Motion monitoring method and device and wearable equipment

By integrating the analysis of pulse signals, signal strength and motion signals, obtaining multiple vital sign data and making fusion judgments, the problems of low data accuracy and high misjudgment rate in the existing technology are solved, and accurate monitoring of swimmers' drowning status is achieved .

CN118986311APending Publication Date: 2024-11-22JIAXING WENXIN INTELLIGENT TECH CO LTD +1
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202310581944.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-05-19
Publication Date
2024-11-22

AI Technical Summary

Technical Problem

The prior art provides low data accuracy and high misjudgment rate when monitoring swimmers due to sports artifact interference, making it difficult to accurately judge the drowning state.

Method used

By integrating the analysis of pulse signals, signal intensity and motion signals, we can obtain various vital sign data such as heart rate, blood oxygen, body movement posture and hazard index, filter motion interference, and reduce the misjudgment rate through the fusion of multiple data.

Benefits of technology

It improves the monitoring accuracy of vital sign data, reduces the misjudgment rate, meets the user's usage needs, and ensures accurate judgment of drowning status.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118986311A_ABST
    Figure CN118986311A_ABST
Patent Text Reader

Abstract

The invention provides a motion monitoring method and device and wearable equipment. The monitoring method comprises the following steps: acquiring a pulse signal, signal intensity and / or a motion signal of a monitoring target; acquiring vital sign data of the monitoring target according to the pulse signal, the signal intensity and / or the motion signal; performing fusion judgment on the vital sign data to obtain a current state of the monitoring target; therefore, the problems of low data precision and high misjudgment rate in the prior art are solved, various vital sign data such as the heart rate, the blood oxygen, the body movement posture and the drowning danger index of the monitored target are obtained through integrated analysis of the pulse signal, the signal intensity and the motion signal, interference of the motion state is filtered out, and the accuracy of the monitored target is improved. The monitoring precision of the vital sign data is improved; according to the method, whether the monitored target is in the drowning state or the suspected drowning state at present is judged through fusion of the multiple vital sign data, so that the misjudgment rate is reduced, and the use requirements of the user are met.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a motion monitoring method, device and wearable device. Background Art

[0002] With the rapid development of social economy and the improvement of people's living standards, swimming, a full-body sport suitable for the public, is becoming more and more popular. However, since swimming is done in water, it is dangerous. Beginners often choke on water and lose balance because they cannot breathe and move freely in the water, which leads to drowning accidents. Sometimes swimmers may also drown due to cramps, collisions, diseases or mental stress.

[0003] In order to solve the drowning problem, drowning is currently determined by monitoring the swimmer's heart rate, breathing rate or blood oxygen. However, the signals collected in the motion state are easily interfered by motion artifacts, and the accuracy is low. In addition, there is a high misjudgment rate when combined with a single judgment method, which not only increases the burden on survivors, but also reduces the user experience. Summary of the invention

[0004] The present application provides a motion monitoring method, apparatus and wearable device to solve the problems of low data accuracy and high misjudgment rate in the prior art. By integrating and analyzing pulse signals, signal strength and / or motion signals, various vital sign data such as heart rate, blood oxygen, body posture and danger index of the monitored target are obtained, thereby filtering out the interference of the motion state and improving the monitoring accuracy of the vital sign data. By fusing and judging various vital sign data, it is possible to determine whether the monitored target is currently in a drowning state or a suspected drowning state, thereby reducing the misjudgment rate and meeting the user's usage needs.

[0005] In a first aspect, the present invention provides a motion monitoring method, which includes: obtaining a pulse signal, a signal strength and / or a motion signal of a monitored target; obtaining vital signs data of the monitored target based on the pulse signal, the signal strength and / or the motion signal; and fusing and judging the vital signs data to obtain a current state of the monitored target.

[0006] Optionally, the vital sign data includes one or more of heart rate data, blood oxygen data, body movement posture and risk index; or / and, the current state includes a static state, a normal movement state, a suspected dangerous state or a dangerous state.

[0007] Optionally, the heart rate data of the monitored target is obtained according to the pulse signal and the motion signal, including: preprocessing and frequency domain transforming the pulse signal to obtain target frequency domain data; obtaining a current frequency domain search range according to the motion signal; and obtaining the heart rate data of the monitored target according to the current frequency domain search range and the target frequency domain data.

[0008] Optionally, the pulse signal is preprocessed and frequency-domain transformed to obtain target frequency-domain data, including: performing fast Fourier transform on the preprocessed pulse signal to obtain initial frequency-domain data; and performing linear filtering on the initial frequency-domain data to obtain the target frequency-domain data.

[0009] Optionally, based on the motion signal, the current frequency domain search range is obtained, including: bandpass filtering the motion signal to obtain a preprocessed motion signal; statistics the preprocessed motion signal in the current time window to obtain cadence data; input the cadence data and historical heart rate information into an adaptive heart rate search model to obtain the current frequency domain search range.

[0010] Optionally, when the historical heart rate information includes a historical spectrum peak index reference and a historical frequency domain search range, the formula of the adaptive heart rate search model is expressed as:

[0011] in, Historical peak index benchmark, is the frequency domain search range at the previous moment, Upper bound variables, is the lower bound variable, St (t) Search lower bound for the current frequency domain, Ed (t) Ns is the upper bound of the current frequency domain search. (t) is the current frequency domain search range.

[0012] Optionally, the heart rate data of the monitored target is obtained according to the current frequency domain search range and the target frequency domain data, including: according to the target frequency domain data, taking the largest frequency domain data in the current frequency domain search range as the main frequency; and obtaining the heart rate data of the monitored target according to the main frequency.

[0013] Optionally, acquiring the blood oxygen data of the monitored target according to the pulse signal includes: filtering the pulse signal of the current time window to obtain a red light DC component, an infrared light DC component, a red light AC component and an infrared light AC component; obtaining a current blood oxygen characteristic value according to the red light DC component, the infrared light DC component, the red light AC component and the infrared light AC component; obtaining the blood oxygen data of the monitored target according to the current blood oxygen characteristic value and the historical blood oxygen average.

[0014] Optionally, a current blood oxygen characteristic value is obtained according to the red light DC component, the infrared light DC component, the red light AC component and the infrared light AC component, including: obtaining a first intermediate value according to the ratio of the infrared light DC component to the red light DC component; performing spectral peak searches on the red light AC component and the infrared light AC component respectively to obtain a red light main spectrum amplitude and an infrared light main spectrum amplitude; obtaining a second intermediate value according to the ratio of the red light main spectrum amplitude to the infrared light main spectrum amplitude; and obtaining the current blood oxygen characteristic value according to the product of the first intermediate value and the second intermediate value.

[0015] Optionally, the blood oxygen data of the monitored target is obtained according to the current blood oxygen characteristic value and the historical blood oxygen average, including: obtaining a blood oxygen calculation value according to the current blood oxygen characteristic value; when the difference between the blood oxygen calculation value and the historical blood oxygen average is within a threshold range, using the blood oxygen calculation value as the blood oxygen data of the monitored target; when the difference between the blood oxygen calculation value and the historical blood oxygen average is not within a threshold range, using the historical blood oxygen average as the blood oxygen data of the monitored target.

[0016] Optionally, obtaining the danger index of the monitored target according to the signal strength and the motion signal includes: performing feature extraction on the motion signal and the signal strength to obtain acceleration characteristic parameters and current water immersion depth characteristic values; inputting the acceleration characteristic parameters and the current water immersion depth characteristic values ​​into a machine learning model for calculation to obtain a current drowning probability; if the current drowning probability is greater than a first threshold, and the drowning probabilities within a continuous preset time period are greater than the first threshold, obtaining the current drowning initial danger index according to the drowning initial danger index of the previous time window and the current drowning probability.

[0017] Optionally, the acceleration characteristic parameters include one or more of a combined acceleration mean, a standard deviation, a skewness, a combined acceleration spectrum entropy, a main frequency and a spectrum peak.

[0018] Optionally, the method also includes: if the current water immersion depth characteristic value is greater than a second threshold, and the water immersion depth characteristic values ​​within a continuous preset time period are greater than the second threshold, calculating the current late drowning hazard index based on the late drowning hazard index of the previous time window and the combined acceleration spectrum entropy.

[0019] Optionally, feature extraction is performed on the motion signal to obtain acceleration feature parameters, including: filtering the motion signal to calculate the combined acceleration; performing Fourier transform on the combined acceleration to obtain frequency domain combined acceleration; and calculating the combined acceleration spectrum entropy based on the frequency domain combined acceleration.

[0020] 15. The motion monitoring method according to claim 14, characterized in that the formula for calculating the spectrum entropy of the combined acceleration is:

[0021]

[0022] Where P(w) represents the total power spectrum, A i (w) represents the amplitude of the ith frequency band in the frequency domain combined acceleration, i = [1, 2, ..., N], N represents the total number of frequency bands, p(i) represents the probability density of the ith frequency band, and psdE represents the spectral entropy of the combined acceleration.

[0023] Optionally, the body posture of the monitored target is obtained according to the motion signal, including: calculating a first angle between the X-axis and the direction of gravity, a second angle between the Y-axis and the direction of gravity, and a third angle between the Z-axis and the direction of gravity according to the motion signal; and obtaining the body posture of the monitored target according to the first angle, the second angle and the third angle.

[0024] Optionally, the body movement posture of the monitored target is obtained according to the motion signal, including: filtering the motion signal and calculating the combined acceleration; taking the absolute value of the combined acceleration minus the gravitational acceleration as the acceleration standard deviation; and obtaining the body movement posture of the monitored target according to the acceleration standard deviation.

[0025] In a second aspect, the present invention provides a motion monitoring device, which includes: a data acquisition module for acquiring a pulse signal, a signal strength and / or a motion signal of a monitored target; a data processing module for acquiring vital signs data of a monitored target based on the pulse signal, the signal strength and / or the motion signal; and a fusion judgment module for performing fusion judgment on the vital signs data to obtain a current state of the monitored target.

[0026] In a third aspect, the present invention provides a wearable device, wherein the wearable device comprises the motion monitoring device.

[0027] Compared with the prior art, this application has the following beneficial effects:

[0028] The present invention obtains various vital sign data of the monitored target, such as heart rate, blood oxygen, body posture and drowning risk index, through integrated analysis of pulse signals, signal strength and / or motion signals, thereby filtering out interference from the motion state, improving the monitoring accuracy of the vital sign data, and providing a guarantee for subsequent data analysis; further, the present invention obtains whether the monitored target is currently in a drowning state or a suspected drowning state by fusing and judging various vital sign data, thereby reducing the misjudgment rate and meeting the user's usage needs. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 The figure is a schematic diagram of a flow chart of a motion monitoring method provided in an embodiment of the present application;

[0030] Figure 2 FIG. 1 is a flow chart of another motion monitoring method provided in an embodiment of the present application;

[0031] Figure 3 The figure shows a flow chart of the second embodiment provided in the embodiment of the present application;

[0032] Figure 4 Another schematic diagram of the process of the second embodiment provided in the embodiment of the present application is shown;

[0033] Figure 5 Shown is a schematic diagram of a PPG frequency domain waveform provided in an embodiment of the present application;

[0034] Figure 6 The figure shows a flow chart of the third embodiment provided in the embodiment of the present application;

[0035] Figure 7 Another schematic diagram of the process of the third embodiment of the present invention is shown;

[0036] Figure 8 Shown is a red light spectrum diagram and an infrared light spectrum diagram provided in an embodiment of the present application;

[0037] Fig. 9 The figure shows a flow chart of the fourth embodiment provided in the embodiments of the present application;

[0038] Fig.10 Another schematic diagram of the process of the fourth embodiment provided in the present application is shown;

[0039] Fig.11 The figure shows a flow chart of the fifth embodiment provided in the embodiment of the present application;

[0040] Fig.12 The figure shows a three-axis direction schematic diagram of an acceleration sensor provided in an embodiment of the present application;

[0041] Fig.13 The figure shows a schematic diagram of the change of the angles between the three axes and the gravity direction provided in an embodiment of the present application;

[0042] Fig.14 Shown is a schematic diagram of a psdE change curve during running provided in an embodiment of the present application. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0044] In a first aspect, the present invention provides a motion monitoring method, which specifically includes the following embodiments:

[0045] Embodiment 1

[0046] Figure 1 FIG. 1 is a flow chart of a motion monitoring method provided by an embodiment of the present application, such as Figure 1 As shown, the monitoring method specifically comprises the following steps:

[0047] Step S101, obtaining a pulse signal, a signal strength and / or a motion signal of a monitoring target.

[0048] It should be noted that a signal acquisition module is worn on the monitored target, and the signal acquisition module integrates a wireless transmission unit, a PPG (Photo Plethysmo Graphy) sensor and an acceleration sensor; wherein the wireless transmission unit is used to obtain the RSSI value of the signal currently received from the fixed base station, the PPG sensor is used to collect the current PPG signal of the monitored target, and the acceleration sensor is used to collect the current acceleration signal; wherein the pulse signal in this embodiment includes but is not limited to the PPG signal, the motion signal includes but is not limited to the acceleration signal, and the signal strength is the RSSI value.

[0049] Step S102, acquiring vital sign data of the monitored target according to the pulse signal, the signal strength and / or the motion signal.

[0050] In this embodiment, the vital signs data include heart rate data, blood oxygen data, body movement posture and risk index; wherein the body movement posture includes but is not limited to static state and moving state, and the risk index includes but is not limited to early drowning risk index and late drowning risk index.

[0051] In this embodiment, if Figure 2As shown, obtaining the vital signs data of the monitored target based on the pulse signal, the signal strength and / or the motion signal specifically includes: obtaining the heart rate data of the monitored target based on the pulse signal and the motion signal; obtaining the blood oxygen data of the monitored target based on the pulse signal; obtaining the danger index of the monitored target based on the signal strength and the motion signal; and obtaining the body movement posture of the monitored target based on the motion signal.

[0052] Step S103, fusing and judging the vital sign data to obtain the current state of the monitored target.

[0053] The current state in this embodiment includes a static state, a normal motion state, a suspected dangerous state or a dangerous state; wherein the suspected dangerous state includes a suspected drowning state, and the dangerous state includes a drowning state.

[0054] It should be noted that the specific process of fusing and judging vital signs data includes:

[0055] (1) If the heart rate data is within a normal preset range, the blood oxygen data is within a normal preset range, and the body posture is in a static state, it is determined that the monitored target is currently in a static state.

[0056] (2) when the heart rate data is within a normal preset range, the blood oxygen data is within a normal preset range, the body movement posture is in a motion state, and the drowning risk index is less than a warning threshold, it is determined that the monitored target is currently in a normal motion state;

[0057] (3) when the heart rate data is within a normal preset range, the blood oxygen data is within a normal preset range, the body movement posture is in motion, and the drowning risk index is greater than a warning threshold, it is determined that the monitored target is currently in a suspected drowning state;

[0058] (4) When the heart rate data is not within the normal preset range, the blood oxygen data is not within the normal preset range, the body movement posture is in motion, and the drowning risk index is greater than the risk threshold, it is determined that the monitored target is currently in a drowning state.

[0059] Compared with the prior art, this embodiment has the following beneficial effects:

[0060] The present invention obtains various vital sign data of the monitored target, such as heart rate, blood oxygen, body posture and drowning risk index, through integrated analysis of pulse signals, signal strength and / or motion signals, thereby filtering out interference from the motion state, improving the monitoring accuracy of the vital sign data, and providing a guarantee for subsequent data analysis; further, the present invention obtains whether the monitored target is currently in a drowning state or a suspected drowning state by fusing and judging various vital sign data, thereby reducing the misjudgment rate and meeting the user's usage needs.

[0061] Embodiment 2

[0062] like Figure 3 As shown, obtaining the heart rate data of the monitored target according to the pulse signal and the motion signal specifically includes the following steps:

[0063] Step S201, preprocessing and frequency domain transformation of the pulse signal to obtain target frequency domain data;

[0064] Step S202, obtaining a current frequency domain search range according to the motion signal;

[0065] Step S203, obtaining the heart rate data of the monitored target according to the current frequency domain search range and the target frequency domain data.

[0066] In this embodiment, the pulse signal is preprocessed and frequency-domain transformed to obtain target frequency-domain data, including: performing fast Fourier transform on the preprocessed pulse signal to obtain initial frequency-domain data; and performing linear filtering on the initial frequency-domain data to obtain the target frequency-domain data.

[0067] In this embodiment, the current frequency domain search range is obtained according to the motion signal, including: bandpass filtering the motion signal to obtain a preprocessed acceleration signal; statistics the preprocessed motion signal in the current time window to obtain cadence data; input the cadence data and historical heart rate information into an adaptive heart rate search model to obtain the current frequency domain search range.

[0068] In this embodiment, the heart rate data of the monitored target is obtained according to the current frequency domain search range and the target frequency domain data, including: according to the target frequency domain data, taking the largest frequency domain data in the current frequency domain search range as the main frequency; and obtaining the heart rate data of the monitored target according to the main frequency.

[0069] It should be noted that if Figure 4 As shown in the figure, first, the acceleration signal is used as a noise reference, and the PPG signal is smoothed by a 0.4-4 Hz bandpass filter; then the smoothed PPG signal is subjected to a fast Fourier transform, and the frequency domain data of 1-3 Hz is intercepted, corresponding to the heart rate variation range of 60-180 bpm; finally, the frequency domain data is filtered again using a Wiener filter to obtain clean PPG frequency domain data.

[0070] In addition, the acceleration data is filtered through a 0.5-5Hz bandpass filter, and the number of peaks of the acceleration signal in the current time window is counted to calculate the cadence data; the cadence and historical heart rate information are input into the adaptive heart rate search model to obtain the current frequency domain search range, and the main frequency in the target frequency domain data is obtained according to the current frequency domain search range. Finally, the heart rate data of the current time window is obtained according to the main frequency.

[0071] It should be noted that the adaptability of the heart rate search model is achieved through the dynamic adjustment of the cadence and historical heart rate information. Cadence reflects the intensity of exercise, and then reflects the trend of heart rate change, which is the first element in the establishment of the adaptive heart rate search model. Combined with the experiment, it is found that the change of cadence can reflect and affect the change of heart rate to a large extent. Determining the range of the current heart rate through the historical heart rate is the second element in the establishment of the adaptive heart rate search model. The change of human heart rate has a certain range, and the difference between the newly calculated heart rate and the previous heart rate will not be too large. Predicting the heart rate change gradient through the change characteristics of exercise heart rate is the third element in the establishment of the adaptive heart rate search model. In daily activities, the range of human heart rate changes is basically maintained between 60 and 180bpm.

[0072] In this embodiment, when the historical heart rate information includes a historical spectrum peak index reference and a historical heart rate search range, the formula of the adaptive heart rate search model is expressed as:

[0073]

[0074]

[0075] N (t) =Ed (t) -St (t) (3)

[0076] Among them, Mid (t-1) Historical peak index benchmark, Ns (t-1) is the frequency domain search range at the previous moment, Upper bound variables, is the lower bound variable, St (t) Search lower bound for the current frequency domain, Ed (t) Ns is the upper bound of the current frequency domain search. (t) is the current frequency domain search range. In order to obtain the optimal and This embodiment uses the IPSO algorithm model to perform parameter optimization.

[0077] In this embodiment, according to the current frequency domain search range and the target frequency domain data, the calculation formula for obtaining the heart rate data of the monitored target is:

[0078] Mid(t) = argmax(FFT PPG [St (t) :Ed (t) ]) (4)

[0079] HR (t) =Mid (t) ×60 (5)

[0080] Among them, Mid (t) Indicates the current main frequency, HR (t) Represents the heart rate data of the current time window.

[0081] It should be noted that the performance of the preprocessed PPG signal in the frequency domain in this window is as follows Figure 5 As shown. For the convenience of observation, the horizontal axis is converted from frequency to heart rate according to formula (5). The adaptive heart rate search model predicts and limits the heart rate range (the upper and lower limits of the limit range are as follows Figure 5 The calculated heart rate is 135bpm, which is very close to the real heart rate of 133bpm.

[0082] This embodiment has the following beneficial effects: this embodiment combines the acceleration cadence data and historical heart rate information, and obtains a reliable frequency domain search range through an adaptive heart rate search model, so that an accurate heart rate value can be calculated based on the reliable frequency domain search range, thereby improving the calculation accuracy of the heart rate.

[0083] Embodiment 3

[0084] like Figure 6 As shown, obtaining the blood oxygen data of the monitored target according to the pulse signal specifically includes the following steps:

[0085] Step S301, filtering the pulse signal of the current time window to obtain a red light DC component, an infrared light DC component, a red light AC component and an infrared light AC component;

[0086] Step S302, obtaining a current blood oxygen characteristic value according to the red light DC component, the infrared light DC component, the red light AC component and the infrared light AC component;

[0087] Step S303, obtaining blood oxygen data of the monitored target according to the current blood oxygen characteristic value and the historical blood oxygen mean value.

[0088] In this embodiment, a current blood oxygen characteristic value is obtained according to the red light DC component, the infrared light DC component, the red light AC component and the infrared light AC component, including: obtaining a first intermediate value according to the ratio of the infrared light DC component to the red light DC component; performing spectral peak searches on the red light AC component and the infrared light AC component respectively to obtain a red light main spectrum amplitude and an infrared light main spectrum amplitude; obtaining a second intermediate value according to the ratio of the red light main spectrum amplitude to the infrared light main spectrum amplitude; and obtaining the current blood oxygen characteristic value according to the product of the first intermediate value and the second intermediate value.

[0089] In this embodiment, blood oxygen data of the monitored target is obtained according to the current blood oxygen characteristic value and the historical blood oxygen average, including: obtaining a blood oxygen calculation value according to the current blood oxygen characteristic value; when the difference between the blood oxygen calculation value and the historical blood oxygen average is within a threshold range, using the blood oxygen calculation value as the blood oxygen data of the monitored target; when the difference between the blood oxygen calculation value and the historical blood oxygen average is not within the threshold range, using the historical blood oxygen average as the blood oxygen data of the monitored target.

[0090] It should be noted that if Figure 7 As shown, firstly, the red light signal and the infrared light signal are subjected to 0.8 Hz low-pass filtering to obtain the DC component, and the first intermediate value val1 is calculated:

[0091] val1=IR dc / R dc (6)

[0092] Among them, R dc is the DC component of red light, IR dc is the DC component of infrared light.

[0093] Then, the red light AC component R is obtained by bandpass filtering from 0.8Hz to 4Hz. ac and infrared light AC component IR ac The AC component is converted to the frequency domain using linear frequency modulation z-transform, and the spectrum peak search is performed to obtain the main frequency amplitude R of the red light. fval , infrared light main frequency amplitude IR fval , calculate the second intermediate value val2:

[0094] val2=R fval / IR fval (7)

[0095] The formula for calculating the current blood oxygen characteristic value r is:

[0096] r=val1×val2 (8)

[0097] According to the current blood oxygen characteristic value, the formula for obtaining the blood oxygen calculation value Spo2 is:

[0098] Spo2=a×r 2 +b×r+c (9)

[0099] Among them, a=-45.060, b=30.354, c=92.845, and the values ​​of a, b, and c are MAXIM.

[0100] Since the blood oxygen value will not change suddenly in a short period of time, adjacent blood oxygen values ​​are usually close. The abnormal blood oxygen value is corrected by the historical blood oxygen value. The specific modification method is: obtain the mean blood oxygen value of the three time windows before the current time window, and output the blood oxygen calculation value if the current blood oxygen calculation value falls within the error range of the historical blood oxygen mean; if the current blood oxygen calculation value does not fall within the error range of the historical blood oxygen mean, output the historical blood oxygen mean.

[0101] For example, taking this window as an example, the pre_spo2 of the current window is 97.4667. The preprocessed PPG signal is low-pass filtered and band-pass filtered respectively, and substituted into formula (6) to obtain val1 = 0.8186. The PPG signal is converted to the frequency domain to obtain the red light and infrared light spectra as follows Figure 8 As shown in Fig. 8a and Fig. 8b, the peak of the red spectrum is 3577.21, and the peak of the infrared spectrum is 6684.75. Substituting into formula (7), val2 = 3577.21 / 6684.75 = 0.5351 is obtained. From formula (8) and formula (9), r = 0.4380 and Spo2 = 97.5 can be obtained. The difference with pre_spo2 falls within the threshold, and it is considered that the current calculated blood oxygen value is valid and no correction is required.

[0102] This embodiment has the following beneficial effects: this embodiment obtains a red light DC component, an infrared light DC component, a red light AC component and an infrared light AC component by performing low-pass filtering and band-pass filtering on the PPG signal respectively, and then performs fusion operation on the red light DC component, the infrared light DC component, the red light AC component and the infrared light AC component to obtain the current blood oxygen value, and finally corrects the current blood oxygen value according to the historical blood oxygen average, so as to obtain high-precision blood oxygen data.

[0103] Embodiment 4

[0104] like Fig. 9 As shown, obtaining the danger index of the monitored target according to the signal strength and the motion signal specifically includes the following steps:

[0105] Step S401, extracting features from the motion signal and the signal strength to obtain acceleration feature parameters and current water entry depth feature values;

[0106] Step S402, inputting the acceleration characteristic parameter and the current water entry depth characteristic value into a machine learning model for calculation to obtain a current drowning probability;

[0107] Step S403, if the current drowning probability is greater than a first threshold, and the drowning probability in a continuous preset time period is greater than the first threshold, the current drowning initial risk index is obtained according to the drowning initial risk index of the previous time window and the current drowning probability.

[0108] It should be noted that the method of obtaining the drowning probability includes but is not limited to the random forest model. In other embodiments, it can be obtained through machine learning models such as CNN convolutional neural network, SVM support vector machine, etc., or through function fitting.

[0109] In another embodiment, the method further includes: if the current water immersion depth characteristic value is greater than a second threshold, and the water immersion depth characteristic values ​​within a continuous preset time period are greater than the second threshold, calculating the current late drowning hazard index based on the late drowning hazard index of the previous time window and the combined acceleration spectrum entropy.

[0110] In this embodiment, feature extraction is performed on the motion signal to obtain acceleration feature parameters, including: filtering the motion signal to calculate the combined acceleration; performing Fourier transform on the combined acceleration to obtain frequency domain combined acceleration; and calculating the combined acceleration spectrum entropy based on the frequency domain combined acceleration.

[0111] In this embodiment, the acceleration characteristic parameters include but are not limited to the combined acceleration mean A, standard deviation std, skew, combined acceleration spectrum entropy psdE, main frequency f and spectrum peak fval.

[0112] It should be noted that if Fig.10 As shown, firstly, the acceleration data is low-pass filtered at 5Hz, and the combined acceleration is calculated by the acceleration of the x, y, and z axes:

[0113]

[0114] In this embodiment, swimming is considered as an orderly movement, and the struggle before drowning is considered as a disordered movement. Compared with swimming, the disorder before drowning increases. The standard deviation can be used as a criterion for disorder. In order to evaluate more accurately, the entropy of the combined acceleration spectrum can be calculated as a feature. First, the signal power spectrum is calculated. N (n) Perform Fourier transform to get A N(w), take the square of its amplitude and divide it by the total number of frequency bands N to obtain A N The power spectrum estimation of (n) is expressed as:

[0115]

[0116] Calculate the proportion of energy in each frequency band to the total energy and get the probability density function p(i), whose formula is:

[0117]

[0118] The formula for calculating the combined acceleration spectrum entropy is:

[0119] psdE=-sum(p(i)×log2p(i)) (13)

[0120] Where P(w) represents the total power spectrum, A i (w) represents the amplitude of the ith frequency band in the frequency domain combined acceleration, i = [1, 2, ..., N], N represents the total number of frequency bands, p(i) represents the probability density of the ith frequency band, and psdE represents the spectral entropy of the combined acceleration.

[0121] Furthermore, in this embodiment, the water entry depth characteristic value includes an RSSI characteristic value, and the distance is estimated by RSSI. The calculation formula for calculating the RSSI characteristic value is:

[0122]

[0123] Among them, RSSI is the received signal strength, H is the signal strength when the transmitter and receiver are 1 meter apart, which can be obtained by actual measurement, and n is the environmental attenuation factor, which needs to be tested and corrected. According to experience, the best range is between 3.25-4.5. Due to different environments, the corresponding parameter values ​​of each transmitter are different, and each parameter in the formula should be obtained through experimental calibration.

[0124] It should be noted that in the early stage of drowning, the human body is in a state of struggle, which can be regarded as a state of disordered movement. The above extracted features are linearly fitted to output the drowning probability and calculate the autonomous consciousness risk index in the early stage of drowning. In the late stage of drowning, the drowning person is unable to struggle, loses the ability to move autonomously, and sinks in the water. The RSSI feature becomes larger and the spectrum entropy decreases. Combining the RSSI feature and the combined acceleration spectrum entropy, the objective consciousness risk index in the late stage of drowning is calculated.

[0125] Taking this window as an example, the tester is currently in a sitting state. The acceleration signal is filtered and features are extracted to obtain V A=0.0423, RSSI=-49.4350, A=1.0423, std=0.0025, skew=-0.1168, d=61.0766, fval=0.0013, f=3.3750, psdE=0.0064. After linear fitting, the drowning probability pro=0 is obtained, which is less than the set threshold of 0.65, and the pro values ​​of the first 10 seconds of the current window are all less than the threshold, and the output rate1=0 is consistent with the actual sitting state. At this time, d=61.0766, which is less than the set threshold of 200, and the d values ​​of the first 10 seconds of the current window are all less than the threshold, and the output rate2=0 is consistent with the actual sitting state.

[0126] Another window is captured, where the tester is struggling in the early stage of simulating drowning. The acceleration signal is filtered and features are extracted to obtain V A =0.2291, RSSI=-46.0100, A=1.1191, std=0.2509, skew=-0.1346, d=50.01476, fval=0.1670, f=1.7500, psdE=5.3372. After linear fitting, the drowning probability pro=0.96 is obtained, which is greater than the set threshold of 0.65, and the pro values ​​of the first 10 seconds of the current window are all greater than the threshold. The output rate1=1 is suspected of drowning. At this time, d=61.0766, which is less than the threshold of 200, and the d values ​​of the first 10 seconds of the current window are all less than the threshold. The output rate2=0 is consistent with the state of suspected drowning but not sinking into the water.

[0127] This embodiment has the following beneficial effects: this embodiment performs feature extraction on the acceleration signal and the RSSI value to obtain acceleration feature parameters and current RSSI feature values, then performs fusion judgment on the acceleration feature parameters and current RSSI feature values, and outputs a danger index for an early stage of drowning and a danger index for a late stage of drowning; this embodiment adopts different judgment methods for the early stage of drowning and the late stage of drowning, thereby improving the judgment accuracy of the drowning state; and analyzes the danger index to ensure the golden rescue time.

[0128] Embodiment 5

[0129] like Fig.11 As shown, according to the motion signal, obtaining the body motion posture of the monitoring target specifically includes the following steps:

[0130] Step S401, calculating, according to the motion signal, a first angle between the X-axis and the direction of gravity, a second angle between the Y-axis and the direction of gravity, and a third angle between the Z-axis and the direction of gravity;

[0131] Step S402: obtaining the body motion posture of the monitored target according to the first angle, the second angle and the third angle.

[0132] It should be noted that the formula for calculating the angle between the x, y, and z axes and the direction of gravity is:

[0133] θ x =arccos(x / g)*180 / π (15)

[0134] θ y =arccos(y / g)*180 / π (16)

[0135] θ z =arccos(z / g)*180 / π (17)

[0136] In order to reduce the amount of calculation and omit the trigonometric function calculation, the angle is mapped to the [-1,1] region. The angle calculation formula between the x, y, and z axes and the gravity direction can also include:

[0137] θ x =x / g (18)

[0138] θ y =y / g (19)

[0139] θ z = z / g (20)

[0140] It should be noted that, taking the motion signal as an acceleration signal and the acceleration sensor worn on the forehead as an example, the acceleration sensor is fixed on the chip in the direction Fig.12 As shown. Fig.14 It can be seen that when the human body is in an upright state, the y-axis is parallel to the direction of gravity, with an angle close to 180°, and the x-axis and z-axis are perpendicular to the direction of gravity, with an angle close to 90°; when the head is lowered, the y-axis and x-axis are perpendicular to the direction of gravity, with an angle close to 90°, and the z-axis is parallel to the direction of gravity, with an angle close to 180°; when the head is raised, the y-axis and x-axis are perpendicular to the direction of gravity, with an angle close to 90°, and the z-axis is parallel to the direction of gravity, with an angle close to 0°. According to the angles between the three directions and gravity, the angular posture of the human body can be judged. For example, when swimming, the human body is in a horizontal state, which can be distinguished from standing vertically.

[0141] In another embodiment, obtaining the body posture of the monitored target based on the motion signal also includes: calculating the combined acceleration after filtering the motion signal; taking the absolute value of the combined acceleration minus the gravity acceleration as the acceleration standard deviation; and obtaining the body posture of the monitored target based on the acceleration standard deviation.

[0142] It should be noted that when the human body is standing still, the acceleration is concentrated on the vertical y-axis, and the y-axis acceleration signal tends to the gravity acceleration g, and there is almost no acceleration in the x-axis and z-axis directions. The standard deviation of acceleration obtained by subtracting the absolute value of g from the total acceleration is calculated as one of the features, and its calculation formula is:

[0143] V A =|A N (n)-g| (21)

[0144] In the static state, only gravity acts on V A tends to 0; therefore, when the acceleration standard deviation V is calculated A When it approaches 0, the body motion posture of the monitored target is static.

[0145] This embodiment has the following beneficial effects: This embodiment adopts a dual calculation method for the acceleration signal, so that the obtained body motion posture is more accurate, providing a guarantee for subsequent data analysis.

[0146] In another embodiment, when the motion state is running, the acceleration signal is extracted to calculate psdE to obtain Fig.14 From the psdE variation curve shown, we can see that the exerciser first sits still for a period of time, and then exercises, with the speed gradually increasing from slow walking to jogging. As the intensity of the exercise increases, the psdE gradually increases.

[0147] In a second aspect, the present invention provides a drowning event monitoring device, the monitoring device comprising:

[0148] A data acquisition module, used to obtain the pulse signal, signal strength and / or motion signal of the monitored target;

[0149] A data processing module, used for acquiring vital sign data of a monitored target according to the pulse signal, the signal strength and / or the motion signal;

[0150] The fusion judgment module is used to perform fusion judgment on the vital sign data to obtain the current state of the monitored target.

[0151] It should be noted that the working principle of the drowning event monitoring device provided in this embodiment is the same as that of the above embodiments one to five, and will not be repeated here.

[0152] In a third aspect, the present invention provides a wearable device, comprising the above-mentioned motion monitoring device; wherein the wearable device includes but is not limited to a swimming cap.

[0153] Finally, it should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0154] The above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solutions of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A motion monitoring method, characterized in that: The monitoring method comprises: Acquire the pulse signal, signal strength and / or motion signal of the monitored target; Acquiring vital sign data of a monitored target according to the pulse signal, the signal strength and / or the motion signal; The vital sign data are fused and judged to obtain the current state of the monitored target.

2. The motion monitoring method according to claim 1, characterized in that: The vital sign data includes one or more of heart rate data, blood oxygen data, body movement posture and risk index; Or / and, the current state includes a static state, a normal motion state, a suspected dangerous state or a dangerous state.

3. The motion monitoring method according to claim 2, characterized in that: Acquiring heart rate data of a monitored target according to the pulse signal and the motion signal, including: Preprocessing and frequency domain transforming the pulse signal to obtain target frequency domain data; According to the motion signal, a current frequency domain search range is obtained; The heart rate data of the monitored target is obtained according to the current frequency domain search range and the target frequency domain data.

4. The motion monitoring method according to claim 3, characterized in that: Preprocessing and frequency domain transforming the pulse signal to obtain target frequency domain data includes: Performing a fast Fourier transform on the preprocessed pulse signal to obtain initial frequency domain data; The initial frequency domain data is linearly filtered to obtain the target frequency domain data.

5. The motion monitoring method according to claim 3, characterized in that: According to the motion signal, a current frequency domain search range is obtained, including: Performing bandpass filtering on the motion signal to obtain a preprocessed motion signal; Perform statistics on the preprocessed motion signal of the current time window to obtain the step frequency data; The step frequency data and historical heart rate information are input into an adaptive heart rate search model to obtain a current frequency domain search range.

6. The motion monitoring method according to claim 5, characterized in that: When the historical heart rate information includes a historical spectrum peak index reference and a historical frequency domain search range, the formula of the adaptive heart rate search model is expressed as: in, Historical peak index benchmark, is the frequency domain search range at the previous moment, The upper bound variable, is the lower bound variable, St (t) Search lower bound for the current frequency domain, Ed (t) Ns is the upper bound of the current frequency domain search. (t) is the current frequency domain search range.

7. The motion monitoring method according to claim 3, characterized in that: Obtaining the heart rate data of the monitored target according to the current frequency domain search range and the target frequency domain data, including: According to the target frequency domain data, the maximum frequency domain data in the current frequency domain search range is used as the main frequency; The heart rate data of the monitored target is obtained according to the main frequency.

8. The motion monitoring method according to claim 2, characterized in that: Acquiring blood oxygen data of the monitored target according to the pulse signal includes: Filtering the pulse signal of the current time window to obtain a red light DC component, an infrared light DC component, a red light AC component and an infrared light AC component; Obtaining a current blood oxygen characteristic value according to the red light DC component, the infrared light DC component, the red light AC component, and the infrared light AC component; The blood oxygen data of the monitored target is obtained according to the current blood oxygen characteristic value and the historical blood oxygen average value.

9. The motion monitoring method according to claim 8, characterized in that: Obtaining a current blood oxygen characteristic value according to the red light DC component, the infrared light DC component, the red light AC component, and the infrared light AC component, including: Obtaining a first intermediate value according to a ratio of the infrared light DC component to the red light DC component; Performing spectrum peak searches on the red light AC component and the infrared light AC component respectively to obtain the main spectrum amplitude of the red light and the main spectrum amplitude of the infrared light; Obtaining a second intermediate value according to a ratio of the amplitude of the main spectrum of the red light to the amplitude of the main spectrum of the infrared light; The current blood oxygen characteristic value is obtained according to the product of the first intermediate value and the second intermediate value.

10. The motion monitoring method according to claim 8, characterized in that: The blood oxygen data of the monitored target is obtained according to the current blood oxygen characteristic value and the historical blood oxygen average value, including: Obtaining a blood oxygen calculation value according to the current blood oxygen characteristic value; When the difference between the calculated blood oxygen value and the historical blood oxygen average value is within a threshold range, using the calculated blood oxygen value as the blood oxygen data of the monitoring target; When the difference between the calculated blood oxygen value and the historical blood oxygen average value is not within a threshold range, the historical blood oxygen average value is used as the blood oxygen data of the monitoring target.

11. The motion monitoring method according to claim 2, characterized in that: The step of obtaining a danger index of a monitoring target according to the signal strength and the motion signal includes: Extracting features of the motion signal and the signal strength to obtain acceleration feature parameters and current water entry depth feature values; Inputting the acceleration characteristic parameter and the current water entry depth characteristic value into the machine learning model for calculation to obtain the current drowning probability; If the current drowning probability is greater than a first threshold, and the drowning probability in a continuous preset time period is greater than the first threshold, the current drowning initial risk index is obtained according to the drowning initial risk index of the previous time window and the current drowning probability.

12. The motion monitoring method according to claim 11, characterized in that: The acceleration characteristic parameters include one or more of a combined acceleration mean, a standard deviation, a skewness, a combined acceleration spectrum entropy, a main frequency, and a spectrum peak.

13. The motion monitoring method according to claim 12, characterized in that: The method further comprises: If the current water immersion depth characteristic value is greater than a second threshold, and the water immersion depth characteristic values ​​within a continuous preset time period are all greater than the second threshold, the current late drowning hazard index is calculated based on the late drowning hazard index of the previous time window and the combined acceleration spectrum entropy.

14. The motion monitoring method according to claim 12, characterized in that: Feature extraction is performed on the motion signal to obtain acceleration feature parameters, including: After filtering the motion signal, a resultant acceleration is calculated; Performing Fourier transformation on the resultant acceleration to obtain a frequency domain resultant acceleration; The frequency spectrum entropy of the combined acceleration is calculated according to the combined acceleration in the frequency domain.

15. The motion monitoring method according to claim 14, characterized in that: According to the frequency domain combined acceleration, the calculation formula for the combined acceleration spectrum entropy is calculated as follows: Where P(w) represents the total power spectrum, A i (w) represents the amplitude of the ith frequency band in the frequency domain combined acceleration, i = [1, 2, ..., N], N represents the total number of frequency bands, p(i) represents the probability density of the ith frequency band, and psdE represents the spectral entropy of the combined acceleration.

16. The motion monitoring method according to claim 2, characterized in that: Acquiring the body motion posture of the monitored target according to the motion signal, including: According to the motion signal, respectively calculate a first angle between the X-axis and the direction of gravity, a second angle between the Y-axis and the direction of gravity, and a third angle between the Z-axis and the direction of gravity; The body motion posture of the monitored target is obtained according to the first angle, the second angle and the third angle.

17. The motion monitoring method according to claim 2, characterized in that: Acquiring the body motion posture of the monitored target according to the motion signal, including: After filtering the motion signal, a resultant acceleration is calculated; The absolute value of the total acceleration minus the acceleration due to gravity is taken as the acceleration standard deviation; The body motion posture of the monitored target is obtained according to the acceleration standard deviation.

18. A motion monitoring device, characterized in that: The monitoring device comprises: A data acquisition module, used to obtain the pulse signal, signal strength and / or motion signal of the monitored target; A data processing module, used for acquiring vital sign data of a monitored target according to the pulse signal, the signal strength and / or the motion signal; The fusion judgment module is used to perform fusion judgment on the vital sign data to obtain the current state of the monitored target.

19. A wearable device, characterized in that: The wearable device comprises the motion monitoring device according to claim 18.

Citation Information

Cited By

  • Indoor and outdoor health monitoring and positioning method

    CN120000183A