A method and device for respiration detection based on four-point CSI signal

Abstract

The application provides a four-point CSI signal-based respiration detection method and device, which comprises the following steps: step 1, four CSI signal transmitting and collecting devices are arranged around a measured object, and each device collects a CSI signal containing respiration information of the measured object; step 2, the collected four CSI signals are respectively preprocessed and signal decomposition is completed; step 3, a signal with a human respiration frequency range is determined in the decomposed four signals, respiration frequency signal screening and judgment are further performed, and it is comprehensively determined whether the respiration of the measured object is detected, and if yes, the corresponding respiration frequency is obtained. The application can accurately and efficiently complete human respiration detection through a CSI signal and output the corresponding frequency, and is simple and low in cost.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of breath detection, in particular to a breath detection method and device based on four-point CSI signal. BACKGROUND

[0002] At present, most of the human breath detection based on CSI signal is based on 5300 network card, especially in various levels of academic papers, 5300 network card is used as a CSI signal acquisition device. The signal acquisition unit deployment of most of the breath frequency detection based on CSI signal is: the first, a single CSI acquisition node; the second, multiple acquisition nodes in the room, the nodes are arranged in the corners of the room; the third, a multi-transmit-multiple-receive node array. The present application adopts four nodes to form a WIFI signal transmission-acquisition module to trigger and collect the CSI signal of WIFI, but the spatial layout is different from the general multi-acquisition node method, and the multi-nodes of other methods are usually located in the corners of the room with a distance of several meters. The existing deployment method has the problems of inaccurate detection, low efficiency and large space occupation. SUMMARY

[0003] In view of the problems in the prior art, a breath detection method and device based on four-point CSI signal are provided, which has four-point CSI signal transmission and acquisition devices with small spacing, which is beneficial to the human breath detection by the CSI signal obtained by the nodes.

[0004] The technical scheme adopted by the present application is as follows: a breath detection method based on four-point CSI signal, comprising:

[0005] Step 1, four CSI signal transmission and acquisition devices are arranged around the measured object, and each device collects the CSI signal containing the breath information of the measured object;

[0006] Step 2, the collected four-way CSI signal is preprocessed and signal decomposition is completed;

[0007] Step 3, the signal with the human breath frequency range is determined in the decomposed four-way signal, and then the breath frequency signal is screened and judged, and whether the breath of the measured object is detected is comprehensively determined, and if detected, the corresponding breath frequency is obtained;

[0008] Among them, the spacing between the four CSI signal transmission and acquisition devices is not more than 1 meter, the four CSI signal transmission and acquisition devices are distributed in a rectangular shape, and the measured object is located in the middle of the four CSI signal transmission and acquisition devices.

[0009] As a preferred solution, in step 1, four CSI signal acquisition devices respectively send PING signals to the WIFI router, and the WIFI router replies to each device, so as to acquire the CSI signals; wherein the signal frequency of each CSI signal acquisition device is 25Hz, and the router replies the signal at a frequency of 100Hz.

[0010] As a preferred solution, in step 2, each CSI signal is subjected to 7-layer wavelet decomposition processing, so as to complete signal decomposition and remove interference signals.

[0011] As a preferred solution, the sub-step of step 3 is:

[0012] Step 3.1, Fourier transform is performed on the low-frequency signals in the decomposed four-channel CSI signals, the zero-frequency component is moved to the center of the array, and the energy signals in the range of 0.15-0.45Hz are taken out;

[0013] Step 3.2, whether there is a significant human respiratory frequency in the four-channel energy signals is detected, and the corresponding respiratory frequency value is determined;

[0014] Step 3.3, the respiratory frequency detection results of the four-channel energy signals are comprehensively judged to determine whether the respiration of the measured object is detected, and if so, the respiratory frequency is comprehensively calculated according to the respiratory frequency values determined in each channel energy signal.

[0015] As a preferred solution, the sub-step of step 3.2 is:

[0016] Step 3.2.1, sliding window filtering is performed on the energy signal Yi of the i-th channel to obtain a sliding window signal Zi;

[0017] Step 3.2.2, the maximum value Emaxi and the corresponding frequency Zmaxi and energy average value Emeani of the sliding window signal Zi are obtained;

[0018] Step 3.2.3, whether 0.2<Zmaxi<0.4 and Emaxi>1.5*Emeani are both true is judged, if so, it is considered that there is a human respiratory frequency in the i-th channel CSI signal, and the corresponding respiratory frequency value is Zmaxi.

[0019] As a preferred solution, the sub-step of step 3.3 is:

[0020] Step 3.3.1, the judgment results of each channel in step 3.2 are sorted to obtain (Fi, Zmaxi), wherein Fi is the detection result of the i-th channel signal, taking the value of 0 or 1, 0 indicating that no respiratory frequency is detected, and 1 indicating that respiratory frequency is detected, and Zmaxi is the respiratory frequency value;

[0021] Step 3.3.2, judging whether the breathing of the measured object is detected according to the comprehensive 4-way judgment result:

[0022]

[0023] Breath_Flag = 1 indicates that the comprehensive judgment is that the breathing frequency is detected, otherwise 0, indicating that no breathing is detected;

[0024] Step 3.3.3, when the breathing of the measured object is detected, the corresponding breathing frequency is calculated:

[0025]

[0026] Breath_Rate represents the breathing frequency.

[0027] The present application provides a kind of based on four point CSI signal breathing detection device, comprising:

[0028] Detection seat, for measured object to sit, the four edge corners of detection seat are all installed with CSI signal acquisition device;CSI signal acquisition device is used to collect the CSI signal containing the breathing information of measured object;

[0029] Preprocessing module, for filtering processing to the collected 4-way signal respectively;

[0030] Respiratory frequency signal extraction module, for wavelet decomposition to the filtered signal;

[0031] Signal screening module, for the signal after 4-way wavelet decomposition respectively do frequency domain transformation to obtain frequency energy signal, according to frequency energy and frequency output each way breathing frequency detection result and corresponding breathing frequency value;

[0032] Signal decision module, for judging whether the breathing of measured object is detected according to the breathing frequency detection result output by signal screening module, and according to the breathing frequency value output by signal screening module, the breathing frequency of measured object is comprehensively calculated.

[0033] As a preferred scheme, the distance between the CSI signal acquisition devices on the detection seat is not more than 1 meter, which is distributed in a rectangular shape, and the measured object is located in the middle of the four CSI signal emission and acquisition devices.

[0034] As a preferred scheme, the working process of the signal screening module is:

[0035] The signals after 4-path wavelet decomposition are respectively subjected to Fourier transform, converted into frequency domain energy signals, and the zero frequency components are moved to the center of the array, and the energy signals in the range of 0.15-045 Hz are taken out; the energy signal Yi of the i-th path is subjected to sliding window filtering to obtain the sliding window signal Zi, and the maximum value Emaxi and the corresponding frequency Zmaxi and energy average value Emeani of the sliding window signal Zi are obtained; it is judged whether 0.2<Zmaxi<0.4 and Emaxi>1.5*Emeani are simultaneously established, if yes, it is considered that the i-th path CSI signal contains the human respiratory frequency, and the corresponding respiratory frequency value is Zmaxi.

[0036] As a preferred scheme, the working process of the signal decision module is:

[0037] The 4-path judgment results are integrated to determine whether the respiratory of the measured object is detected:

[0038]

[0039] When Breath_Flag=1, it indicates that the comprehensive judgment is that the respiratory frequency is detected, otherwise it is 0, indicating that the respiratory is not detected;

[0040] When the respiratory of the measured object is detected, the corresponding respiratory frequency is calculated:

[0041]

[0042] Wherein, Breath_Rate represents the respiratory frequency, Fi is the detection result of the i-th path signal, taking the value of 0 or 1, 0 represents that the respiratory frequency is not detected, 1 represents that the respiratory frequency is detected, and Zmaxi is the respiratory frequency value.

[0043] Compared with the prior art, the beneficial effects of the above technical scheme are:

[0044] 1、(Through the reasonable arrangement of the 4 CSI acquisition and transmission devices on the seat position, it is ensured that the Phinery curve of at least one signal is in a relatively good relative position relationship with the human respiratory movement, which helps to improve the effectiveness of the respiratory detection algorithm.

[0045] 2、The steps in the respiratory detection algorithm, especially the wavelet signal decomposition layer number, respiratory signal acquisition and frequency calculation, are associated with the CSI signal transmission and acquisition frequency and the detection target, and are a complete whole.

[0046] 3、The signal screening and decision take into account the sensitivity and robustness, as long as there is one path signal to effectively detect the respiratory signal, it is considered that there is a respiratory signal, which also corresponds to the reasonable arrangement of the 4 points.

[0047] 4. The scheme is logically clear, low in implementation cost, and easy to mass produce. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 A flow chart of the breath detection method according to the present application.

[0049] Figure 2 A schematic diagram of the signal acquisition device according to an embodiment of the present application.

[0050] Figure 3 A schematic diagram of the breath detection device according to an embodiment of the present application.

[0051] Figure 4 A schematic diagram of the signal acquisition device installation according to an embodiment of the present application. DETAILED DESCRIPTION

[0052] Embodiments of the present application are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar notations represent the same or similar modules or modules with the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application, and cannot be understood as a limitation of the present application. On the contrary, the embodiments of the present application include all changes, modifications and equivalents within the spirit and scope of the appended claims.

[0053] Embodiment 1

[0054] As shown in the figure, the present embodiment provides a breath detection method based on four-point CSI signals, comprising: Figure 1

[0055] Step 1, four CSI signal emission and acquisition devices are arranged around the measured object, each device acquires CSI signals containing the breath information of the measured object;

[0056] Step 2, the collected four-way CSI signals are respectively preprocessed and signal decomposition is completed;

[0057] Step 3, the signal with the human breath frequency range is determined in the decomposed four-way signal, and then the breath frequency signal is screened and judged, and whether the breath of the measured object is detected is comprehensively determined, and if detected, the corresponding breath frequency is obtained;

[0058] Among them, the distance between the four CSI signal emission and acquisition devices is not more than 1 meter, which is distributed in a rectangular shape, and the measured object is located in the middle of the four CSI signal emission and acquisition devices.

[0059] ​In the embodiment, the 4 CSI signal acquisition devices in step 1 form the signal acquisition module of the whole scheme, wherein the CSI signal acquisition device is an ESP32 unit, which is different from the 5300 network card in the existing scheme, and the ESP32 unit has significant differences in acquisition rate and data characteristics. As shown in Figure 2 each ESP32 can independently receive and transmit WIFI signals, they send PING signals to the WIFI router, prompting the WIFI router to reply to the ESP32 unit, thereby capturing the CSI frame.

[0060] Since the sampling frequency of the signal acquisition module needs to reach 100 Hz to achieve respiratory rate detection, in the embodiment, the 4 ESP32 units respectively send signals with a frequency of 25 Hz to the WIFI router, and the 4 ESP32 units add up to 100 Hz; the router sends response signals at a frequency of 100 Hz, and since WIFI wireless signals have broadcast nature, at this time, for the 4 ESP32 units, the received signals all have the frequency characteristic of 100 Hz, that is, the frequency condition for respiratory detection.

[0061] Since the 4 ESP32 units in the embodiment are in the middle of the measured object, the feedback signal contains human body information, and human activity information causes the signal propagation path to change, and the greater the activity, the more obvious the change. Human respiration belongs to weak motion and is not easy to detect, and the Fresnel characteristic curve of the radio signal needs to be combined to amplify the signal difference caused by the motion.

[0062] In the embodiment, the 4 ESP32 units are deployed on the armrests of the seat, as shown in Figure 4 , so that they are in close proximity to each other, and the measured object is in the middle of the 4 points, no matter what direction the human body is in, the Fresnel curve of some or some points and the weak respiratory motion are tangent, and at the same time, the distance from the human body is close, thereby amplifying the influence of respiratory motion on the signal, which is helpful for subsequent extraction and processing of respiratory characteristics.

[0063] The normal respiratory rate of the human body is 16-20 times per minute, and the embodiment appropriately widens the range to about 12-24 times per minute, that is, 0.2-0.4 Hz. The frequency of the 4 CSI signals obtained is about 100 Hz, in order to extract the respiratory signal of 0.2-0.4 Hz from the 100 Hz CSI signal, the embodiment uses 7-layer wavelet transform for processing, and the approximate range of the low-frequency signal and the high-frequency detail signal corresponding to each layer is shown in Table 1:

[0064] Table 1 shows the results of wavelet decomposition of CSI

[0065] Number of layers Low frequency signal frequency High frequency detail signal frequency 1 0-50 Hz 50-100 Hz 2 0-25 Hz 25-50 Hz 3 0-12.5 Hz 12.5-25 Hz 4 0-6.25 Hz 6.25-12.5 Hz 5 0-3.125 Hz 3.125-6.25 Hz 6 0-1.56 Hz 1.56-3.125 Hz 7 0-0.78 Hz 0.78-1.56 Hz

[0066] After 7-layer wavelet decomposition of the CSI signal, most of the interference signals are removed in the low-frequency signal components, which can be used for subsequent respiratory frequency analysis. As preferred, the wavelet basis used in this embodiment is db4.

[0067] It should be noted that in this embodiment, data preprocessing including Hampel filtering is required before wavelet decomposition of the CSI signal.

[0068] After wavelet decomposition, a signal close to the respiratory frequency range has been obtained. However, due to the complexity and uncertainty of key points such as the signal itself, motion amplitude and wavelet transform, the wavelet transform may not necessarily reflect the human respiratory frequency. Therefore, the 4 low-frequency signals after wavelet transform need to be screened and judged to determine whether there is a respiratory signal and the approximate frequency of respiration if there is. The specific process is as follows:

[0069] Step 3.1, Fourier transform is performed on the low-frequency signals of the 4 decomposed CSI signals, respectively, to convert them into frequency domain energy signals, and the zero-frequency components are moved to the center of the array, and the energy signals in the range of 0.15-0.45 Hz are taken out;

[0070] In this embodiment, Yi = SELECT(fftshift(abs(fft(Xi)))) represents the energy signal;

[0071] Where Xi is the low-frequency time series amplitude signal of the i-th output, fft is the Fourier transform, abs takes the modulus of a complex number, and fftshift is the Fourier shift frequency transform. This patent only focuses on human respiratory signals, and considering further retention of redundancy, the SELECT operation takes signals between 0.15-0.45 Hz, and the result Yi is the energy signal in the 4 low-frequency regions between 0.15-0.45 Hz.

[0072] Step 3.2, detect whether there is a significant human respiratory frequency in the 4 energy signals and determine the corresponding respiratory frequency value; specifically:

[0073] In order to more accurately obtain the detection result, in step 3.2.1, the energy signal Yi is subjected to sliding window filtering to further eliminate interference noise,

[0074] In this embodiment, the sliding window length is 7, and a relatively smoother sliding window signal Zi = WIN_FILTER(Yi) is obtained.

[0075] Step 3.2.2, after obtaining the sliding window signal, calculate the maximum value Emaxi and its corresponding frequency Zmaxi, the energy average value Emeani.

[0076] In this embodiment, it is considered that Zmaxi in the range of 0.15-0.45 Hz may correspond to the human respiratory signal, it may also be a false interference signal, and it may also be that there is no human respiratory signal in the range. Therefore, whether Zmaxi is a human respiratory signal frequency needs to be further judged.

[0077] In step 3.3.2, the obtained parameters are judged as follows:

[0078] Condition 1: Zmaxi itself appears in the common range of human respiratory frequency, that is, the following formula is established:

[0079] 0.2<Zmaxi<0.4

[0080] Condition 2: The energy Emax corresponding to Zmaxi is particularly significant, that is, the following formula is established:

[0081] Emaxi>1.5*Emeani

[0082] If conditions 1 and 2 are met at the same time, it is considered that there is a human respiratory frequency in the ith CSI signal, and Zmaxi is the corresponding respiratory frequency value.

[0083] Finally, considering the weakness of the human respiratory signal and the complexity brought by it, not all CSI signal sources can detect the respiratory frequency. Therefore, a data fusion method is used to make a comprehensive decision. Specifically:

[0084] In this embodiment, (Fi, Zmaxi) is used to represent the detection result of each CSI signal, where i is the ith signal detection result, taking the value of 0 or 1, 0 indicating that no respiratory frequency is detected, and 1 indicating that a respiratory frequency is detected, and Zmaxi is the corresponding respiratory frequency value. The decision process is as follows:

[0085] (1) Decide whether to detect the subject's breath:

[0086]

[0087] When Breath_Flag=1, it indicates that the comprehensive decision is that the respiratory frequency is detected, otherwise it is 0, indicating that no breath is detected.

[0088] (2) When the subject's breath is detected, the corresponding respiratory frequency is calculated:

[0089]

[0090] Wherein, Breath_Rate represents the respiratory frequency.

[0091] At this point, the subject's breath detection and respiratory frequency detection are completed.

[0092] Embodiment 2

[0093] As Figure 3 shown, the embodiment proposes a four-point CSI signal-based respiration detection device, comprising:

[0094] A detection seat for a subject to sit on, and four edge corners of the detection seat are each provided with a CSI signal acquisition device; the CSI signal acquisition device is configured to acquire a CSI signal containing respiration information of the subject;

[0095] A preprocessing module configured to perform filtering processing on the acquired four signals respectively;

[0096] A respiration frequency signal extraction module configured to perform wavelet decomposition on the filtered signals;

[0097] A signal screening module configured to perform frequency domain transformation on the four wavelet-decomposed signals to obtain frequency energy signals, and output respiration frequency detection results and corresponding respiration frequency values according to the frequency energy and frequency;

[0098] A signal decision module configured to comprehensively judge whether to detect the respiration of the subject according to the respiration frequency detection results output by the signal screening module, and comprehensively calculate the respiration frequency of the subject according to the respiration frequency values output by the signal screening module.

[0099] In the embodiment, the CSI signal acquisition devices on the detection seat are spaced apart by no more than 1 meter, are distributed in a rectangular shape, and the subject is located in the middle of the four CSI signal emission and acquisition devices.

[0100] The CSI signal acquisition device is an ESP32 unit, which is different from the 5300 network card in the prior art, and the ESP32 unit has significant differences in acquisition rate and data characteristics. As Figure 2 shown, each ESP32 can independently receive and transmit WIFI signals, and they send PING signals to the WIFI router to prompt the WIFI router to reply to the ESP32 unit, thereby capturing the CSI frame.

[0101] Since the sampling frequency of the signal acquisition module needs to reach 100 Hz to achieve respiration frequency detection, in the embodiment, the four ESP32 units respectively send signals with a frequency of 25 Hz to the WIFI router, and the four ESP32 units add up to 100 Hz; the router sends a response signal with a frequency of 100 Hz, and since the WIFI wireless signal has a broadcast nature, at this time, the signals received by the four ESP32 units all have a frequency characteristic of 100 Hz, i.e., they have the frequency condition for respiration detection.

[0102] The preprocessing module implements data preprocessing including Hampel filtering.

[0103] The normal breathing frequency of human body is 16-20 times per minute, and the embodiment appropriately broadens the range to about 12-24 times per minute, i.e. 0.2-0.4 Hz. The frequency of the acquired 4-channel CSI signal is about 100 Hz, and in order to extract the breathing signal of 0.2-0.4 Hz from the 100 Hz CSI signal, the embodiment uses 7-layer wavelet decomposition to complete signal decomposition by means of the breathing frequency signal extraction module, and the low-frequency signal component after decomposition has removed most of the interference signals and can be used for subsequent breathing frequency analysis. As a preferred embodiment, the wavelet base used in the embodiment is db4.

[0104] After wavelet decomposition, the signal close to the breathing frequency interval has been obtained, but due to the complexity and uncertainty of key points such as the signal itself, motion amplitude and wavelet transform, the wavelet transform may not necessarily reflect the breathing frequency of the human body. Therefore, the 4-channel low-frequency signals after wavelet transform need to be screened and judged to determine whether there is a breathing signal, and if so, the approximate frequency of the breathing is determined. In the embodiment, the signal screening module and the signal judgment module are used to achieve the above, and the details are as follows:

[0105] The working process of the signal screening module is as follows:

[0106] The Fourier transform is performed on the 4-channel signals after wavelet decomposition to convert them into frequency energy signals, and the zero-frequency component is moved to the center of the array, and the energy signals in the range of 0.15-0.45 Hz are taken out; the sliding window filtering is performed on the energy signal Yi of the i-th channel to obtain the sliding window signal Zi, and the maximum value Emaxi and the corresponding frequency Zmaxi and the energy average value Emeani of the sliding window signal Zi are obtained; it is judged whether 0.2<Zmaxi<0.4 and Emaxi>1.5*Emeani are established at the same time, if so, it is considered that the i-th channel CSI signal contains the breathing frequency of the human body, and the corresponding breathing frequency value is Zmaxi.

[0107] The working process of the signal judgment module in the embodiment is as follows:

[0108] The judgment results of the 4 channels are combined to determine whether the breathing of the measured object is detected:

[0109]

[0110] When Breath_Flag=1, it indicates that the comprehensive judgment is that the breathing frequency is detected, otherwise it is 0, indicating that the breathing is not detected;

[0111] When the breathing of the measured object is detected, the corresponding breathing frequency is calculated:

[0112]

[0113] Wherein, Breath_Rate represents the breathing rate, Fi is the detection result of the ith signal, and the value is 0 or 1, 0 represents that the breathing rate is not detected, and 1 represents that the breathing rate is detected, and Zmaxi is the breathing rate value.

[0114] It should be noted that in the description of the embodiments of the present application, unless otherwise explicitly specified and limited, the terms "set", "connected" should be understood broadly, for example, can be fixedly connected, can also be detachably connected, or integrally connected; can be directly connected, or indirectly connected through an intermediate medium. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances; the drawings in the embodiments are used to clearly and completely describe the technical solutions in the embodiments of the present application, obviously, the described embodiments are part of the embodiments of the present application, not all the embodiments. The components of the embodiments of the present application described and shown in the drawings herein can be arranged and designed in various different configurations.

[0115] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments within the scope of the present application.

Claims

1. A respiratory detection method based on four-point CSI signals, characterized in that, include: Step 1: Set up 4 CSI signal transmitting and acquiring devices around the subject, and each device will collect CSI signals containing the subject's respiratory information. Step 2: Preprocess the four acquired CSI signals and perform signal decomposition. Step 3: Determine the signal with the human respiratory frequency range from the four decomposed signals, then filter and judge the respiratory frequency signals, and comprehensively determine whether the breathing of the subject is detected. If it is detected, obtain the corresponding respiratory frequency. The four CSI signal transmitting and acquiring devices are spaced no more than 1 meter apart, arranged in a rectangular pattern, and the object under test is located in the middle of the four CSI signal transmitting and acquiring devices. In step 1, the four CSI signal acquisition devices send PING signals to the WIFI router, and the WIFI router responds to each device, thereby acquiring the CSI signal; wherein, the signal frequency emitted by each CSI signal acquisition device is 25Hz, and the router responds with a signal at a frequency of 100Hz. The sub-steps of step 3 are as follows: Step 3.1: Perform Fourier transform on the low-frequency signals in the four decomposed CSI signals to convert them into frequency domain energy signals, and move the zero-frequency component to the center of the array to extract the energy signals in the range of 0.15-045Hz; Step 3.2: Detect whether there is a significant human respiratory frequency in the four energy signals and determine the corresponding respiratory frequency value; Step 3.3: Make a comprehensive judgment on the human respiratory frequency detection results of the four energy signals to determine whether the breathing of the tested object is detected. If it is detected, calculate the respiratory frequency based on the respiratory frequency value determined in each energy signal. The sub-steps of step 3.2 are as follows: Step 3.2.1: Perform sliding window filtering on the energy signal Yi of the i-th channel to obtain the sliding window signal Zi; Step 3.2.2: Obtain the maximum value Emaxi of the sliding window signal Zi and its corresponding frequency Zmaxi and average energy value Emeani; Step 3.2.3: Determine if 0.2 is true. <Zmaxi< 0.4、Emaxi> 1.5 * Emeani. If both conditions are met, then the i-th CSI signal is considered to contain human respiratory frequency and the corresponding respiratory frequency value is Zmaxi. The sub-steps of step 3.3 are as follows: Step 3.3.1: Organize the judgment results of each channel in step 3.2 to obtain (Fi, Zmaxi), where Fi is the detection result of the i-th signal, taking a value of 0 or 1, where 0 indicates that no respiratory rate was detected and 1 indicates that a respiratory rate was detected, and Zmaxi is the respiratory rate value; Step 3.3.2: Combine the judgment results of the four channels to determine whether the breathing of the tested object was detected: When Breath_Flag=1, it indicates that a respiratory rate has been detected; otherwise, it is 0, indicating that no breathing has been detected. Step 3.3.3: When breathing is detected in the subject, calculate the corresponding respiratory rate: Breath_Rate represents the respiratory rate.

2. The respiratory detection method based on four-point CSI signals according to claim 1, characterized in that, In step 1, the four CSI signal acquisition devices send PING signals to the WIFI router, and the WIFI router responds to each device, thereby acquiring the CSI signal; wherein, the signal frequency emitted by each CSI signal acquisition device is 25Hz, and the router responds with a signal at a frequency of 100Hz.

3. A respiratory detection device based on four-point CSI signals, characterized in that, include: The testing seat is used for the object to be tested to sit on. CSI signal acquisition devices are installed at all four corners of the testing seat. The CSI signal acquisition device is used to acquire CSI signals containing respiratory information of the subject being tested; The preprocessing module is used to filter the four acquired signals respectively. The respiratory rate signal extraction module is used to perform wavelet decomposition on the filtered signal; The signal filtering module is used to perform frequency domain transformation on the four wavelet decomposed signals to obtain frequency domain energy signals, and output the respiratory frequency detection results and corresponding respiratory frequency values ​​of each channel according to the frequency domain energy and frequency. The signal decision module is used to comprehensively determine whether to detect the breathing of the subject based on the respiratory rate detection results output by the signal filtering module, and to comprehensively calculate the respiratory rate of the subject based on the respiratory rate value output by the signal filtering module. The CSI signal acquisition devices on the testing seat are spaced no more than 1 meter apart, arranged in a rectangular pattern, and the object under test is located in the middle of the four CSI signal transmission and acquisition devices. The signal filtering module operates as follows: Perform Fourier transforms on the four wavelet decomposition signals to convert them into frequency domain energy signals. Move the zero-frequency component to the center of the array and extract the energy signal within the range of 0.15-0.45Hz. Perform sliding window filtering on the i-th energy signal Yi to obtain the sliding window signal Zi. Obtain the maximum value Emaxi of the sliding window signal Zi and its corresponding frequency Zmaxi, as well as the average energy value Emeani. Determine if 0.2... <Zmaxi< 0.4、Emaxi> 1.5 * Emeani. If both conditions are met, then the i-th CSI signal is considered to contain human respiratory frequency and the corresponding respiratory frequency value is Zmaxi. The working process of the signal decision module is as follows: The determination of whether the subject's breathing was detected is based on the combined results of the four assessments: When Breath_Flag=1, it indicates that a respiratory rate has been detected by the overall judgment; otherwise, it is 0, indicating that no breathing has been detected. When breathing is detected in the subject, the corresponding respiratory rate is calculated: Where Breath_Rate represents the respiratory rate, Fi is the detection result of the i-th signal, and the value is 0 or 1, where 0 indicates that no respiratory rate was detected and 1 indicates that the respiratory rate was detected, and Zmaxi is the respiratory rate value.

Images