Paralysis patient active alarm method based on respiratory signal characteristic monitoring

Abstract

The invention discloses a paralyzed patient active alarm method based on respiratory signal characteristic monitoring, which comprises the following steps: detecting a human body signal by a PVDF film sensor, carrying out zero equalization and normalization processing, extracting a respiratory signal BR [k] in S [k], searching for a wave crest and a wave trough in the BR [k], and carrying out active alarm on a paralyzed patient according to the wave crest and the wave trough. The breathing signal acquisition module is used for acquiring a breathing signal of a single period from the beginning of current expiration to the beginning of next expiration, performing comparison calculation, setting a notification and warning event number according to a comparison result, and transmitting the notification and warning number to the notification and alarm module. The system has the advantages that the demand information transmitted by the patient can be detected according to different breathing modes of the patient, the demand of the patient is actively sensed, the efficiency and quality of nursing work are effectively improved, and the nursing cost is reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of paralysis monitoring, and particularly relates to a paralysis patient active alarm method based on respiratory signal feature monitoring. BACKGROUND

[0002] The monitoring device for paralysis patients has been relatively mature. Generally, the device uses sensor technology and embedded instrument technology to monitor physiological parameters of the patient so as to diagnose the current physiological state of the patient.

[0003] The monitoring device on the market currently generally has the functions of monitoring physiological parameters such as blood pressure, blood oxygen saturation, electrocardiogram, body temperature, pulse and respiration, and is mainly applied to hospital wards and family nursing. In addition, there is a kind of intelligent mattress system which can control the patient to turn over through a motor, and cooperates with an electrocardiograph, a sphygmomanometer, a blood glucose meter and an electronic thermometer to assist the nursing personnel to pay attention to the physiological condition of the patient and detect various physiological parameters.

[0004] However, the above device is expensive and too professional in function, and generally adopts a passive monitoring method. In other words, they cannot actively intervene or respond to the needs of the patient, and only trigger an alarm notification after an emergency occurs. Passive monitoring cannot prompt the nursing personnel to take measures in time before the problem occurs or when the patient has an active appeal, and an alarm is only sent after the problem has occurred, which will lead to a delayed response of the monitoring and requires the nursing personnel to check frequently.

[0005] In the rehabilitation stage of the paralysis patient, the disease is generally relatively stable, and the patient cannot move freely, so the patient mainly depends on the care of family members or nursing personnel in this stage. In this case, the traditional monitoring device has certain limitations because the monitoring device cannot actively deliver signals to respond to the physiological needs or emergency of the patient. In addition, the monitoring device may be mis-triggered, which greatly reduces the actual use efficiency.

[0006] Therefore, the focus of daily monitoring is not on the monitoring of complex physiological parameters, but on the subjective and timely demand information of the patient such as the basic life state, environmental comfort, bowel movement, turning over time and the like.

[0007] The demand information is usually actively proposed by the patient, but the traditional monitoring device cannot actively collect and respond to the demand, which leads to the passivity of the monitoring work, and the passivity limits the timeliness and effectiveness of the monitoring work. SUMMARY

[0008] The purpose of the present application is to overcome the deficiency that passive monitoring cannot actively collect and respond to the demand of the paralysis patient, and provide a paralysis patient active alarm method based on respiratory signal feature monitoring.

[0009] In order to achieve the above object, the present application adopts the following technical solutions:

[0010] A paralyzed patient active alarm method based on respiratory signal feature monitoring, comprising a PVDF film sensor, a signal acquisition module, a signal processing module and a notification alarm module connected in sequence; comprising the following steps:

[0011] Step 1, the patient lies on the mattress, and the PVDF film sensor is located below the upper body of the human body, the PVDF film sensor detects the analog electric signal of the human body, and the analog electric signal is transmitted to the signal acquisition module;

[0012] Step 2, the signal acquisition module amplifies, samples and converts the analog electric signal to digital electric signal, selects the T-second length signal S[k] in the digital electric signal, N is the signal length of S[k], k=1, 2, 3, …, N;

[0013] S[k] includes body motion signal A[k], ballistocardiogram BCG[k], respiratory signal BR[k] and noise signal NOI[k].

[0014] Step 3, the signal processing module carries out zero mean and normalization processing on S[k], and extracts the respiratory signal BR[k] in S[k] by using a filter;

[0015] Step 4, the signal processing module finds the peak P i and the trough T i in BR[k], wherein NP is the number of peaks and troughs in BR[k] with T-second length, i=1, 2, 3, …, NP; the respiratory signal BR_single[i] of a single cycle from the current expiration to the next expiration is obtained, and BR_single[i] is empty when i=NP;

[0016] Step 5, the signal processing module extracts features from BR_single[i] to obtain a respiratory pattern feature set BR_F[i];

[0017] Step 6, a C-class respiratory feature fingerprint Fin is pre-stored in the respiratory fingerprint library, the signal processing module compares the respiratory feature set BR_F[i] with each Fin in the respiratory fingerprint library; a threshold distinguishing method is adopted, the notification warning event number is set according to the comparison result, and the notification warning number is transmitted to the notification alarm module;

[0018] Step 7, the notification alarm module sends the notification warning information corresponding to the notification warning event number to the patient's guardian or family member.

[0019] The PVDF film sensor of the present application is used to convert signals containing the patient's heart shock signal, body movement signal, respiratory signal and noise signal into analog electric signals, and deliver the analog electric signals to the signal acquisition module;

[0020] The signal acquisition module is used to sample and amplify the analog electric signals, so as to convert them into digital electric signals, and deliver the digital electric signals to the signal processing module;

[0021] The signal processing module is used to process the digital electric signals from the signal acquisition module, convert the processed signals into notification warning event numbers in different states, and deliver the notification warning event numbers to the notification alarm module;

[0022] The notification alarm module sends notification warning information corresponding to the notification warning event numbers to the patient's guardian or family member.

[0023] The signal acquisition module and the signal processing module can be placed on the side of the mattress, the ground, the bottom or other positions which are not easy to be touched.

[0024] After using the present application, when the patient has active needs, the patient can actively send the needs by changing the breathing mode, so as to realize the active perception of the patient's needs, effectively avoid many events, improve the efficiency and quality of the nursing work, reduce the nursing cost, and improve the survival quality of the patient; the possibility of the mis-triggering of the traditional passive method is reduced, and the actual use efficiency is greatly improved.

[0025] The present application encodes different alarm information INF in the respiratory fingerprint library, analyzes the needs of the patient to change the breathing mode spontaneously, greatly simplifies the understanding of the patient's needs by the patient's guardian and family member, and helps the guardian and family member to arrange the nursing time more effectively.

[0026] The portable sensor, signal acquisition processing and alarm module are adopted, the expensive medical monitoring equipment is not needed, the working performance is reliable, and the high performance price ratio is provided for the rehabilitation patients at home or in the nursing home.

[0027] As preferred, step 5 comprises the following specific steps:

[0028] The signal processing module calculates the respiratory frequency BR_R[t] by using the following formula: BR_R[t] = (NP-1) / 10, NP>1;

[0029] The signal processing module calculates the average respiratory energy BR_E[t] by using the following formula:

[0030]

[0031] The signal processing module calculates the average respiratory rhythm BR_Y[t] by using the following formula:

[0032]

[0033] The signal processing module calculates the average inhalation time BR_IN[t] using the following formula:

[0034]

[0035] The signal processing module calculates the average exhalation time BR_OUT[t] using the following formula:

[0036]

[0037] The signal processing module calculates the average respiratory flow rate BR_S[t] using the following formula:

[0038]

[0039] The respiratory pattern feature set BR_F[t] at time t is obtained:

[0040] BR_F[t] = {BR_R[t], BR_E[t], BR_Y[t], BR_IN[t], BR_OUT[t], BR_S[t]}.

[0041] As a preference, step 6 includes the following steps:

[0042] Step 6-1, matrix dimension is increased for BR_F[t] using the following formula to obtain the high-dimensional feature set BR_FX[t]:

[0043] Let BR_F[t] = {x1, x2, x3, x4, x5, x6};

[0044] Then

[0045] where k(x j ,x z ) is a kernel function, and the values of x j ,x z are x1, x2, x3, x4, x5, or x6.

[0046] Step 6-2, the spearman similarity of BR_FX[t] and the high-dimensional features Fin_X of each type of Fin is calculated using the following formula to obtain C similarities D(t): D(t) = Max(spearman_cor(BR_FX[t], Fin_X))

[0047] Step 6-3, when D(t) < TH, TH ∈ (0.5, 0.8), it indicates that the patient at time t is in a normal respiratory state, and the warning event number is set to 0.

[0048] Step 6-4, when D(t) > TH, calculate the maximum value D(t) in C similarity D(t) max , select the class Fin corresponding to D(t) max Select the notification warning event number corresponding to the class Fin.

[0049] As a preferred, the kernel function is Gaussian kernel function:

[0050] σ>0 is the bandwidth of Gaussian kernel function, and the bandwidth is selected in the range of 0.3-0.7.

[0051] As a preferred, step 3 comprises the following steps:

[0052] The zero-mean signal S rm [k] is calculated by the following formula:

[0053]

[0054] The normalized signal S nm [k] is calculated by the following formula:

[0055]

[0056] Wherein, std() is the standard deviation calculation operator.

[0057] As a preferred, the step of recording the breath pattern feature set BR_F[t] into the breath fingerprint library comprises:

[0058] A plurality of breath pattern feature sets BR_F[t] are recorded into the breath fingerprint library, each breath pattern feature set BR_F[t] is marked as a breath fingerprint Fin and corresponds to a notification warning event number, and the value of the notification warning event number is 0, 1, 2, 3, …, NF; NF is an integer greater than 3, representing the maximum value of the notification warning event number.

[0059] As a preferred, the PVDF film sensor is located below the heart of the patient, and the distance between the PVDF film sensor and the bed head is 50-60 cm.

[0060] Therefore, the present application has the following beneficial effects: different demand information transmitted by the patient can be detected according to different breath patterns of the patient, active perception of the patient's demand is realized, many events are effectively avoided, the efficiency and quality of the nursing work are effectively improved, the nursing cost is reduced, and the monitoring cost is reduced. BRIEF DESCRIPTION OF DRAWINGS

[0061] Figure 1 is a flow chart of the present application;

[0062] Figure 2 is a schematic diagram of the present application;

[0063] Figure 3 is a schematic diagram of the signal breath fingerprint library of the present application. DETAILED DESCRIPTION

[0064] The present application is further described below in conjunction with the accompanying drawings and specific embodiments.

[0065] As Figures 1-3 shown in the embodiment is a kind of paralysis patient active alarm method based on respiratory signal feature monitoring, including the PVDF film sensor, signal acquisition module, signal processing module and notification alarm module connected in turn;Including the following steps:

[0066] Step 1, the patient lies on the mattress, so that the PVDF film sensor is located below the upper body of the human body, the PVDF film sensor detects the analog electric signal of the human body, and the analog electric signal is transmitted to the signal acquisition module;

[0067] Step 2, the signal acquisition module amplifies, samples and converts the analog electric signal to digital electric signal, selects the 10-second length signal S [k] in the digital electric signal;S [k] includes: body movement signal A [k], ballistocardiogram BCG [k], respiratory signal BR [k] and noise signal NOI [k], N is the signal length of S [k], k=1, 2, 3, …, N;N=1250 is the length of 10 seconds signal;

[0068] Step 3, the signal processing module carries out zero mean and normalization processing to S [k], and extracts the respiratory signal BR [k] in S [k] by using filter;

[0069] The zero-mean signal S rm [k] is calculated by using the following formula:

[0070]

[0071] The normalized signal S nm [k] is calculated by using the following formula:

[0072]

[0073] Wherein, std () is the standard deviation calculation operator.

[0074] S nm [k] is filtered by using fourth-order Butterworth filter 0.1Hz-3Hz, and the signal data of non-respiratory signal band of S nm [k] is filtered out;

[0075] Step 4, the signal processing module finds the peaks P in BR[k] i and the troughs T i where NP is the number of peaks and troughs in BR[k] of 10 seconds length, i = 1, 2, 3, …, NP; obtain the single cycle of breath signal BR_single[i] from the current expiration start to the next expiration start, BR_single[i] is empty when i = NP;

[0076] Step 5, the signal processing module extracts features from BR_single[i] to obtain the breath pattern feature set BR_F[i]:

[0077] The signal processing module calculates the breath frequency BR_R[t] using the following formula: BR_R[t] = (NP-1) / 10, NP>1;

[0078] The signal processing module calculates the average breath energy BR_E[t] using the following formula:

[0079]

[0080] The signal processing module calculates the average breath rhythm BR_Y[t] using the following formula:

[0081]

[0082] The signal processing module calculates the average inspiration time BR_IN[t] using the following formula:

[0083]

[0084] The signal processing module calculates the average expiration time BR_OUT[t] using the following formula:

[0085]

[0086] The signal processing module calculates the average breath flow rate BR_S[t] using the following formula:

[0087]

[0088] Obtain the breath pattern feature set BR_F[t] at time t:

[0089] BR_F[t] = {BR_R[t], BR_E[t], BR_Y[t], BR_IN[t], BR_OUT[t], BR_S[t]}.

[0090] Step 6, the 6 types of breath feature fingerprints Fin are pre-stored in the breath fingerprint library, and the signal processing module compares the breath feature set BR_F[i] with each type of Fin in the breath fingerprint library; threshold division method is adopted, the notification warning event number is set according to the comparison result, and the notification warning number is transmitted to the notification alarm module:

[0091] Before real-time detection, the breath feature set BR_F[i] corresponding to the 6 breath patterns of the patient with paralysis is obtained by using steps 1-6, each breath pattern feature set BR_F[t] is recorded into the breath fingerprint library, the breath pattern feature set BR_F[t] is labeled as a breath fingerprint Fin and corresponds to a notification warning event number, and the 6 notification warning event numbers are 0, 1, 2, 3, 4 and 5.

[0092] Step 6-1, the matrix dimension is increased for BR_F[t] by using the following formula to obtain a high-dimensional feature set BR_FX[t]:

[0093] BR_F[t] = {x1, x2, x3, x4, x5, x6};

[0094]

[0095] Wherein, k(x j ,x z ) is a kernel function, and the values of x j ,x z are x1, x2, x3, x4, x5 or x6; the kernel function in the embodiment is a Gaussian kernel function:

[0096] σ>0 is the bandwidth of the Gaussian kernel function, and the bandwidth selection range is 0.3-0.7.

[0097] Step 6-2, the spearman similarity of BR_FX[t] and the high-dimensional features Fin_X of each type of Fin is calculated by using the following formula to obtain C similarities D(t): D(t) = Max(spearman_cor(BR_FX[t], Fin_X))

[0098] Step 6-3, when D(t) < TH, TH ∈ (0.5, 0.8), it indicates that the patient at time t is in a normal breathing state, and the notification warning event number is set to 0;

[0099] Step 6-4, when D(t) > TH, the maximum value D(t) max in the C similarities D(t) is calculated, the type of Fin corresponding to D(t) max is selected, and the notification warning event number corresponding to the type of Fin is selected. ​

[0100] When the notification warning event number = 0, it means that the patient is currently in a normal breathing mode, at which time the notification alarm module will not start.

[0101] When the notification warning event number > 0, it means that the patient is currently not in a normal breathing mode, and the notification alarm module notifies the patient's guardian or family member in the form of short message, app notification, alarm sound, alarm phone, etc.

[0102] Step 7, the notification alarm module sends the notification warning information corresponding to the notification warning event number to the patient's guardian or family member.

[0103] As shown in Figure 2 , the PVDF film sensor is located on the mattress and below the patient's heart, and the PVDF film sensor is 50 cm away from the bed head.

[0104] As shown in Figure 3 , the different notification warning signals corresponding to different breathing states recorded in the respiratory fingerprint library of the paralyzed patient of the embodiment. Figure 3 recorded 0s-10s, 10s-20s, 20s-30s, 30s-40s, 40s-50s, 50s-60s 6 different breathing modes, wherein 0-10s is the normal breathing mode of the patient, 10s-20s is the periodic breathing mode of the patient with deep and shallow breathing alternately, 20s-30s is the slow breathing mode of the patient, 30s-40s is the shallow breathing mode of the patient, 40s-50s is the deep breathing mode of the patient, and 50s-60s is the rapid breathing mode of the patient. Corresponding to the notification warning event number with the number 0 to 5. The patient's guardian or family member will respond appropriately according to the warning information corresponding to the different notification warning event numbers received, in order to assist the patient to perform corresponding actions to meet their needs.

[0105] May include providing additional support, adjusting environmental conditions or starting necessary medical procedures, etc. By responding to the warning information in a timely and effective manner, the guardian can better meet the patient's life and medical needs and ensure that they receive appropriate support.

[0106] The above only describes the preferred embodiments of the present application and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for monitoring a paralytic patient to actively alert, based on a breathing signal feature, characterized in that, It comprises a PVDF film sensor, a signal acquisition module, a signal processing module and a notification alarm module connected in sequence; and comprises the following steps. Step 1, the patient lies on the mattress, and the PVDF film sensor is located below the upper body of the human body, the PVDF film sensor detects the analog electric signal of the human body, and the analog electric signal is transmitted to the signal acquisition module; Step 2, the signal acquisition module amplifies, samples and converts the analog electric signal to a digital electric signal, selects a signal S[k] of T seconds in length in the digital electric signal, N is the signal length of S[k], k = 1, 2, 3, …, N; Step 3, the signal processing module performs zero mean and normalization processing on S[k], and extracts a respiratory signal BR[k] in S[k] by using a filter; Step 4, the signal processing module finds the peaks P in BR[k] i and the troughs T i where NP is the number of peaks and troughs in BR[k] of T seconds in length, i = 1, 2, 3, …, NP; the breath signal BR_single[i] of a single cycle from the current breath start to the next breath start is obtained, BR_single[i] is empty when i = NP; Step 5, the signal processing module extracts features from BR_single[i] to obtain a respiratory pattern feature set BR_F[i]; Step 6, a respiratory fingerprint library pre-stores C types of respiratory feature fingerprints Fin, the signal processing module compares the respiratory feature set BR_F[i] with each type of Fin in the respiratory fingerprint library; a threshold distinguishing method is adopted, a notification warning event number is set according to the comparison result, and the notification warning number is transmitted to the notification alarm module; Step 7, the notification alarm module sends notification warning information corresponding to the notification warning event number to the patient's guardian or family member.

2. The method of claim 1, wherein the method further comprises: Step 5 comprises the following specific steps: The signal processing module calculates the respiratory frequency BR_R[t] by using the following formula: BR_R[t] = (NP-1) / 10, NP>1; The signal processing module calculates the average respiratory energy BR_E[t] by using the following formula: The signal processing module calculates the average respiratory rhythm BR_Y[t] by using the following formula: The signal processing module calculates the average inspiration time BR_IN[t] by using the following formula: The signal processing module calculates the average expiration time BR_OUT[t] by using the following formula: The signal processing module calculates the average respiratory flow rate BR_S[t] by using the following formula: The respiratory pattern feature set BR_F[t] at time t is obtained: BR_F[t] = {BR_R[t], BR_E[t], BR_Y[t], BR_IN[t], BR_OUT[t], BR_S[t]}.

3. The active alarm method for paralysis patients based on respiratory signal feature monitoring according to claim 2; wherein: Step 6 comprises the following steps: Step 6-1, the matrix dimension of BR_F[t] is increased by using the following formula to obtain a high-dimensional feature set BR_FX[t]: BR_F[t] = {x1, x2, x3, x4, x5, x6}; then where k(x j ,x z ) is a kernel function, and x j ,x z each take the value of x1, x2, x3, x4, x5, or x6. Step 6-2, the high-dimensional feature Fin_X of BR_FX[t] and each type of Fin is calculated by using the following formula to obtain C similarities D(t): D(t) = Max(spearman_cor(BR_FX[t], Fin_X)) Step 6-3, when D(t) < TH, TH ∈ (0.5, 0.8), it indicates that the patient at time t is in a normal respiratory state, and the notification warning event number is set to 0; Step 6-4, when D(t) > TH, calculate the maximum value D(t) in C similarities D(t) max , select a class Fin corresponding to D(t) max , select the notification warning event number corresponding to the class Fin.

4. The method of claim 3, wherein the method further comprises: The kernel function is a Gaussian kernel function: σ>0 is the bandwidth of the Gaussian kernel function, with a bandwidth selection range of 0.3-0.

7.

5. The method of claim 1, wherein the method further comprises: Step 3 comprises the following steps: The zero-meaned signal S is calculated using the following equation rm [k]: The normalized signal S is calculated using the following equation nm [k]: Wherein std() is a standard deviation calculation operator.

6. The method of claim 1, wherein the method further comprises: The step of entering the breath pattern feature set BR_F[t] into the breath fingerprint library: A plurality of breath pattern feature sets BR_F[t] are entered into the breath fingerprint library, each breath pattern feature set BR_F[t] is marked as a breath fingerprint Fin and corresponds to a notification warning event number, the value of the notification warning event number is 0, 1, 2, 3, …, NF; NF is an integer greater than 3, representing the maximum value of the notification warning event number.

7. The paralysis patient active alarm method based on respiratory signal feature monitoring according to claim 1 or 2 or 3 or 4 or 5 or 6; characterized in that, The PVDF film sensor is located below the patient's heart, and the distance between the PVDF film sensor and the bed head is 50-60 cm.

Images