them, f H is the heartbeat frequency, bpm H is the total number of heartbeats of an individual in one minute.

[0167] In this application, an accumulator can be used to simulate the displacement of the chest cavity of an individual caused by respiration and heartbeat. The accumulator accumulates the phase at each sampling time point t i , then the displacement caused by respiration is as shown in Formula 16, and the displacement caused by heartbeat is as shown in Formula 17:

[0168] d B (t i ) = A B · εB (t i )·sin(acc B (t i )),(Formula 16)

[0169] Wherein, d B (t i ) is the displacement caused by breathing at time t i , A B is the preset breathing displacement amplitude, ε B (t i ) is the preset breathing amplitude change factor, acc B (t i ) is the phase accumulation value corresponding to the breathing accumulated by the accumulator at time t i .

[0170] d H (t i ) = A H ·ε H (t i )·sin(acc H (t i )),(Formula 17)

[0171] Wherein, d H (t i ) is the displacement caused by heartbeat at time t i , A H is the preset heartbeat displacement amplitude, ε H (t i ) is the preset heartbeat amplitude change factor, acc H (t i ) is the phase accumulation value corresponding to the heartbeat accumulated by the accumulator at time t i .

[0172] In some possible embodiments, the update method of acc B (t i ) is as shown in Formula 18:

[0173]

[0174] Wherein, acc B (t i ) is the phase accumulation value corresponding to the breathing accumulated by the accumulator at time t i , acc B (t i-1 ) is the phase accumulation value corresponding to the breathing accumulated by the accumulator last time, f B is the breathing frequency, τ B (t i ) is the preset time change factor corresponding to the breathing, fs is the sampling frequency of the radar.

[0175] In some possible embodiments, acc H (t i ) is updated as shown in Equation 19:

[0176]

[0177] where acc H (t i ) is the phase accumulation value corresponding to the respiration accumulated by the accumulator at time t, acc i (t H ) is the phase accumulation value corresponding to the previous respiration accumulated by the accumulator, f i-1 is the respiration frequency, τ H (t H ) is the preset time variation factor corresponding to the respiration, and f i is the sampling frequency of the radar. s is the sampling frequency of the radar.

[0178] where τ B (t i ) and τ H (t i ) are used to simulate the time difference between respiration and heartbeat.

[0179] In summary, the phase change caused by the displacement generated by respiration, i.e., the respiration phase, is as shown in Equation 20, where:

[0180]

[0181] where is the respiration phase, λ is the wavelength of the radar, and d B (t) is the displacement caused by respiration.

[0182] The phase change caused by the displacement generated by heartbeat, i.e., the heartbeat phase, is as shown in Equation 21, where:

[0183]

[0184] where is the heartbeat phase, λ is the wavelength of the radar, and d H (t) is the displacement caused by heartbeat.

[0185] In this application, in order to make the phase closer to the real situation, an interference term is added in this application. The total phase in this application is as shown in Equation 22:

[0186]

[0187] where is the total phase, is the respiration phase, is the heartbeat phase, is the interference term.

[0188] For example: Figure 9B In it are three groups of measured respiration signals, heartbeat signals, and radar phase signals, Figure 9C are three groups of simulation signals generated according to step 901. By comparing with the real signals, it can be seen that the simulation signals can well simulate the phase changes generated by respiration and heartbeat, and at the same time also show similar interference and fluctuations to the real signals.

[0189] In step 902: Construct a training sample set based on the total phase.

[0190] In this application, in order to ensure the accuracy of the constructed overall dictionary matrix, the constructed overall dictionary matrix can be updated. In order to enrich the training sample set, step 901 can be executed a preset number of times to obtain a preset number of total phases, and then a training sample set is constructed based on the preset number of total phases.

[0191] In step 903: Update the overall dictionary matrix using the training sample set, and use the updated overall dictionary matrix as the overall dictionary matrix.

[0192] In some possible embodiments, the pre-constructed overall dictionary matrix is updated using the training sample set, and the specific implementation can be as Figure 10 shown in the steps, where:

[0193] In step 1001: Based on the overall dictionary matrix and the target phase matrix, obtain the sparse representation of the target phase matrix on the overall dictionary matrix.

[0194] Among them, the target phase matrix is constructed based on the total phase in the training sample set.

[0195] The sparse representation of the target phase matrix on the overall dictionary matrix is obtained using formula 23:

[0196] Φ = DX, (formula 23)

[0197] Among them, Φ is the target phase matrix, D is the overall dictionary matrix, and X is the sparse representation of the target phase matrix on the overall dictionary matrix.

[0198] Since the sparse representation X is a k×m matrix, where k = k B + k H + k N , k B is the preset sampling quantity corresponding to respiration, k H is the preset sampling quantity corresponding to heartbeat, k NFor the preset sampling quantity corresponding to interference, and m is the quantity of the total phases in the training sample set. Therefore, Equation 23 can be expanded based on Equation 7, Equation 9, and Equation 11, and then Equation 24 can be obtained:

[0199]

[0200] Among them, the basis vector Among them is sampled based on the frequency interval in the preset breathing frequency range, and the basis vector Among them is sampled based on the frequency interval in the preset heartbeat frequency range, and the basis vector Among them is sampled based on the frequency interval in the preset interference frequency range.

[0201] Based on Equation 7, Equation 9, and Equation 11, Equation 24 can be rewritten as Equation 25:

[0202]

[0203] Among them, X B is the sparse representation of the dictionary matrix D B corresponding to the breathing signal, and X H is the sparse representation of the dictionary matrix D H corresponding to the heartbeat signal, and X N is the sparse representation of the dictionary matrix D N corresponding to the interference signal.

[0204] In step 1002: The orthogonal matching pursuit method is used to update the sparse representation to obtain the updated sparse representation.

[0205] It can be seen from Equation 25 that the target phase signal is decomposed into a linear combination of the breathing signal, the heartbeat signal, and the interference signal. In this application, the convergence of the training of the overall dictionary matrix can be determined by setting the number of iterations. Updating the overall dictionary matrix once represents one iteration.

[0206] For example, if the number of iterations is set to τ, the convergence condition is as shown in Equation 26:

[0207]

[0208] Among them, x i is the row vector of the sparse representation X, representing the coefficient of the overall dictionary matrix D; ‖x i ‖0 is the l0 norm of x i and represents the number of non-zero elements, and τ is the number of iterations.

[0209] In some possible embodiments, the Lagrange multiplier method can be used to transform the constrained optimization problem in Equation 26 into an unconstrained optimization problem, as shown in Equation 27:

[0210]

[0211] where λ is the regularization parameter, and ‖x i ‖1 is the l1 norm of x i . The l1 norm of x is used to replace the l0 norm to simplify the calculation. Based on Equation 27, the orthogonal matching pursuit method can be used to update the sparse representation X. i

[0212] In step 1003: Based on the updated sparse representation, the singular value decomposition method is used to update the overall dictionary matrix to obtain the updated overall dictionary matrix.

[0213] After obtaining the updated sparse representation, Equation 28 can be obtained:

[0214]

[0215] where D fixed is a matrix representing the respiratory signal preset based on the overall dictionary matrix, X fixed is a matrix representing the heartbeat signal preset based on the overall dictionary matrix, d k x k is the interference signal obtained based on the overall dictionary matrix, D fixed , and X fixed . Based on Equation 28, the minimization problem of the reconstruction error can be rewritten as Equation 29:

[0216]

[0217] Based on Equation 29, the singular value decomposition method can be used to perform singular value decomposition on the residual E i to obtain Equation 30:

[0218] U∑V T = SVD(E i ), (Equation 30)

[0219] where U is the first positive definite matrix, V is the second positive definite matrix, and V T is the transpose of the second positive definite matrix, and ∑ is a diagonal matrix

[0220] In some possible embodiments, the optimal rank is composed of the first column of U, the first element of ∑, and the first column of V T . Therefore, Equations 31 and 32 can be obtained:

[0221] d​k = u1, (Equation 31)

[0222]

[0223] where u1 is the first column of U, σ1 is the first singular value of ∑ and also the largest singular value in ∑, is the first column of V T .

[0224] After obtaining d k and x k , Equation 31 and Equation 32 can be substituted into Equation 24 to obtain the updated overall dictionary matrix.

[0225] Step 702: Based on the overall dictionary matrix and the target phase signal, obtain the sparse representation of the target phase signal on the overall dictionary matrix.

[0226] The specific implementation method of this step is the same as that of Step 1001 and will not be elaborated here.

[0227] Step 703: Obtain the respiration signal based on the overall dictionary matrix and obtain the heartbeat signal based on the sparse representation.

[0228] In some possible embodiments, obtaining the respiration signal based on the overall dictionary matrix can be specifically implemented as the steps Figure 11 shown, where:

[0229] In Step 1101: Perform a fast Fourier transform on each column in the overall dictionary matrix to obtain the frequency of each column.

[0230] For example: Denote the i-th column in the overall dictionary matrix as d i (t), and the frequency corresponding to the i-th column obtained by performing a fast Fourier transform on the i-th column is:

[0231] In Step 1102: For the frequency of each column, determine whether the frequency belongs to a preset respiration frequency range; if it belongs, determine the frequency as the first frequency.

[0232] In this application, after obtaining the frequency of each column, the frequency corresponding to the respiration signal can be filtered out through the preset respiration frequency range.

[0233] In Step 1103: Construct a first support set based on the first frequency.

[0234] In this application, Equation 33 can be used to construct the first support set:

[0235]

[0236] where Ω Bis the first support set, is the upper limit of the preset breathing frequency range, is the lower limit of the preset breathing frequency range, is the i-th breathing frequency.

[0237] In step 1104: Obtain the breathing signal based on the first support set.

[0238] After obtaining the first support set, it is necessary to find the corresponding columns from the overall dictionary matrix according to the first support set to construct the breathing signal. The specific implementation is shown in Formula 34:

[0239]

[0240] where, is the breathing signal, is the corresponding column determined from the overall dictionary matrix according to the first support set, is the corresponding row determined from the sparse representation according to the first support set.

[0241] For example: The first support set includes: the first column, the second column, and the third column, then are the first column, the second column, and the third column in the overall dictionary matrix, are the first row, the second row, and the third row in the sparse representation.

[0242] In some other possible embodiments, the heartbeat signal is obtained based on the sparse representation. The specific implementation can be the steps shown in Figure 12 where:

[0243] In step 1201: Perform a fast Fourier transform on each row in the sparse representation to obtain the frequency of each row.

[0244] For example: Denote the i-th row in the sparse representation as x i (t), and the frequency corresponding to the i-th row obtained by performing a fast Fourier transform on the i-th row is:

[0245] In step 1202: For the frequency of each row, determine whether the frequency belongs to the preset heartbeat frequency range; if it belongs, then determine the frequency as the second frequency.

[0246] In this application, after obtaining the frequency of each row, the frequency corresponding to the heartbeat signal can be filtered out through the preset heartbeat frequency range.

[0247] In step 1203: Construct the second support set based on the second frequency.

[0248] In this application, Formula 35 can be used to construct the second support set:

[0249]

[0250] Among them, Ω H is the second support set, is the upper limit of the preset heart rate range, is the lower limit of the preset heart rate range, is the i-th heart rate.

[0251] In step 1204: Obtain the heartbeat signal based on the second support set.

[0252] After obtaining the second support set, it is necessary to find the corresponding rows from the sparse representation according to the second support set to construct the heartbeat signal. The specific implementation is shown in formula 36:

[0253]

[0254] Among them, is the heartbeat signal, is the corresponding column determined from the overall dictionary matrix according to the second support set, is the corresponding row determined from the sparse representation according to the second support set.

[0255] For example: The second support set includes: the first row, the second row, and the third row, then are the first column, the second column, and the third column in the overall dictionary matrix, are the first row, the second row, and the third row in the sparse representation.

[0256] In some possible embodiments, after determining the respiration signal and the heartbeat signal, in order to facilitate the user to more intuitively understand the physical condition, therefore, after implementing the steps shown in Figure 3 the respiration rate can be obtained based on the respiration signal and the preset respiration rate range, and the heart rate can be obtained based on the heartbeat signal and the preset heart rate range.

[0257] In some possible embodiments, obtaining the respiration rate based on the respiration signal and the preset respiration rate range can be specifically implemented as the steps shown in Figure 13 Among them:

[0258] In step 1301: Perform DC filtering on the respiration signal to obtain the filtered respiration signal.

[0259] Performing DC filtering on the respiration signal can be carried out using formula 37, where:

[0260]

[0261] Among them, is the filtered respiration signal, is the respiration signal, and N is the number of slow time dimensions.

[0262] In step 1302: Perform a fast Fourier transform on the filtered respiration signal to obtain the first respiration frequency.

[0263] Performing a fast Fourier transform on the filtered respiration signal can be carried out using formula 38, where:

[0264]

[0265] where, is the first respiration frequency, is the filtered respiration signal.

[0266] In step 1303: Determine the first respiration frequencies within the preset respiration frequency range as the second respiration frequencies.

[0267] In the present application, is the upper limit of the preset respiration frequency range, is the lower limit of the preset respiration frequency range. According to and the second respiration frequencies within the preset respiration frequency range can be determined, and the second respiration frequencies are denoted as

[0268] In step 1304: Take the maximum value among the second respiration frequencies as the respiration frequency.

[0269] In the present application, considering that the higher the frequency, the higher the representation intensity, so the maximum value is selected as the respiration frequency among the second respiration frequencies.

[0270] To simplify the calculation process, steps 1302 - 1304 can be simplified to formula 38:

[0271]

[0272] where, f B is the respiration frequency, is the lower limit of the preset respiration frequency range, is the filtered respiration signal.

[0273] In some possible embodiments, based on the heartbeat signal and the preset heartbeat frequency range, the heartbeat frequency is obtained, which can be specifically implemented as the steps shown in Figure 14 as follows, where:

[0274] In step 1401: Perform a DC filtering process on the heartbeat signal to obtain the filtered heartbeat signal.

[0275] The DC filtering process of the heartbeat signal can be carried out using Equation 39, where:

[0276]

[0277] Among them, is the filtered heartbeat signal, is the heartbeat signal, and N is the number of slow time dimensions.

[0278] In step 1402: Perform a fast Fourier transform on the filtered heartbeat signal to obtain the first heartbeat frequency.

[0279] The fast Fourier transform of the filtered heartbeat signal can be carried out using Equation 40, where:

[0280]

[0281] Among them, is the first heartbeat frequency, is the filtered heartbeat signal.

[0282] In step 1403: Determine the first heartbeat frequency within the preset heartbeat frequency range as the second heartbeat frequency.

[0283] In this application, is the upper limit of the preset heartbeat frequency range, is the lower limit of the preset heartbeat frequency range. According to and the second heartbeat frequency within the preset heartbeat frequency range can be determined. Denote the second heartbeat frequency as

[0284] In step 1404: Take the maximum value in the second heartbeat frequency as the heartbeat frequency.

[0285] In this application, considering that the higher the frequency, the higher the representation intensity, so the maximum value in the second heartbeat frequency is selected as the heartbeat frequency.

[0286] To simplify the calculation process, steps 1302 - 1304 can be simplified to Equation 38:

[0287]

[0288] Among them, f H is the heartbeat frequency, is the upper limit of the preset heartbeat frequency range, is the lower limit of the preset heartbeat frequency range, is the filtered heartbeat signal.

[0289] For example: The target phase signal is as shown in Figure 15A Adopt this applicationFigure 3 The respiratory signal obtained by the steps shown is as Figure 15B shown, and the obtained heartbeat signal is as Figure 15C shown.

[0290] By Figure 15A - Figure 15C , it can be seen that in Figure 15A the heartbeat information is almost completely masked by the respiratory information, and only the waveform of respiration can be seen. Figure 15B And Figure 15C are the respiratory signal and the heartbeat signal processed by the method of this application. It can be seen that the respiratory signal and the heartbeat signal are well reconstructed and the waveforms are relatively smooth without obvious interference components.

[0291] In this application, the construction of the overall dictionary matrix is divided into three parts, representing respiration, heartbeat, and interference respectively. The frequencies of different signals are different, and the discontinuity of the three parts of the overall dictionary matrix can better separate respiration, heartbeat, and interference.

[0292] To further improve the accuracy of signal detection, therefore, a signal detection method is also provided in this application, as Figure 16 shown:

[0293] Step 1601: Send a radar signal and receive an echo signal; the echo signal is formed after the radar signal is reflected by the human body.

[0294] The specific implementation method of this step is the same as that of step 301, and will not be elaborated here.

[0295] Step 1602: Perform preprocessing on the echo signal to obtain a target phase signal.

[0296] The specific implementation method of this step is the same as that of step 302, and will not be elaborated here.

[0297] Step 1603: Obtain the pre-constructed overall dictionary matrix, and perform update processing on the pre-constructed overall dictionary matrix to obtain an updated overall dictionary matrix.

[0298] In some possible embodiments, the update processing of the pre-constructed overall dictionary matrix can be specifically implemented as the steps shown in Figure 17 , where:

[0299] In step 1701: Based on the overall dictionary matrix and the target phase signal, obtain the sparse representation of the target phase signal on the overall dictionary matrix.

[0300] The sparse representation of the target phase signal on the overall dictionary matrix is obtained using Equation 39:

[0301]

[0302] Among them, is the target phase signal, D is the overall dictionary matrix, and X is the sparse representation of the target phase signal on the overall dictionary matrix.

[0303] In step 1702: The orthogonal matching pursuit method is used to update the sparse representation to obtain the updated sparse representation.

[0304] The specific implementation method of this step is the same as that of step 1002, and at this time, τ is equal to 1.

[0305] In step 1703: Based on the updated sparse representation, the singular value decomposition method is used to update the overall dictionary matrix to obtain the updated overall dictionary matrix.

[0306] The specific implementation method of this step is the same as that of step 1003, and will not be elaborated here.

[0307] Step 1604: Based on the target phase signal and the updated overall dictionary matrix, the respiration signal and the heartbeat signal are obtained.

[0308] The specific implementation method of this step is the same as that of step 303, and will not be elaborated here.

[0309] Step 1605: Based on the respiration signal and the preset respiration frequency range, the respiration frequency is obtained, and based on the heartbeat signal and the preset heartbeat frequency range, the heartbeat frequency is obtained.

[0310] The specific implementation method of this step is the same as that of step 304, and will not be elaborated here.

[0311] For the convenience of further understanding, the overall process of a signal detection method provided by the embodiments of the present application is described below, as Figure 18 shown, where:

[0312] In step 1801: A radar signal is sent and an echo signal is received.

[0313] In step 1802: Coherent accumulation is performed on the fast time dimension in the echo signal to obtain the coherently accumulated signal.

[0314] In step 1803: Based on the coherently accumulated signal, a phase extraction operation is performed to obtain phase information.

[0315] In step 1804: The phase information is corrected to obtain the target phase signal.

[0316] In step 1805: The updated overall dictionary matrix based on the training sample set is obtained.

[0317] In step 1806: The overall dictionary matrix is updated to obtain the updated overall dictionary matrix.

[0318] In step 1807: Based on the target phase signal and the updated overall dictionary matrix, the respiration signal and the heartbeat signal are obtained.

[0319] In step 18051: Obtain the pre-constructed overall dictionary matrix and the training sample set.

[0320] In step 18052: Obtain the training sample set, and construct a target phase matrix based on the total phase in the training sample set.

[0321] In step 18053: Based on the overall dictionary matrix and the target phase matrix, obtain the sparse representation of the target phase matrix on the overall dictionary matrix.

[0322] In step 18054: Use the orthogonal matching pursuit method to update the sparse representation to obtain the updated sparse representation.

[0323] In step 18055: Based on the updated sparse representation, use the singular value decomposition method to update the overall dictionary matrix to obtain the updated overall dictionary matrix.

[0324] In step 1808: Based on the respiration signal and the preset respiration frequency range, obtain the respiration frequency, and based on the heartbeat signal and the preset heartbeat frequency range, obtain the heartbeat frequency.

[0325] Based on the same inventive concept, after introducing a signal detection method provided by an embodiment of the present application, as Figure 19 shown, a terminal device control device 1900 provided by an embodiment of the present application is described below. The device includes:

[0326] A receiving module 19001, configured to obtain an echo signal;

[0327] A preprocessing module 19002, configured to preprocess the echo signal to obtain a target phase signal;

[0328] A signal separation module 19003, configured to obtain a respiration signal and a heartbeat signal based on the target phase signal and the pre-constructed overall dictionary matrix obtained.

[0329] Corresponding to the above embodiment, the present application also provides an electronic device. Figure 20 The structure diagram of an electronic device provided by an embodiment of the present invention. The electronic device 2000 may include: a processor 2001, a memory 2002, and a communication unit 2003. A radar component (not shown in the figure) is further included in the electronic device, and the radar component has the ability to transmit and receive radar signals.

[0330] These components communicate via one or more buses. Those skilled in the art can understand that the structure of the electronic device shown in the figure does not limit the embodiments of the present invention. It can be a bus structure, a star structure, or include more or fewer components than shown, or combine certain components, or have a different component arrangement.

[0331] Among them, the communication unit 2003 is used to establish a communication channel so that the electronic device can communicate with other devices, receive user data sent by other devices, or send user data to other devices.

[0332] The processor 2001 is the control center of the electronic device, connecting various parts of the entire electronic device through various interfaces and lines. By running or executing software programs and / or modules stored in the memory 2002, and by invoking the data stored in the memory, it performs various functions of the electronic device and / or processes data. The processor can be composed of an integrated circuit (IC). For example, it can be composed of a single packaged IC, or composed of multiple packaged ICs with the same or different functions connected together. For example, the processor 2001 can include only a central processing unit (CPU). In the embodiments of the present invention, the CPU can be a single arithmetic core or include multiple arithmetic cores.

[0333] The memory 2002 is used to store the execution instructions of the processor 2001. The memory 2002 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, a magnetic disk, or an optical disc.

[0334] When the execution instructions in the memory 2002 are executed by the processor 2001, the electronic device 2000 can execute Figure 3 Some or all of the steps in the illustrated embodiments.

[0335] In a specific implementation, the present invention also provides a computer storage medium. Among them, the computer storage medium can store a program, and when the program is executed, it can include some or all of the steps in the embodiments of the signal detection method provided by the present invention. The storage medium can be a magnetic disk, an optical disc, a read-only memory (ROM), or a random access memory (RAM), etc.

[0336] Those skilled in the art can clearly understand that the technologies in the embodiments of the present invention can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solutions in the embodiments of the present invention, in essence, or the parts that contribute to the prior art can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disc, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of the present invention.

[0337] For the same or similar parts among the various embodiments in this specification, reference can be made to each other. In particular, for the device embodiments and the terminal embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and for the relevant parts, reference can be made to the descriptions in the method embodiments.

Claims

1. A signal detection method, characterized in that, The method includes: Obtaining an echo signal; Preprocessing the echo signal to obtain a target phase signal; Based on the target phase signal and a pre-constructed overall dictionary matrix, obtaining a respiration signal and a heartbeat signal; the overall dictionary matrix is a discontinuous matrix constructed according to the dictionary matrix corresponding to the respiration signal and the dictionary matrix corresponding to the heartbeat signal.

2. The method according to claim 1, wherein The echo signal includes a fast time dimension and a slow time dimension. The preprocessing the echo signal to obtain a target phase signal includes: Performing coherent accumulation on the fast time dimension in the echo signal to obtain a coherently accumulated signal; Performing a phase extraction operation based on the coherently accumulated signal to obtain phase information; Performing a correction process on the phase information to obtain the target phase signal.

3. The method according to claim 2, wherein The performing a correction process on the phase information to obtain the target phase signal includes: Performing a linear regression process on the phase information to obtain the slope and intercept corresponding to the phase information; Performing an amplitude-phase process on the phase information based on the slope and intercept to obtain a processed phase signal; Performing a normalization process on the processed phase signal to obtain a target phase signal.

4. The method according to claim 1, characterized in that, The obtaining a respiration signal and a heartbeat signal based on the target phase signal and a pre-constructed overall dictionary matrix includes: Obtaining the pre-constructed overall dictionary matrix; Based on the overall dictionary matrix and the target phase signal, obtaining a sparse representation of the target phase signal on the overall dictionary matrix; Obtaining the respiration signal based on the overall dictionary matrix and obtaining the heartbeat signal based on the sparse representation.

5. The method according to claim 4, characterized in that, The obtaining the respiration signal based on the overall dictionary matrix includes: Performing a fast Fourier transform on each column in the overall dictionary matrix to obtain the frequency of each column; For the frequency of each column, determining whether the frequency belongs to a preset respiration frequency range; if it belongs, determining the frequency as a first frequency; Constructing a first support set based on the first frequency; Obtaining a respiration signal based on the first support set.

6. The method according to claim 4, characterized in that The obtaining the heartbeat signal based on the sparse representation includes: Performing a fast Fourier transform on each row in the sparse representation to obtain the frequency of each row; For the frequency of each row, determining whether the frequency belongs to a preset heartbeat frequency range; if it belongs, determining the frequency as a second frequency; Constructing a second support set based on the second frequency; Obtaining a heartbeat signal based on the second support set.

7. The method according to claim 1, wherein Before the obtaining a respiration signal and a heartbeat signal by the target phase signal and a pre-constructed overall dictionary matrix, the method further includes: Performing an update process on the pre-constructed overall dictionary matrix to obtain an updated overall dictionary matrix; The obtaining a respiration signal and a heartbeat signal based on the target phase signal and the obtained pre-constructed overall dictionary matrix includes: Obtaining a respiration signal and a heartbeat signal based on the target phase signal and the updated overall dictionary matrix.

8. The method according to claim 7, characterized in that, The performing an update process on the pre-constructed overall dictionary matrix includes: Based on the overall dictionary matrix and the target phase signal, obtain the sparse representation of the target phase signal on the overall dictionary matrix; Use the orthogonal matching pursuit method to update the sparse representation to obtain the updated sparse representation; Based on the updated sparse representation, use the singular value decomposition method to update the overall dictionary matrix to obtain the updated overall dictionary matrix.

9. The method according to claim 1, wherein After obtaining the respiration signal and the heartbeat signal, the method further includes: Based on the respiration signal and a preset respiration frequency range, obtain the respiration frequency, and based on the heartbeat signal and a preset heartbeat frequency range, obtain the heartbeat frequency.

10. The method according to claim 9, wherein The obtaining the respiration frequency based on the respiration signal and a preset respiration frequency range includes: Perform DC filtering on the respiration signal to obtain the filtered respiration signal; Perform fast Fourier transform on the filtered respiration signal to obtain the first respiration frequency; Determine the first respiration frequencies within the preset respiration frequency range as the second respiration frequencies; Take the maximum value among the second respiration frequencies as the respiration frequency.

11. The method according to claim 9, wherein The obtaining the heartbeat frequency based on the heartbeat signal and a preset heartbeat frequency range includes: Perform DC filtering on the heartbeat signal to obtain the filtered heartbeat signal; Perform fast Fourier transform on the filtered heartbeat signal to obtain the first heartbeat frequency; Determine the first heartbeat frequencies within the preset heartbeat frequency range as the second heartbeat frequencies; Take the maximum value among the second heartbeat frequencies as the heartbeat frequency.

12. The method according to any one of claims 1-11, characterized in that, The pre-constructed overall dictionary matrix is constructed according to the following method: Based on a preset respiration frequency range and a preset number of samples, obtain the dictionary matrix corresponding to the respiration signal; Based on a preset heartbeat frequency range and a preset number of samples, obtain the dictionary matrix corresponding to the heartbeat signal; Based on a preset interference frequency range and a preset number of samples, obtain the dictionary matrix corresponding to the interference signal; Based on the dictionary matrix corresponding to the respiration signal, the dictionary matrix corresponding to the heartbeat signal, and the dictionary matrix corresponding to the interference signal, obtain the overall dictionary matrix.

13. The method according to claim 12, wherein After obtaining the overall dictionary matrix, the method further includes: Determine the respiration phase and the heartbeat phase through an accumulator; based on the respiration phase, the heartbeat phase, and a preset interference phase, obtain the total phase; Construct a training sample set based on the total phase; Use the training sample set to update the overall dictionary matrix, and take the updated overall dictionary matrix as the overall dictionary matrix.

14. The method according to claim 1, characterized in that, Before obtaining the echo signal, the method further includes: Transmit a radar signal; the radar signal is used to form an echo signal after being reflected by the human body.

15. A signal detection device, characterized in that, The device includes: A receiving module that obtains an echo signal; A preprocessing module for preprocessing the echo signal to obtain a target phase signal; A signal separation module for obtaining a respiration signal and a heartbeat signal based on the target phase signal and the pre-constructed overall dictionary matrix obtained; the overall dictionary matrix is a discontinuous matrix constructed according to the dictionary matrix corresponding to the respiration signal and the dictionary matrix corresponding to the heartbeat signal.

16. An electronic device, characterized in that, It includes a memory for storing computer program instructions and a processor for executing the program instructions. When the computer program instructions are executed by the processor, the electronic device is triggered to execute the method described in any one of claims 1-14.

17. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program. When the program runs, it controls the device where the computer-readable storage medium is located to execute the method described in any one of claims 1-14.

Images