Accelerometer-based respiration rate detection method, device and wearable device

Abstract

The application discloses a respiration rate detection method and device based on an accelerometer and a wearable device. The respiration rate detection device comprises an accelerometer, a forward acceleration acquisition module, a respiration wave reconstruction module, a respiration rate calculation and output module, and realizes the respiration rate detection method. The accelerometer is used for measuring the acceleration of a user in each axis direction. The forward acceleration acquisition module is connected with the accelerometer, acquires the acceleration in each axis direction measured by the accelerometer, and superimposes the components of each acceleration in the forward direction to obtain a forward acceleration signal. The respiration wave reconstruction module is connected with the forward acceleration acquisition module, receives the forward acceleration signal, and reconstructs a respiration wave according to the forward acceleration signal. The respiration rate calculation and output module is connected with the respiration wave reconstruction module, receives the respiration wave, and outputs a respiration rate according to the respiration wave. The accuracy and precision of the respiration wave reconstructed based on the accelerometer are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of wearable devices, and specifically relates to a respiratory rate detection method and device based on an accelerometer, and a wearable device. Background Art

[0002] Respiratory monitoring is one of the most important means of assessing vital status, and its importance cannot be overstated. The occurrence of respiratory disorders can gradually or continuously manifest through respiratory parameters during daily life or exercise. Therefore, monitoring respiratory parameters during regular exercise can help detect respiratory disorders early, allowing for early detection and treatment, thus preventing severe consequences and even life-threatening consequences of severe respiratory disorders.

[0003] Chinese patent CN102753095A discloses a method and apparatus for determining respiratory signals, which involves performing frequency domain spectrum analysis on data from acceleration axes with large amplitude responses, identifying frequency bands with high acceleration power as non-respiratory response bands. Frequency spectrum analysis is also performed on data from acceleration axes with small amplitude responses, identifying frequency bands with high acceleration power as respiratory response bands. Adaptive filtering is then used to extract data from the respiratory frequency bands. When the respiratory frequency and motion frequency intersect, motion response errors can occur in the respiratory wave, reducing the accuracy of respiratory rate calculation. Summary of the Invention

[0004] The present invention provides a respiratory rate detection method, device and wearable device based on an accelerometer. According to the different characteristics of the user's respiratory state in different states and the different acceleration responses, different signal reconstruction methods and systems are designed to extract the respiratory wave from the acceleration, thereby realizing the calculation of the respiratory rate and improving the calculation accuracy.

[0005] In order to solve the above technical problems, the present invention adopts the following technical solutions:

[0006] A respiratory rate detection method based on an accelerometer, comprising:

[0007] Get forward acceleration signal;

[0008] Performing spectrum analysis on the forward acceleration signal to generate a forward acceleration spectrum;

[0009] The user state is divided into multiple motion levels according to the maximum response frequency and acceleration amplitude in the forward acceleration spectrum; the maximum response frequency is the frequency range of the maximum acceleration amplitude distribution;

[0010] Generate filtered acceleration signals by filtering the forward acceleration signals at different motion levels with corresponding frequencies;

[0011] The respiratory wave is reconstructed according to the periodic characteristics of the filtered acceleration signal.

[0012] In one embodiment, the exercise levels include three, namely, non-exercise state, normal exercise state, and high-intensity exercise state; a preset first frequency, a second frequency greater than the first frequency, a first acceleration, and a second acceleration greater than the first acceleration;

[0013] The maximum response frequency corresponding to the non-motion state is smaller than the first frequency, and the acceleration amplitude is smaller than the first acceleration;

[0014] The maximum response frequency corresponding to the normal motion state is between the first frequency and the second frequency, and the acceleration amplitude is smaller than the second acceleration;

[0015] The maximum response frequency corresponding to the high-intensity exercise state is greater than the second frequency, and the acceleration amplitude is greater than the second acceleration.

[0016] In one embodiment, a third frequency is preset, which is greater than the second frequency; the filtered acceleration signal includes a first filtered acceleration signal, a second filtered acceleration signal, and a third filtered acceleration signal;

[0017] The forward acceleration signal corresponding to the non-motion state is subjected to low-pass filtering at the first frequency to obtain the first filtered acceleration signal;

[0018] The forward acceleration signal corresponding to the normal motion state is subjected to low-pass filtering at the third frequency to obtain the second filtered acceleration signal;

[0019] The forward acceleration signal corresponding to the high-intensity motion state is subjected to low-pass filtering at the third frequency or band-pass filtering from the second frequency to the third frequency to obtain the third filtered acceleration signal.

[0020] In one embodiment, the first frequency is 0.5 Hz; the second frequency is 1 Hz; the third frequency is 2 Hz;

[0021] The first acceleration is 0.01g; the second acceleration is 0.1g.

[0022] In some embodiments, the forward acceleration signal is subjected to spectrum analysis by Fourier transform.

[0023] In some embodiments, the respiratory wave is reconstructed by detecting peak points of the first filtered acceleration, the second filtered acceleration, or the third filtered acceleration, and marking a timestamp of each peak point;

[0024] The respiratory wave is reconstructed according to each peak point with the time stamp.

[0025] In some embodiments, each of the discrete peak points with the timestamp is used to generate an envelope using a cubic spline function, and the envelope is made consistent with the sampling frequency of the forward acceleration signal to reconstruct a response curve;

[0026] The response curve in the non-exercise state is low-pass filtered at the first frequency to obtain the respiratory wave; the continuous curve in the normal exercise state is low-pass filtered at the second frequency to obtain the respiratory wave; and the continuous curve in the high-intensity exercise state is low-pass filtered at the third frequency to obtain the respiratory wave.

[0027] In some embodiments, the formula

[0028] b_r = Ts*60 / △t (1)

[0029] Solve for breathing rate;

[0030] Wherein, b_r is the respiratory rate, the unit is bpm (times / minute);

[0031] Ts is the sampling frequency of the forward acceleration signal;

[0032] △t is the epoch interval of each respiratory cycle.

[0033] A respiratory rate detection device, comprising an accelerometer, a forward acceleration acquisition module, a respiratory wave reconstruction module, and a respiratory rate calculation and output module, to implement the above-mentioned accelerometer-based respiratory rate detection method;

[0034] The accelerometer is a single-axis or multi-axis accelerometer, used to measure the acceleration of the user in a single-axis or multi-axis direction;

[0035] The forward acceleration acquisition module is connected to the accelerometer, acquires the acceleration measured by the accelerometer, and superimposes the components of the acceleration in the forward direction to obtain a forward acceleration signal;

[0036] The respiratory wave reconstruction module is connected to the forward acceleration acquisition module, receives the forward acceleration signal, and reconstructs the respiratory wave according to the forward acceleration signal;

[0037] The respiratory rate calculation and output module is connected to the respiratory wave reconstruction module, receives the respiratory wave, and outputs the respiratory rate according to the respiratory wave.

[0038] A wearable device comprises the above-mentioned respiratory rate detection device.

[0039] Compared with the prior art, the advantages and positive effects of the present invention are: the accelerometer-based respiratory rate detection method, device and wearable device of the present invention obtain the acceleration signal in the breathing direction, that is, the forward acceleration signal, excludes the acceleration signal in the direction other than the chest rise and fall, reduces the number of interference signals, and improves the detection accuracy.

[0040] A forward acceleration signal is subjected to spectral analysis to obtain a spectrogram, and the forward acceleration signal is determined to be from different user states based on the respiratory frequency range and acceleration amplitude range in different user states; the forward acceleration signal is filtered at the frequency corresponding to the user state, and noise above the respiratory frequency is filtered out to obtain a filtered acceleration signal of the respiratory response; and a respiratory wave is reconstructed based on the periodic characteristics of the filtered acceleration signal. The present invention classifies the user state into different motion levels based on the maximum response frequency and acceleration amplitude, and performs filtering at different frequencies at different motion levels, thereby isolating the respiratory frequency from the motion frequency to the greatest extent possible, reducing the overlapping interference of the motion acceleration signal on the respiratory acceleration, and improving the accuracy and precision of the reconstructed respiratory wave signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0042] Figure 1 1 is a flow chart of a respiratory rate detection method based on an accelerometer proposed in the present invention;

[0043] Figure 2 1 is a schematic diagram of the composition of a respiratory rate detection device proposed by the present invention;

[0044] Figure 3 is a schematic diagram of an example of a forward acceleration signal;

[0045] Figure 4 This is a schematic diagram of the acceleration amplitude spectrum and amplitude sequence distribution obtained by Fourier transforming the forward acceleration;

[0046] Figure 5 is a schematic diagram comparing a forward acceleration signal and a filtered acceleration signal after the forward acceleration signal is filtered according to the corresponding motion level;

[0047] Figure 6 is a schematic diagram of peak sampling of filtered acceleration signal;

[0048] Figure 7 It is a schematic diagram of the respiratory wave reconstructed based on the periodic characteristics;

[0049] Figure 8 It is a schematic diagram comparing the reconstructed respiratory wave and the measured respiratory wave;

[0050] Figure 9 It is a schematic diagram comparing the detected respiratory rate and the measured respiratory rate;

[0051] Figure 10 It is a schematic diagram of the filter function model structure.

[0052] In the figure,

[0053] 1. Accelerometer; 2. Forward acceleration acquisition module; 3. Respiratory wave reconstruction module; 4. Respiratory rate calculation and output module. DETAILED DESCRIPTION

[0054] The following describes embodiments of the present invention in detail, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and are not to be construed as limiting the present invention.

[0055] The present invention discloses a respiratory rate detection method, device and wearable device based on an accelerometer. The device and the wearable device are equipped with an accelerometer 1 and a processing module and are worn on the chest of a user to realize the calculation of the user's respiratory wave and respiratory rate.

[0056] The key to calculating the respiratory rate through accelerometer 1 is to detect the chest fluctuation caused by breathing; however, the chest fluctuation responds very little to the vertical changes of accelerometer 1 and is easily affected by noise, so it is difficult to detect the chest fluctuation signal through accelerometer 1 and directly process it to obtain an accurate respiratory wave.

[0057] Reference Figure 1 The respiratory rate detection method based on accelerometer of the present invention includes the following steps.

[0058] S1: Detect the user's acceleration in a single axis or multiple axes through the accelerometer 1, and separate and superimpose the acceleration in the single axis or multiple axes in the chest and back directions through the forward acceleration acquisition module 2 to obtain a forward acceleration signal.

[0059] S2: Performing spectrum analysis on the forward acceleration signal and classifying the user's state into three levels of motion: non-motion state, normal motion state, and high-intensity motion state, based on the user's different responses to acceleration in different states. The forward acceleration signals in different states are filtered at corresponding frequencies to obtain filtered acceleration signals.

[0060] In the forward acceleration signal spectrum, the response of the user state to acceleration is reflected in the maximum response frequency and acceleration amplitude; the maximum response frequency is the frequency range of the maximum acceleration amplitude distribution.

[0061] S3: Reconstructing the respiratory wave according to the periodic characteristics of the filtered acceleration signal, and then obtaining the respiratory rate according to the reconstructed respiratory wave.

[0062] Reference Figure 2 The respiratory rate detection device of the present invention includes an accelerometer 1, a forward acceleration acquisition module 2, a respiratory wave reconstruction module 3, and a respiratory rate calculation and output module 4, which are used to implement the above-mentioned respiratory rate detection method.

[0063] The accelerometer 1 is a single-axis accelerometer or a multi-axis accelerometer, and is used to measure the acceleration of the user in a single-axis or multi-axis direction.

[0064] The forward acceleration acquisition module 2 is connected to the accelerometer 1 to acquire the uniaxial acceleration or the acceleration in multiple axes measured by the accelerometer 1 and to superimpose the components of each acceleration in the forward direction to obtain a forward acceleration signal.

[0065] The respiratory wave reconstruction module 3 is connected to the forward acceleration acquisition module 2, receives the forward acceleration signal, and reconstructs the respiratory wave according to the periodic characteristics of the forward acceleration.

[0066] Specifically, the forward acceleration signal is spectrally analyzed. Based on the user's varying responses to acceleration, the user's states are categorized as non-exercise, normal exercise, and high-intensity exercise. This is reflected in the forward acceleration spectrum as the difference in maximum response frequency and acceleration amplitude. The forward acceleration signal for each different exercise state is filtered at the corresponding frequency to generate a filtered acceleration signal. Respiratory waves are then reconstructed based on the periodic characteristics of the filtered acceleration. The maximum response frequency of the spectrum is the frequency range within which the maximum acceleration amplitude is distributed.

[0067] The wearable device includes the above-mentioned respiratory rate detection device.

[0068] In the prior art, calculating the respiratory rate generally requires wearing a respiratory sensor device worn on the nose / mouth or wearing a blood oxygen detection device; however, wearing the above-mentioned sensor devices will affect the exercise experience.

[0069] The accelerometer-based respiratory rate detection method, device, and wearable device of the present invention obtain acceleration signals in the respiratory direction, i.e., forward acceleration signals, exclude acceleration signals in directions other than chest undulation, reduce the number of interference signals, and improve detection accuracy.

[0070] A forward acceleration signal is subjected to spectral analysis to obtain a spectrogram, and the forward acceleration signal is determined to be from different user states based on the respiratory frequency range and acceleration amplitude range in different user states; the forward acceleration signal is filtered at the frequency corresponding to the user state, and noise above the respiratory frequency is filtered out to obtain a filtered acceleration signal of the respiratory response; and a respiratory wave is reconstructed based on the periodic characteristics of the filtered acceleration signal. The present invention classifies the user state into different motion levels based on the maximum response frequency and acceleration amplitude, and performs filtering at different frequencies at different motion levels, thereby isolating the respiratory frequency from the motion frequency to the greatest extent possible, reducing the overlapping interference of the motion acceleration signal on the respiratory acceleration, and improving the accuracy and precision of the reconstructed respiratory wave signal.

[0071] The wearable device of the present invention can be worn on the chest via a bandage, which can enhance the exercise experience compared to a respiratory sensor device worn on the nose / mouth or a blood oxygen detection device.

[0072] The following describes in detail the specific process, structure and principle of the respiratory rate detection method, device and wearable device of the present invention through specific embodiments.

[0073] In one embodiment, referring to Figure 1 、 Figure 2 In the respiratory rate detection method, the user states include three types, namely, non-exercise state, normal exercise state, and high-intensity exercise state; and the first frequency, second frequency, first acceleration, and second acceleration are preset according to the corresponding user's respiratory frequency and acceleration amplitude in the non-exercise state, normal exercise state, and high-intensity exercise state; and the first frequency is less than the second frequency; and the first acceleration is less than the second acceleration.

[0074] When the maximum response frequency of the spectrum of the forward acceleration is less than the first frequency and the acceleration amplitude is less than the first acceleration, it is determined that the user is in a non-exercise state.

[0075] When the maximum response frequency of the spectrum of the forward acceleration is between the first frequency and the second frequency and the acceleration amplitude is less than the second acceleration, it is determined that the user is in a normal movement state.

[0076] When the maximum response frequency of the spectrum of the forward acceleration is greater than the second frequency and the acceleration amplitude is greater than the second acceleration, it is determined that the user is in a high-intensity exercise state.

[0077] A third frequency is preset, which is greater than the second frequency. The filtered acceleration signal includes a first filtered acceleration signal, a second filtered acceleration signal, and a third filtered acceleration signal. The forward acceleration signal corresponding to the non-motion state is low-pass filtered at the first frequency to obtain the first filtered acceleration signal; the forward acceleration signal corresponding to the normal motion state is low-pass filtered at the third frequency or band-pass filtered from the second frequency to the third frequency to obtain the second filtered acceleration signal; the forward acceleration signal corresponding to the high-intensity motion state is low-pass filtered at the third frequency to obtain the third filtered acceleration signal. The filtering makes the periodic characteristics of the first, second, and third filtered acceleration signals more pronounced.

[0078] The peak values ​​of the filtered acceleration signal and the timestamps corresponding to the peak values ​​are obtained; and the respiratory wave is reconstructed based on the discrete peak points with the timestamps.

[0079] The respiratory wave reconstruction module 3 of the respiratory rate detection device includes a processor module, a first-frequency low-pass filter module, a second-to-third-frequency band-pass filter module, and a third-frequency low-pass filter module. The processor module is configured with the first frequency, the second frequency, the third frequency, the first acceleration, and the second acceleration. The forward acceleration signal acquired by the forward acceleration acquisition module 2 is subjected to spectral analysis, and the user's status is determined based on the maximum response frequency and acceleration amplitude of the spectrum. Based on the determination result, the first-frequency low-pass filter module, the second-to-third-frequency band-pass filter module, or the third-frequency low-pass filter module is selected to filter the forward acceleration signal to generate a filtered acceleration signal. The respiratory wave reconstruction module 3 obtains the peak values ​​of the filtered acceleration signal and the timestamps of each peak value, and reconstructs the respiratory wave based on each timestamp-containing peak value.

[0080] The respiratory rate output module is connected to the respiratory wave reconstruction module 3, receives the generated respiratory wave, and calculates the respiratory rate output according to the respiratory wave.

[0081] The wearable device includes the above-mentioned respiratory rate detection device.

[0082] The respiratory rate detection method, apparatus, and wearable device of this embodiment classify the normal user state into a non-exercise state, a normal exercise state, and a high-intensity exercise state based on the response of the forward acceleration signal frequency. The measured forward acceleration signal is then determined based on its frequency spectrum to determine the exercise state in which it was measured and is filtered for the corresponding frequency. That is, the forward acceleration signal corresponding to each exercise state is filtered for the respiratory wave frequency in the corresponding state, thereby improving the targeted signal processing, reducing the number of interfering signals, and improving measurement and detection accuracy.

[0083] Reference Figure 3 、 Figure 4 、 Figure 5, which are the forward acceleration signal, spectrum analysis diagram and filtered acceleration signal under normal motion state.

[0084] In a specific embodiment, the filter module of the respiratory rate detection method of the present invention is constructed using a bilinear model, and its objective function is output = f(acc, b_n, a_n); where acc represents the forward original acceleration data, and b_n and a_n are the parameters of the filter respectively.

[0085] Based on the data extracted by filtering, the periodic characteristics are checked and the detected peak points are cached. The peak point detection and extraction results refer to Figure 6 .

[0086] In a specific embodiment, referring to Figure 4 , the forward acceleration signal is transformed into an amplitude spectrum through Fourier transform.

[0087] The filtered acceleration signal is reconstructed using a cubic spline function. Specifically, the envelope of each discrete peak point with a time stamp is generated using a cubic spline function to obtain a response curve. The response curve is filtered at the corresponding frequency to generate a respiratory wave. Figure 7 , which is the respiratory wave generated during normal exercise. Figure 8 、 Figure 9 , respectively, are schematic diagrams for comparing the respiratory wave calculated by the respiratory rate detection method of the present invention and the measured respiratory wave under normal exercise conditions, and schematic diagrams for comparing the calculated respiratory rate and the measured respiratory rate.

[0088] The timestamp is the time of the detected peak point. The frequency of the envelope is consistent with the sampling frequency of the forward acceleration.

[0089] Specifically, after obtaining the peak point sequence of the filtered acceleration signal, cubic spline processing is performed based on the acceleration amplitude and the epoch interval between two points to obtain an envelope consisting of the peak points. This envelope is aligned with the original sampling frequency of the forward acceleration signal. The objective function of the cubic spline is output = f(peak_indexs, peak_points, Ts), where output is the output discrete envelope sequence; peak_indexs is the index of the peak point, that is, the time stamp sequence of the peak point, which is used to calculate the epoch interval of each cycle of the response curve; peak_points represents the acceleration amplitude of the peak point, and Ts is the original acceleration sampling frequency.

[0090] The response curve in the non-exercise state is low-pass filtered at a first frequency to obtain a respiratory wave; the continuous curve in the normal exercise state is low-pass filtered at a second frequency to obtain a respiratory wave; the continuous curve in the high-intensity exercise state is low-pass filtered at a third frequency to obtain the respiratory wave.

[0091] The respiratory wave reconstruction module 3 of the respiratory rate detection device implements the reconstruction of the response curve and filters the response curve through the first frequency low-pass filtering module, the second frequency low-pass filtering module, or the third frequency low-pass filtering module to obtain the respiratory wave; the wearable device includes the above-mentioned respiratory rate detection device.

[0092] The respiratory rate detection method, device and wearable device of this embodiment have periodicity according to the respiratory action, and the response of the respiratory action to acceleration is also periodic, and the waveform presented by the acceleration is an approximate trigonometric function waveform. By detecting the peak point of the filtered acceleration signal, the start and end time of the waveform can be obtained, that is, the start and end time of each respiratory cycle. The peak point of the discrete point data format is obtained as [peak point timestamp, peak point amplitude]; the discrete peak points are interpolated using a cubic spline function to reconstruct the response curve, and then the response curve is low-pass filtered to obtain a reconstructed respiratory wave curve. While eliminating the high-frequency noise caused by the peak error, the data of the respiratory frequency band is retained, thereby improving the precision and accuracy of the respiratory wave and respiratory rate.

[0093] In a specific embodiment, the first frequency is 0.5 Hz; the second frequency is 1 Hz; the third frequency is 2 Hz; the first acceleration is 0.01 g; and the second acceleration is 0.1 g.

[0094] In a specific embodiment, the first frequency low-pass filtering module, the second frequency low-pass filtering module, the third frequency low-pass filtering module, and the second-to-third frequency band-pass filtering module use the following filter functions:

[0095]

[0096] Where input represents the input acceleration sequence, output represents the output filtered value, a and b represent the denominator and numerator vectors of the corresponding filter. The order of the bandpass filter is M, and n represents the length of the sequence. This is a bilinear model with nested input and output. For details on the model structure, refer to Figure 10 .

[0097] The cubic spline function processes the peak point as follows.

[0098] The node that the cubic spline function needs to interpolate is x i ,y i (i=0,1,……,n) and the set of horizontal coordinates xx that need to be interpolated; calculate h for i=0,1,……,n-1 i =x i+1 –x i .

[0099] Calculate the intermediate value α i and β i(i=0,1,……,n):

[0100] The first type of boundary conditions:

[0101] α0=0,β0=2m0

[0102] α n =1,β n =2m n

[0103] The second type of boundary conditions:

[0104]

[0105]

[0106] Middle part:

[0107]

[0108] Next, calculate a, b, and m:

[0109] When i=0:

[0110]

[0111] When i=1,2,...,n:

[0112]

[0113] Calculate m:

[0114] m n =b n (i=n);

[0115] m i =a i m i+1 +b i (i=n-1,...,1,0);

[0116] Finally, the interval [x i ,x i+1 ](i=0,1,…,n-1), then find the value of s(x) and output:

[0117]

[0118] After obtaining the timestamp of the peak point and the acceleration amplitude sequence of the peak point, the timestamp of the peak point is xi, and the acceleration amplitude is yi. Substitute them into the cubic spline function, and the final output value yy is the envelope sequence, and xx is the timestamp interpolated according to the sampling frequency.

[0119] In a specific embodiment, by formula

[0120] b_r = Ts*60 / △t (1)

[0121] Solve for breathing rate;

[0122] Wherein, b_r is the respiratory rate;

[0123] Ts is the sampling frequency of the forward acceleration signal;

[0124] △t is the epoch interval of each respiratory cycle.

[0125] △t is the time epoch interval of each respiratory cycle, △t / Ts is the time interval in seconds, the reciprocal of △t / Ts Ts / △t is the number of breaths in one second, and multiplying it by 60 is the number of breaths in one minute, which is the standard unit of respiratory rate bpm (times / minute).

[0126] In the description of the present invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of such features. In the description of the present invention, "plurality" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0127] In the present invention, unless otherwise specified or limited, the terms "installed," "connected," "connect," "fixed," etc. should be understood in a broad sense. For example, they can refer to fixed connection, detachable connection, or integration; mechanical connection, electrical connection; direct connection, or indirect connection through an intermediate medium; internal communication between two components, or interaction between two components, unless otherwise specified. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0128] In the description of this specification, the reference terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" mean that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0129] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A respiratory rate detection method based on an accelerometer, characterized in that: include: Get forward acceleration signal; Performing spectrum analysis on the forward acceleration signal to generate a forward acceleration spectrum; The user state is divided into multiple motion levels according to the maximum response frequency and acceleration amplitude in the forward acceleration spectrum; the maximum response frequency is the frequency range of the maximum acceleration amplitude distribution; Generate filtered acceleration signals by filtering the forward acceleration signals at different motion levels with corresponding frequencies; The respiratory wave is reconstructed according to the periodic characteristics of the filtered acceleration signal.

2. The respiratory rate detection method according to claim 1, wherein The exercise levels include three: non-exercise state, normal exercise state, and high-intensity exercise state; a preset first frequency, a second frequency greater than the first frequency, a first acceleration, and a second acceleration greater than the first acceleration; The maximum response frequency corresponding to the non-motion state is smaller than the first frequency, and the acceleration amplitude is smaller than the first acceleration; The maximum response frequency corresponding to the normal motion state is between the first frequency and the second frequency, and the acceleration amplitude is smaller than the second acceleration; The maximum response frequency corresponding to the high-intensity exercise state is greater than the second frequency, and the acceleration amplitude is greater than the second acceleration.

3. The respiratory rate detection method according to claim 2, wherein: A third frequency is preset, which is greater than the second frequency; the filtered acceleration signal includes a first filtered acceleration signal, a second filtered acceleration signal, and a third filtered acceleration signal; The forward acceleration signal corresponding to the non-motion state is subjected to low-pass filtering at the first frequency to obtain the first filtered acceleration signal; The forward acceleration signal corresponding to the normal motion state is subjected to low-pass filtering at the third frequency to obtain the second filtered acceleration signal; The forward acceleration signal corresponding to the high-intensity motion state is subjected to low-pass filtering at the third frequency or band-pass filtering from the second frequency to the third frequency to obtain the third filtered acceleration signal.

4. The respiratory rate detection method according to claim 3, wherein: The first frequency is 0.5 Hz; the second frequency is 1 Hz; the third frequency is 2 Hz; The first acceleration is 0.01g; the second acceleration is 0.1g.

5. The respiratory rate detection method according to any one of claims 1 to 4, characterized in that: The forward acceleration signal is subjected to spectrum analysis through Fourier transform.

6. The respiratory rate detection method according to claim 3 or 4, characterized in that: The respiratory wave reconstruction first detects the peak points of the first filtered acceleration signal, the second filtered acceleration signal, or the third filtered acceleration signal, and marks the timestamps of the peak points; Then, the respiratory wave is reconstructed according to each peak point with the time stamp.

7. The respiratory rate detection method according to claim 6, characterized in that: The discrete peak points with the timestamp are used to generate an envelope using a cubic spline function, and the envelope is made consistent with the sampling frequency of the forward acceleration signal to reconstruct a response curve; The response curve in the non-exercise state is subjected to low-pass filtering at the first frequency to obtain the respiratory wave; The response curve in the normal exercise state is subjected to low-pass filtering at the second frequency to obtain the respiratory wave; the response curve in the high-intensity exercise state is subjected to low-pass filtering at the third frequency to obtain the respiratory wave.

8. The respiratory rate detection method according to claim 7, characterized in that: Pass-through b_r = Ts*60 / △t (1) Solve for breathing rate; Wherein, b_r is the respiratory rate, the unit is bpm (times / minute); Ts is the sampling frequency of the forward acceleration signal; △t is the epoch interval of each respiratory cycle.

9. A respiratory rate detection device, characterized in that: It includes an accelerometer, a forward acceleration acquisition module, a respiratory wave reconstruction module, and a respiratory rate calculation and output module, and implements the respiratory rate detection method based on the accelerometer according to any one of claims 1 to 8; The accelerometer is a single-axis or multi-axis accelerometer, used to measure the acceleration of the user in a single-axis or multi-axis direction; The forward acceleration acquisition module is connected to the accelerometer, acquires the acceleration measured by the accelerometer, and superimposes the components of the acceleration in the forward direction to obtain a forward acceleration signal; The respiratory wave reconstruction module is connected to the forward acceleration acquisition module, receives the forward acceleration signal, and reconstructs the respiratory wave according to the forward acceleration signal; The respiratory rate calculation and output module is connected to the respiratory wave reconstruction module, receives the respiratory wave, and outputs the respiratory rate according to the respiratory wave.

10. A wearable device, characterized in that: Including the respiratory rate detection device according to claim 9.

Images