Static user respiration detection method and apparatus, electronic device, and storage medium

Abstract

The present application relates to the technical field of signal analysis, and particularly relates to a static user breathing detection method and device, electronic equipment and storage medium. The method comprises the following steps: firstly, a plurality of channel state information data sets are acquired; then, data decomposition and reconstruction are performed on the plurality of channel state information data sets respectively to obtain a plurality of reconstructed data sets; then, a plurality of high-sensitivity reconstructed data sets are selected from the plurality of reconstructed data sets, and the plurality of high-sensitivity reconstructed data sets are fused to obtain a fused data set, wherein the periodicity of the high-sensitivity reconstructed data set corresponds to a human breathing frequency range; finally, the breathing frequency is determined according to the fused data set. In the data preprocessing part, the present application adds an environment detection part to extract a feature vector from the CSI signal to exclude environmental interference. In the high-frequency noise filtering part, the present application proposes an improved algorithm to solve the problems of excessive residual noise and the existence of pseudo modes, so as to better reconstruct the breathing signal and filter out high-frequency noise.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of signal analysis, and in particular to a static user breathing detection method and device, electronic equipment and a storage medium. BACKGROUND

[0002] Breathing detection usually requires wearing expensive equipment, which may have the risk of invasion and privacy leakage. In contrast, non-contact methods are more attractive. The latest research in this field explores the possibility of applying ubiquitous WiFi signals to breathing monitoring without installing equipment on the target.

[0003] Wi-Fi has been widely covered and deployed in various regions across the country. Using Wi-Fi signals to detect breathing of static target users not only becomes an important research field in the academic community, but also becomes a necessary technology to realize smart cities. Researchers have mainly used RSS-based breathing detection in Wi-Fi in the early stage. Later studies found that collecting and processing channel state information (CSI) in Wi-Fi signals can achieve passive human perception, which can more finely perceive the weak fluctuations of signals on the line-of-sight or non-line-of-sight path than RSS. Wi-Sleep first started using wireless signals for breathing perception. It detects people's sleep by extracting the rhythm pattern related to breathing from CSI. In subsequent studies, a new rhythm estimation algorithm was proposed, which can achieve breathing frequency estimation in different sleep postures. Wang et al. described the Fresnel domain model in the WiFi signal in detail, established a theoretical model of human breathing and wireless signal changes. Yang et al. combined the Fresnel domain model, modeled the breathing movement of the human body during sleep as a Gaussian model, converted the posture change into vertical displacement, optimized the transceiver deployment, and realized sleep breathing detection.

[0004] However, since CSI breathing perception is in the fine-grained category, it is very important to complete accurate vital sign detection of static target users through CSI data, and the existing literature's breathing detection model still has certain room for improvement. This paper will study the human static breathing monitoring based on WiFi all-in-one, and propose a feasible breathing detection model and algorithm to realize the feasibility and stability of breathing detection, which can more effectively face more complex scenarios.

[0005] Therefore, it is necessary to develop and design a static user breathing detection method. SUMMARY

[0006] The embodiment of the present application provides a static user breathing detection method, device, electronic equipment and storage medium, and is used for solving the problem of poor stability of breathing detection in the prior art.

[0007] In a first aspect, the embodiment of the present application provides a static user breathing detection method, comprising:

[0008] Obtaining a plurality of channel state information data sets;

[0009] Respectively performing data decomposition and reconstruction on the plurality of channel state information data sets to obtain a plurality of reconstructed data sets, wherein the plurality of reconstructed data sets correspond to the plurality of channel state information data sets, and a reconstructed data set is obtained by reconstruction according to a first component data set in a human respiratory frequency range, and the first component data set is obtained based on data decomposition of the channel state information data set;

[0010] Selecting a plurality of high-sensitivity reconstructed data sets from the plurality of reconstructed data sets, and fusing the plurality of high-sensitivity reconstructed data sets to obtain a fused data set, wherein the periodicity of a high-sensitivity reconstructed data set corresponds to the human respiratory frequency range;

[0011] Determining a respiratory frequency according to the fused data set.

[0012] In a possible implementation manner, the respectively performing data decomposition and reconstruction on the plurality of channel state information data sets to obtain a plurality of reconstructed data sets comprises:

[0013] For each data set in the plurality of channel state information data sets, the following steps are performed:

[0014] Removing outliers deviating from an original signal track by a filtering algorithm to obtain a first data set;

[0015] Decomposing the first data set by a plurality of time domain signal decomposition manners to obtain a plurality of first component data sets corresponding to a plurality of frequencies;

[0016] Selecting a plurality of data sets from the plurality of first component data sets as a plurality of second component data sets according to a human respiratory frequency range and a frequency corresponding to the first component data set;

[0017] Reconstructing according to the plurality of second component data sets to obtain a reconstructed data set.

[0018] In a possible implementation manner, the decomposing the first data set by a plurality of time domain signal decomposition manners to obtain a plurality of first component data sets corresponding to a plurality of frequencies comprises:

[0019] Obtaining a plurality of first Gaussian white noise arrays, wherein the first Gaussian white noise array comprises Gaussian white noise data that are opposite to each other.

[0020] performing empirical mode decomposition on the plurality of first Gaussian white noise arrays respectively to obtain a plurality of component values, wherein the component value is an average of two intrinsic mode component data, and the two intrinsic mode component data are obtained based on the empirical mode decomposition of the first Gaussian white noise array;

[0021] adding the plurality of component values to the first data set respectively to obtain a third data set;

[0022] calculating a set average value of the third data set to obtain a residual signal data set and a modal component data set, and taking the modal component data set as a first component data set;

[0023] if the residual signal data set is not monotonic, taking the residual signal data set as a first data set, performing empirical mode decomposition on the plurality of component values, taking a plurality of results obtained by the empirical mode decomposition as a plurality of component values, and jumping to the step of adding the plurality of component values to the first data set respectively to obtain a third data set.

[0024] In a possible implementation manner, the calculating a set average value of the third data set to obtain a residual signal data set and a modal component data set comprises:

[0025] determining the residual signal data set according to the third data set and a first formula, wherein the first formula is:

[0026]

[0027] wherein, r i (k+1) is a residual signal data set obtained in the k+1th iteration, M i is a set average value calculation function, ε (k+1) is a coefficient in the k+1th iteration, IMF (k+1) (n i ) is the i th component value in the k+1th iteration;

[0028] determining the modal component data set according to the residual signal data set and a second formula, wherein the second formula is:

[0029] IMF (k+1) = r (k) -r (k+1)

[0030] wherein, IMF (k+1) is the modal component data set.

[0031] In a possible implementation, the selecting a plurality of high-sensitivity reconstruction data sets from the plurality of reconstruction data sets and fusing the plurality of high-sensitivity reconstruction data sets to obtain a fused data set comprises:

[0032] performing Fourier transform on the plurality of reconstruction data sets respectively to obtain a plurality of frequency domain data sets, wherein the plurality of frequency domain data sets correspond to the plurality of reconstruction data sets;

[0033] determining the plurality of high-sensitivity reconstruction data sets from the plurality of reconstruction data sets according to a human respiratory frequency range and the plurality of frequency domain data sets;

[0034] removing direct current components in the plurality of high-sensitivity reconstruction data sets;

[0035] combining the plurality of high-sensitivity reconstruction data sets from which the direct current components are removed to obtain the fused data set.

[0036] In a possible implementation, the determining the respiratory frequency according to the fused data set comprises:

[0037] determining a plurality of peak values according to the fused data set;

[0038] obtaining a sliding window;

[0039] taking a plurality of data from the fused data set by using the sliding window;

[0040] judging whether a peak value in the plurality of taken data is a true peak value, wherein the true peak value is the maximum in the plurality of taken data;

[0041] filtering out false peak values in the plurality of taken data;

[0042] if the sliding window does not reach the end of the fused data set, sliding the sliding window and jumping to the step of taking the plurality of data from the fused data set by using the sliding window;

[0043] determining a plurality of frequency components by using the plurality of peak values;

[0044] calculating an average value of the plurality of frequency components as the respiratory frequency.

[0045] In a possible implementation, before the performing data decomposition and reconstruction on the plurality of channel state information data sets respectively, the method comprises:

[0046] performing normalization processing on the plurality of channel state information data sets by using a tangent function;

[0047] performing clustering on the plurality of channel state information data sets after the normalization processing, to determine whether there is environmental interference;

[0048] If there is no environmental interference, the filtering algorithm is used to remove outliers deviating from the original signal trajectory from the plurality of channel state information data sets.

[0049] In a second aspect, embodiments provide a static user respiration detection device for implementing the static user respiration detection method according to the first aspect or any possible implementation of the first aspect, the static user respiration detection device comprising:

[0050] a data acquisition module configured to acquire a plurality of channel state information data sets;

[0051] a data reconstruction module configured to perform data decomposition and reconstruction on the plurality of channel state information data sets respectively to obtain a plurality of reconstructed data sets, wherein the plurality of reconstructed data sets correspond to the plurality of channel state information data sets, and a reconstructed data set is obtained by reconstruction according to a first component data set in a human respiration frequency range, and the first component data set is obtained based on data decomposition of the channel state information data set;

[0052] a data fusion module configured to select a plurality of high-sensitivity reconstructed data sets from the plurality of reconstructed data sets, and fuse the plurality of high-sensitivity reconstructed data sets to obtain a fused data set, wherein a periodicity of a high-sensitivity reconstructed data set corresponds to the human respiration frequency range;

[0053] and,

[0054] a respiration frequency determination module configured to determine a respiration frequency according to the fused data set.

[0055] In a third aspect, embodiments provide an electronic device, comprising a memory and a processor, wherein the memory stores a computer program capable of running on the processor, and the processor implements the steps of the method according to the first aspect or any possible implementation of the first aspect when executing the computer program.

[0056] In a fourth aspect, embodiments provide a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program implements the steps of the method according to the first aspect or any possible implementation of the first aspect when executed by a processor.

[0057] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0058] The embodiment of the application discloses a static user breathing detection method, which comprises the following steps: firstly, a plurality of channel state information data sets are acquired; then, the plurality of channel state information data sets are respectively subjected to data decomposition and reconstruction to obtain a plurality of reconstructed data sets, wherein the plurality of reconstructed data sets correspond to the plurality of channel state information data sets, and a reconstructed data set is obtained according to a first component data set of a human breathing frequency range, and the first component data set is obtained based on data decomposition of the channel state information data set; then, a plurality of high-sensitivity reconstructed data sets are selected from the plurality of reconstructed data sets, and the plurality of high-sensitivity reconstructed data sets are fused to obtain a fused data set, wherein the periodicity of the high-sensitivity reconstructed data set corresponds to the human breathing frequency range; finally, the breathing frequency is determined according to the fused data set. The embodiment of the application adds an environment detection part in the data preprocessing part, extracts a feature vector from a CSI breathing signal to exclude environmental interference; in the high-frequency noise filtering part, the embodiment of the application proposes an improved EEMD algorithm to solve the problems of excessive residual noise and the existence of pseudo modes, and further better reconstructs the breathing signal and filters the high-frequency noise. BRIEF DESCRIPTION OF DRAWINGS

[0059] In order to more clearly illustrate the technical solutions in the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute some embodiments of the application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0060] Figure 1 It is a flow chart of the static user breathing detection method provided by the embodiment of the application.

[0061] Figure 2 It is a general model structure diagram of the static user breathing detection provided by the embodiment of the application.

[0062] Figure 3 It is a decomposition diagram based on the improved EEMD algorithm provided by the embodiment of the application.

[0063] Figure 4 It is a breathing signal diagram reconstructed based on the EEMD algorithm provided by the embodiment of the application.

[0064] Figure 5 It is a residual white noise comparison diagram provided by the embodiment of the application.

[0065] Figure 6 It is a functional block diagram of the static user breathing detection device provided by the embodiment of the application.

[0066] Figure 7 It is a functional block diagram of the electronic device provided by the embodiment of the application. DETAILED DESCRIPTION

[0067] In the following description, specific details such as particular system structures and techniques are provided for purposes of illustration, not limitation, to facilitate a thorough understanding of the embodiments of the present invention. However, it will be apparent to those skilled in the art that the present invention may be practiced in alternative embodiments without these specific details. In other instances, detailed descriptions of well-known systems, devices, and methods are omitted so as not to obscure the description of the present invention with unnecessary detail.

[0068] In order to make the objectives, technical solutions and advantages of the present invention more clear, the following will be described through specific implementation methods in conjunction with the accompanying drawings.

[0069] The following is a detailed description of an embodiment of the present invention. This example is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process, but the protection scope of the present invention is not limited to the following embodiment.

[0070] Figure 1 This is a flow chart of a static user breathing detection method provided by an embodiment of the present invention.

[0071] like Figure 1 As shown, it shows a flowchart of the implementation method of the static user breathing detection method provided by an embodiment of the present invention, which is detailed as follows:

[0072] In step 101, a plurality of channel state information data sets are obtained.

[0073] For example, using ubiquitous WiFi signals for respiratory monitoring is extremely valuable. However, CSI respiratory perception is fine-grained and easily overwhelmed by the external environment, making it difficult to capture CSI respiratory signals. To accurately extract CSI respiratory information, the present invention integrates WiFi and synaesthesia to perform static human respiratory monitoring, establishes a more robust and stable vital sign monitoring system model, detects human activity in indoor environments and monitors respiratory status, and completes research on respiratory signal acquisition, data preprocessing, and respiratory rate estimation, thereby improving perception sensitivity, expanding the perception area, and enhancing perception reliability.

[0074] The overall architecture of the respiratory monitoring research based on WiFi synaesthesia integration is shown in the figure. It is mainly divided into four stages, namely data extraction, data preprocessing, subcarrier selection, and respiratory signal extraction.

[0075] The core idea of the present application is to provide a static user breathing detection method based on CSI, and different from other breathing detection models, the embodiment of the present application further adds an environment detection part in the data preprocessing part, uses the two characteristics of dynamic path change and signal periodicity to extract a feature vector, and performs normalization processing, so as to exclude environmental interference and improve the detection efficiency; in the high-frequency noise filtering part, the embodiment of the present application proposes an improved EEMD (Ensemble Empirical Mode Decomposition, Ensemble Empirical Mode Decomposition) algorithm to solve the problems of excessive residual noise and existence of pseudo-mode, and further better reconstructs the breathing signal and filters the high-frequency noise.

[0076] In the data extraction stage, the embodiment of the present application uses a WiFi sending and receiving device to obtain the CSI measurement sequence of the target in an indoor environment.

[0077] First, a system for collecting CSI raw data is built, which is composed of a host computer equipped with three antennas and an Intel 5300 wireless network card and a router supporting IEEE 802.11n protocol. The router serves as a sending device; the host computer serves as a receiving device, and the CSI Tool software is installed under the Ubuntu system. The host computer is connected to the network emitted by the router, then performs Ping operation on the IP address of the router at a certain period, and the host computer additionally opens a terminal to use the log_to_file command to obtain CSI data and save it in a dat file.

[0078] In step 102, the plurality of channel state information data sets are respectively subjected to data decomposition and reconstruction to obtain a plurality of reconstructed data sets, wherein the plurality of reconstructed data sets correspond to the plurality of channel state information data sets, and the reconstructed data set is obtained by reconstructing a first component data set according to a human respiratory frequency range, and the first component data set is obtained based on data decomposition of the channel state information data set.

[0079] In some embodiments, the step 102 includes:

[0080] For each data set in the plurality of channel state information data sets, the following steps are performed:

[0081] Abnormal values deviating from the original signal trajectory are removed by a filtering algorithm to obtain a first data set;

[0082] The first data set is decomposed by multiple time domain signal decomposition to obtain a plurality of first component data sets corresponding to a plurality of frequencies;

[0083] selecting a plurality of data sets from the plurality of first component data sets as a plurality of second component data sets according to a human respiratory frequency range and a frequency corresponding to the first component data sets;

[0084] reconstructing according to the plurality of second component data sets to obtain a reconstructed data set.

[0085] In some embodiments, the decomposing the first data set to obtain a plurality of first component data sets corresponding to a plurality of frequencies comprises:

[0086] obtaining a plurality of first Gaussian white noise arrays, wherein the first Gaussian white noise array comprises Gaussian white noise data that are opposite to each other;

[0087] respectively performing empirical mode decomposition on the plurality of first Gaussian white noise arrays to obtain a plurality of component values, wherein the component value is an average value of two intrinsic mode component data, and the two intrinsic mode component data are obtained based on the empirical mode decomposition of the first Gaussian white noise array;

[0088] adding the plurality of component values to the first data set respectively to obtain a third data set;

[0089] calculating a set average value of the third data set to obtain a residual signal data set and a modal component data set, and taking the modal component data set as the first component data set;

[0090] If the residual signal data set is not monotonic, taking the residual signal data set as the first data set, performing empirical mode decomposition on the plurality of component values, taking a plurality of results obtained by the empirical mode decomposition as the plurality of component values, and jumping to the step of adding the plurality of component values to the first data set respectively to obtain a third data set.

[0091] In some embodiments, the calculating a set average value of the third data set to obtain a residual signal data set and a modal component data set comprises:

[0092] determining the residual signal data set according to the third data set and a first formula, wherein the first formula is:

[0093]

[0094] wherein, r i (k+1) is a residual signal data set obtained in the k+1th iteration, M i is a set average value calculation function, ε (k+1) is a coefficient of the k+1th iteration, IMF (k+1) (n i) is the i-th component value of the k+1-th iteration;

[0095] determining the modal component dataset according to the residual dataset and a second formula, wherein the second formula is:

[0096] IMF (k+1) = r (k) -r (k+1)

[0097] wherein IMF (k+1) is the modal component dataset.

[0098] In some embodiments, before the data decomposition and reconstruction of the plurality of channel state information datasets respectively, comprising:

[0099] normalizing the plurality of channel state information datasets by using tangent function;

[0100] clustering the normalized plurality of channel state information datasets to determine whether there is environmental interference;

[0101] if there is no environmental interference, using filtering algorithm to remove outliers deviating from the original signal trajectory from the plurality of channel state information datasets.

[0102] Exemplarily, for the captured CSI information, we need to do further processing to obtain more robust feature information.

[0103] A environment detection

[0104] determining whether there is a static target user.

[0105] (i) feature extraction

[0106] Ideally, the CSI received by the receiving end is:

[0107]

[0108] that is, the CSI is the sum of the static component Hs(f,t) and the dynamic component Hd(f,t), the static component is the reflection path of the static object, and the dynamic component is the path reflected from the human target, which is the combination of the complex amplitude A(f,t) after fading and the phase shift caused by path propagation , wherein d(t) is the length of the propagation path, f represents the propagation frequency, and λ is the wavelength. When the human body moves or moves slightly, because the signal amplitude is determined by the path length, A(f,t) in the dynamic component can be regarded as a constant.

[0109] Therefore we will dynamic path changes as part of the feature data, to measure the degree of interference or the respiratory signal environment. In addition, because the respiratory signal is periodic changes, using dynamic path changes and signal periodicity of the two features to extract feature vectors.

[0110] (ii) normalization process

[0111] Using normalization method for different feature data normalization, so that the feature data between the comparability speed up the analysis.

[0112] Nonlinear normalization method is suitable for data difference is larger, because the respiratory signal is weak, belongs to small scale changes, and other environmental noise and so have obvious fluctuations, so the application of nonlinear normalization method in the respiratory detection scene is more suitable. In this paper, through the tangent function, the original signal is mapped:

[0113]

[0114] x is the original signal, y is the signal obtained after mapping.

[0115] (iii) clustering

[0116] Using clustering algorithm to draw feature sample scatter plot, and then determine whether there is a larger environmental disturbance, to exclude the human body movement of weak respiratory motion disturbance, to ensure the correctness of the respiratory rate estimation.

[0117] B outlier removal

[0118] Using Hampel filter algorithm to remove outliers deviating from the original signal trajectory.

[0119] For the median of the original measurement data x is m, then the median absolute deviation is:

[0120] s=m(|x-m|)

[0121] Any point falling outside the closed interval [m-rs, m+rs] is declared as an outlier, where r is the threshold value and s is the median absolute deviation.

[0122] In signal processing technology, EMD (Empirical Mode Decomposition) is a means of signal decomposition based on time domain, compared with the traditional wavelet decomposition, without selecting the base function only rely on the original signal itself, but there is a problem of modal aliasing. EEMD algorithm is based on the optimization of EMD algorithm, using noise auxiliary analysis, but there is residual auxiliary noise, but the above algorithm has its shortcomings.

[0123] Therefore, the embodiment of the present application proposes to solve the problems of excessive residual noise and false modal based on an improved algorithm, and further better reconstruct the respiratory signal and filter high-frequency noise. Unlike the direct addition of Gaussian white noise in the decomposition process before, the innovation of the embodiment of the present application is:

[0124] (1) Selecting I groups of Gaussian white noise with a mean value of 0, each group is a pair of opposite Gaussian white noise sequences, and each group of noise is obtained by adding the average IMF component after EMD decomposition to the original signal for overall EMD decomposition, thereby improving the decomposition efficiency and solving the problem that the number of noises is difficult to align in set average.

[0125] (2) The residual signal obtained in the previous step is further subjected to EMD decomposition, and the Gaussian white noise added in each group is I groups of Gaussian white noise obtained by second EMD decomposition and set average of itself.

[0126] Therefore, the EMD decomposition process is first introduced:

[0127] (1) Calculate the mean value of the upper and lower envelope lines of the signal.

[0128] (2) Subtract the mean envelope line from the original signal.

[0129] (3) Repeat the iteration until the signal satisfies.

[0130] After the signal is decomposed by EMD, each layer signal component is called intrinsic mode component IMF (intrinsic mode component, Intrinsic Mode Function). Each IMF k contains only one oscillation mode, which represents the kth IMF component of the signal after EMD decomposition.

[0131] The improved EEMD algorithm process is as follows:

[0132] (1) There are I groups of Gaussian white noise n1, n2...n I Each group is composed of a pair of positive and negative Gaussian white noise which are opposite to each other, and there are 2I groups. The 2I groups of Gaussian white noise are subjected to EMD decomposition to obtain IMF components. Then, the two IMF components in each group are subjected to set average:

[0133]

[0134] i=0, 1...I, indicating I groups of Gaussian white noise, (1) indicating the first iteration.

[0135] (2) Let the signal obtained by Hampel filtering of the original signal x be z. Add the IMF (1) (n1), IMF (1) (n2),...IMF(1) (n I ), respectively, into the signal z, we have:

[0136] z i = z + ε (1) IMF (1) (n i ) (3)

[0137] ε (1) represents the weight coefficient of the Gaussian white noise in the first iteration.

[0138] (3) The mean value of z i is calculated to obtain M(z i ), and the residual signal is obtained by ensemble averaging:

[0139]

[0140] The first modal component is:

[0141] IMF (1) = x - r (1) (5)

[0142] (4) After the first iteration, the Gaussian white noise component obtained in this iteration is decomposed by EMD to obtain IMF (2) (n1), IMF (2) (n2),..., IMF (2) (n I ), which are added to the residual signal r (1) , respectively, to obtain:

[0143] r i = r (1) + ε (2) IMF (2) (n i ) (6)

[0144] i = 0, 1,..., I, indicating I sets of

[0145] (5) Then, EMD decomposition is performed, and the second iteration is performed, and finally the residual signal is obtained:

[0146]

[0147] The second modal component is:

[0148] IMF (2) = r (1) - r (2) (8)

[0149] Then, iteration is performed in turn until the residual signal is a monotonic signal, i.e., it cannot be decomposed by EMD, and then the iteration is stopped.

[0150] Figure 3 Decomposition figure obtained based on the algorithm provided in the embodiment of the present application, wherein the first one is the original signal, the following five components obtained after EMD decomposition are called IMF1-IMF5 in turn, and the last one is the residual signal.

[0151] These IMFs represent the frequency components in the original signal respectively, and are arranged in order from high frequency to low frequency, and it can be seen that the algorithm provided in the present application has better modal decomposition results and reduces a large number of low-frequency IMF components which are not significant for signal analysis.

[0152] Finally, the IMF components in the human respiratory frequency range of 0.167 Hz-0.617 Hz are selected, and they are added to reconstruct the respiratory signal, as shown in Figure 4 The reconstructed respiratory signal is more periodic, the signal waveform is smoother and more regular, and the influence of high-frequency noise is reduced to a certain extent.

[0153] The residual noise of the algorithm provided in the embodiment of the present application is compared with that of the EEMD algorithm, and it can be seen from Figure 5 that the residual noise of the improved EEMD algorithm is much smaller than that of the EEMD algorithm by many orders of magnitude, which provides accurate reconstruction of the original signal and better spectral separation, and good completeness can be achieved at a smaller average number.

[0154] In step 103, a plurality of high-sensitivity reconstructed data sets are selected from the plurality of reconstructed data sets, and the plurality of high-sensitivity reconstructed data sets are fused to obtain a fusion data set, wherein the periodicity of the high-sensitivity reconstructed data set corresponds to the human respiratory frequency range.

[0155] In some embodiments, the step 103 comprises:

[0156] The plurality of reconstructed data sets are respectively subjected to Fourier transform to obtain a plurality of frequency domain data sets, wherein the plurality of frequency domain data sets correspond to the plurality of reconstructed data sets;

[0157] According to the human respiratory frequency range and the plurality of frequency domain data sets, the plurality of high-sensitivity reconstructed data sets are determined from the plurality of reconstructed data sets;

[0158] The direct current components in the plurality of high-sensitivity reconstructed data sets are removed;

[0159] The plurality of high-sensitivity reconstructed data sets with the direct current components removed are combined to obtain the fusion data set.

[0160] Exemplarily,

[0161] The CSI change caused by breathing is a periodic change, and the human breathing frequency does not change much in a short time. The higher the periodicity of the subcarrier, the more sensitive the subcarrier is to breathing, and therefore a subcarrier with a higher periodicity needs to be selected as the subcarrier most sensitive to the breathing state. The periodicity can be measured by the breathing signal-to-noise ratio BNR.

[0162] The method for judging the periodicity of the CSI breathing signal is to calculate the breathing signal-to-noise ratio BNR of each subcarrier. First, the breathing data is subjected to FFT, and then the maximum energy in the human breathing frequency range of 0.167-0.617 Hz, i.e. the breathing energy, is found as the breathing energy u, and then the energy sum v is calculated, and then:

[0163]

[0164] We select the data set of the subcarrier signal with the largest BNR from the reconstructed data set of multiple subcarriers as the high-sensitivity reconstructed data set.

[0165] Finally, the energy sum of the high-sensitivity reconstructed data set of the selected subcarriers after removing the direct current component is calculated. In order to solve the problem of unstable measurement of a single subcarrier, the high-sensitivity reconstructed data sets of the subcarriers are combined by using the maximum ratio fusion strategy, so as to obtain a fusion data set.

[0166] In step 104, the breathing frequency is determined according to the fusion data set.

[0167] In some embodiments, the step 104 includes:

[0168] determining a plurality of peak values according to the fusion data set;

[0169] obtaining a sliding window;

[0170] taking a plurality of data from the fusion data set by using the sliding window;

[0171] judging whether the peak value in the taken plurality of data is a true peak value, wherein the true peak value is the maximum value in the taken plurality of data;

[0172] filtering out false peak values in the taken plurality of data;

[0173] if the sliding window does not reach the end of the fusion data set, the sliding window is shifted, and the step of taking a plurality of data from the fusion data set by using the sliding window is jumped to;

[0174] determining a plurality of frequency components by using the plurality of peak values;

[0175] calculating the average value of the plurality of frequency components as the breathing frequency.

[0176] Exemplarily, for the respiratory frequency determination aspect, the embodiment of the present application determines by a peak seeking algorithm and a frequency estimation.

[0177] Peak seeking algorithm:

[0178] False peaks are filtered by using a sliding window. Firstly, all peaks of the sub-carrier signal are calculated, and a sliding window about the peaks is defined. The window size is set to 60 sampling points, and the peaks are compared with the data in the window to determine whether the peak is greater than all the data in the window. If not, the peak is a false peak and needs to be filtered out, otherwise, it is retained.

[0179] Frequency estimation:

[0180] Suppose that M peaks are obtained by the peak seeking algorithm, and the M frequency components corresponding to the peaks are averaged to obtain the final respiratory frequency.

[0181] Optionally, after determining the respiratory frequency according to the fusion data set, the embodiment can determine the sleep of the user based on the respiratory frequency, and realize sleep respiration detection.

[0182] The embodiment of the static user respiration detection method of the present application first acquires a plurality of channel state information data sets; then performs data decomposition and reconstruction on the plurality of channel state information data sets respectively to obtain a plurality of reconstructed data sets, wherein the plurality of reconstructed data sets correspond to the plurality of channel state information data sets, and a reconstructed data set is obtained by reconstruction according to a first component data set of a human respiration frequency range, and the first component data set is obtained based on data decomposition of the channel state information data set; then a plurality of high-sensitivity reconstructed data sets are selected from the plurality of reconstructed data sets, and a fusion data set is obtained by fusing the plurality of high-sensitivity reconstructed data sets, wherein the periodicity of a high-sensitivity reconstructed data set corresponds to the human respiration frequency range; finally, the respiratory frequency is determined according to the fusion data set. The embodiment of the present application increases an environment detection part in the data preprocessing part to extract a feature vector from the CSI respiration signal to exclude environmental interference; in the high-frequency noise filtering part, the embodiment of the present application proposes an improved EEMD algorithm to solve the problems of excessive residual noise and the existence of pseudo modes, and thus better reconstructs the respiration signal and filters the high-frequency noise.

[0183] It should be understood that the size of the serial number of each step in the above embodiment does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0184] The following is the device embodiment of the present application, and for details not described in detail, reference can be made to the corresponding method embodiments described above.

[0185] Figure 6is a functional block diagram of a static user respiration detection device provided by an embodiment of the present application, referring to Figure 6 The static user respiration detection device comprises a data acquisition module 601, a data reconstruction module 602, a data fusion module 603, and a respiration frequency determination module 604, wherein:

[0186] The data acquisition module 601 is configured to acquire a plurality of channel state information data sets.

[0187] The data reconstruction module 602 is configured to perform data decomposition and reconstruction on the plurality of channel state information data sets respectively, to obtain a plurality of reconstructed data sets, wherein the plurality of reconstructed data sets correspond to the plurality of channel state information data sets, and a reconstructed data set is obtained by reconstruction according to a first component data set of a human respiration frequency range, and the first component data set is obtained based on data decomposition of the channel state information data set.

[0188] The data fusion module 603 is configured to select a plurality of high-sensitivity reconstructed data sets from the plurality of reconstructed data sets, and fuse the plurality of high-sensitivity reconstructed data sets to obtain a fused data set, wherein the periodicity of a high-sensitivity reconstructed data set corresponds to the human respiration frequency range.

[0189] The respiration frequency determination module 604 is configured to determine a respiration frequency according to the fused data set.

[0190] Figure 7 is a functional block diagram of an electronic device provided by an embodiment of the present application. As shown in Figure 7 The electronic device 7 of the embodiment comprises a processor 700 and a memory 701, and the memory 701 stores a computer program 702 which can run on the processor 700. The processor 700 implements the steps in the above-mentioned various static user respiration detection methods and embodiments when executing the computer program 702, such as the steps 101 to 104 shown in Figure 1

[0191] For example, the computer program 702 can be divided into one or more modules / units, which are stored in the memory 701 and executed by the processor 700 to complete the present application.

[0192] The electronic device 7 can be a desktop computer, a notebook computer, a palm computer, a cloud server, and other computing devices. The electronic device 7 can include, but is not limited to, a processor 700 and a memory 701. Those skilled in the art can understand that Figure 7 ​The electronic device 7 is merely an example and does not constitute a limitation on the electronic device 7, and can include more or fewer components than illustrated, or combine certain components, or different components, for example, the electronic device 7 can also include an input / output device, a network access device, a bus, etc.

[0193] The processor 700 can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0194] The memory 701 can be an internal storage unit of the electronic device 7, for example, a hard disk or a memory of the electronic device 7. The memory 701 can also be an external storage device of the electronic device 7, for example, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Further, the memory 701 can include both the internal storage unit and the external storage device of the electronic device 7. The memory 701 is used to store the computer program 702 and other programs and data required by the electronic device 7. The memory 701 can also be used to temporarily store data that has been output or will be output.

[0195] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be realized in the form of hardware or software function unit. In addition, the specific name of each functional unit and module is only for the convenience of mutual distinction, and does not limit the protection scope of the present application. The specific working process of the unit and module in the above system can refer to the corresponding process in the foregoing method embodiment, which will not be repeated here.

[0196] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can be referred to the related description of other embodiments.

[0197] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. A person skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0198] In the embodiments provided by the present application, it should be understood that the disclosed apparatus / equipment and method can be implemented in other ways. For example, the above-described apparatus / equipment embodiments are merely schematic, for example, the division of the modules or units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed each other can be indirect coupling or communication connection through some interface, apparatus or unit, which can be electrical, mechanical or other forms.

[0199] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the present embodiment scheme.

[0200] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0201] The integrated module / unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, all or part of the processes in the above-mentioned embodiment methods can also be completed by a computer program instructing related hardware, and the computer program can be stored in a computer readable storage medium. When the computer program is executed by a processor, the steps of each method and device embodiment described above can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable files or some intermediate forms. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0202] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for part of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A static user breathing detection method, characterized in that: include: Acquiring multiple channel state information data sets; Decomposing and reconstructing the multiple channel state information data sets to obtain multiple reconstructed data sets includes: The multiple reconstructed data sets correspond to the multiple channel state information data sets, the reconstructed data set is reconstructed based on a first component data set within a human respiratory frequency range, the first component data set is obtained based on data decomposition of the channel state information data set, and for each of the multiple channel state information data sets, the following steps are performed: Using a filtering algorithm to remove outliers that deviate from the original signal trajectory, a first data set is obtained; Decomposing the first data set by multiple time domain signal decompositions to obtain multiple first component data sets corresponding to multiple frequencies; selecting, according to a human respiratory frequency range and a frequency corresponding to the first component data set, a plurality of data sets from the plurality of first component data sets as a plurality of second component data sets; Reconstructing according to the plurality of second component data sets to obtain a reconstructed data set; selecting a plurality of high-sensitivity reconstructed data sets from the plurality of reconstructed data sets, and fusing the plurality of high-sensitivity reconstructed data sets to obtain a fused data set, wherein the periodicity of the high-sensitivity reconstructed data sets corresponds to a range of human respiratory frequency; determining a respiratory rate based on the fused dataset; The step of decomposing the first data set by multiple time domain signal decompositions to obtain multiple first component data sets corresponding to multiple frequencies includes: Acquire a plurality of first Gaussian white noise arrays, wherein the first Gaussian white noise arrays include Gaussian white noise data that are opposite numbers of each other; Performing empirical mode decomposition on the plurality of first white Gaussian noise arrays to obtain a plurality of component values, wherein the component value is an average value of two intrinsic modal component data, and the two intrinsic modal component data are obtained based on the empirical mode decomposition of the first white Gaussian noise array; adding the plurality of component values ​​to the first data set respectively to obtain a third data set; Calculating a collective mean value for the third data set to obtain a residual signal data set and a modal component data set, and using the modal component data set as the first component data set; If the residual signal data set is not monotonic, the residual signal data set is used as the first data set, empirical mode decomposition is performed on the multiple component values, multiple results obtained by the empirical mode decomposition are used as multiple component values, and the process jumps to the step of adding the multiple component values ​​to the first data set respectively to obtain a third data set.

2. The static user breathing detection method according to claim 1, characterized in that: Calculating the ensemble mean value of the third data set to obtain the residual signal data set and the modal component data set includes: The residual signal data set is determined according to the third data set and the first formula, wherein the first formula is: Where, For the The residual signal data set obtained by the iteration is is the ensemble mean calculation function, For the The coefficient of the iteration, For the The first iteration component values; The modal component dataset is determined according to the residual signal dataset and a second formula, wherein the second formula is: Where, is the modal component dataset.

3. The static user breathing detection method according to claim 1, characterized in that: The selecting a plurality of high-sensitivity reconstructed data sets from the plurality of reconstructed data sets, and fusing the plurality of high-sensitivity reconstructed data sets to obtain a fused data set, includes: Performing Fourier transform on the multiple reconstructed data sets respectively to obtain multiple frequency domain data sets, wherein the multiple frequency domain data sets correspond to the multiple reconstructed data sets; Determining the plurality of high-sensitivity reconstructed data sets from the plurality of reconstructed data sets according to a human respiratory frequency range and the plurality of frequency domain data sets; removing DC components from the plurality of high-sensitivity reconstructed data sets; The multiple high-sensitivity reconstructed data sets with DC components removed are combined to obtain the fused data set.

4. The static user breathing detection method according to claim 1, characterized in that: Determining the respiratory rate according to the fused data set includes: a plurality of peaks determined from the fused data set; Get the sliding window; Retrieving a plurality of data from the fused data set using the sliding window; Determine whether a peak value among the multiple data points retrieved is a true peak value, wherein the true peak value has the largest value among the multiple data points retrieved; Filter out false peaks in the extracted multiple data; If the sliding window has not reached the end of the fused data set, sliding the sliding window and jumping to the step of extracting a plurality of data from the fused data set using the sliding window; Determining multiple frequency components using multiple peaks; An average value of the plurality of frequency components is calculated as the respiratory frequency.

5. The static user breathing detection method according to any one of claims 1 to 4, characterized in that: Before respectively decomposing and reconstructing the plurality of channel state information data sets, the method further comprises: Performing normalization processing on the multiple channel state information data sets using a tangent function; Clustering the multiple channel state information data sets after normalization to determine whether there is environmental interference; If there is no environmental interference, a filtering algorithm is used to remove abnormal values ​​that deviate from the original signal trajectory from the multiple channel state information data sets.

6. A static user breathing detection device, characterized in that: For implementing the static user breathing detection method according to any one of claims 1 to 5, the static user breathing detection device comprises: A data acquisition module, configured to acquire multiple channel state information data sets; a data reconstruction module, configured to perform data decomposition and reconstruction on the multiple channel state information data sets to obtain multiple reconstructed data sets, wherein the multiple reconstructed data sets correspond to the multiple channel state information data sets, and the reconstructed data sets are reconstructed based on a first component data set within a human respiratory frequency range, and the first component data set is obtained based on data decomposition of the channel state information data set; a data fusion module, configured to select a plurality of high-sensitivity reconstructed data sets from the plurality of reconstructed data sets, and fuse the plurality of high-sensitivity reconstructed data sets to obtain a fused data set, wherein the periodicity of the high-sensitivity reconstructed data sets corresponds to a range of human respiratory frequency; as well as, A respiratory rate determination module is configured to determine the respiratory rate based on the fused data set.

7. An electronic device comprising a memory and a processor, wherein the memory stores a computer program that can be run on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 5 are implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images