Portable respiration monitoring method and device based on piezoelectric film sensor

Abstract

The invention discloses a portable respiration monitoring method and device based on a piezoelectric film sensor, and relates to the technical field of medical health monitoring. The portable respiration monitoring method comprises the following steps that the piezoelectric film sensor made of a high-molecular polymer sensitive material is attached to the outer side of a nasal cavity; breathing airflow vibration enables lattice polarity displacement in the sensor to generate surface charges, and electric signals are output; transmitting the electric signal to a charge amplifier circuit through a shielded wire, and sequentially carrying out band-pass filtering and low-pass filtering; the filtered signal is converted into a digital signal, and a special buffer area is developed to realize non-blocking transmission; performing waveform smoothing on the converted and transmitted digital signal; based on the smoothed waveform, de-averaging the signal and detecting a zero crossing point interval sequence, and calculating a complexity parameter; adopting a baseline threshold discrimination method to accumulate the inspiration and expiration time, and calculating to obtain the respiratory rate; and dynamically displaying the obtained respiratory rate and related waveform parameters and transmitting the respiratory rate and the related waveform parameters to an upper computer through Bluetooth.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of medical health monitoring, in particular to a portable respiration monitoring method and device based on a piezoelectric film sensor. BACKGROUND

[0002] Respiration rate as a key physiological indicator of the human body, its real-time and accurate monitoring has important clinical value for early screening, disease assessment and prognosis of sleep apnea syndrome and other respiratory diseases. The current mainstream respiration monitoring technology has problems to be solved:

[0003] The traditional contact sensor has insufficient sensitivity when collecting the weak airflow signal in the nasal cavity, and the rigid contact can easily cause discomfort to the user, resulting in poor monitoring compliance, especially difficult to meet the long-term night sleep monitoring demand, the respiration signal itself has a narrow characteristic frequency range, and is easily affected by the superposition of baseline drift, 50Hz power frequency interference and motion pseudo-noise, so that the signal-to-noise ratio of the original signal is extremely low, which brings difficulties to the subsequent processing.

[0004] In the signal processing link, the existing equipment generally lacks a targeted filtering mechanism, and the traditional RC filter circuit is difficult to effectively suppress wideband interference; in the feature extraction algorithm, the peak detection method is easily affected by the pseudo-peak interference caused by poor respiration, resulting in large frequency calculation error, therefore we propose a portable respiration monitoring method and device based on a piezoelectric film sensor. SUMMARY

[0005] To solve the above technical problems, the present application is realized by the following technical scheme:

[0006] The present application is a portable respiration monitoring method and device based on a piezoelectric film sensor, comprising the following steps:

[0007] Step S1: a piezoelectric film sensor with a high polymer sensitive material is attached to the outside of the nasal cavity, ensuring that the electrode plate is perpendicular to the airflow direction; the respiration airflow vibration causes the internal lattice polarity displacement of the sensor to generate surface charge, and the electric signal is output through the potential difference between the upper and lower electrode layers;

[0008] Step S2: transmit the electric signal to the charge amplifier circuit through the shielding line, and sequentially perform 0.05 to 3Hz band-pass filtering and 5Hz low-pass filtering;

[0009] Step S3: convert the filtered signal into a digital signal using an ADC with a sampling rate of not less than 100Hz, configure a DMA cyclic transmission mode to open a dedicated buffer area to realize non-blocking transmission;

[0010] Step S4: using Kalman filter to smooth the waveform of the converted and transmitted digital signal, updating the estimation error covariance through the prediction stage and calculating the Kalman gain through the update stage to realize the optimal estimation;

[0011] Step S5: based on the smoothed waveform, the signal is de-meaned by zero point analysis method and the zero-crossing interval sequence is detected, and the complexity parameter is calculated;

[0012] Step S6: setting the baseline voltage, using the baseline threshold discrimination method to accumulate the inhalation and exhalation time according to the parameters obtained by feature extraction, and the timer interrupt frequency is 100Hz, and the frequency is calculated when the respiratory cycle is completed;

[0013] Step S7: the calculated respiratory frequency and related waveform parameters are dynamically displayed by hardware IIC driving OLED, and are packaged and transmitted to the upper computer through Bluetooth protocol.

[0014] Further, the step S1 comprises the following steps:

[0015] Step S11, the sensor output potential difference is proportional to the input charge, satisfying the charge-voltage conversion relationship: V out ∝ΔQ;

[0016] Wherein, V out is the potential difference output by the sensor, ∝ is the proportional relationship, and ΔQ is the surface charge change amount generated by the sensor due to the vibration of the respiratory airflow;

[0017] Step S12, the nasal airflow direction is perpendicular to the electrode plate to improve the sensitivity (exhalation / inhalation airflow model as shown in Figure 3 ).

[0018] Further, the step S2 comprises the following steps:

[0019] Step S21, the charge amplifier circuit realizes impedance matching, which is used to analyze the feedback impedance characteristics of the charge amplifier to realize impedance matching, and the calculation formula of the feedback impedance is as follows:

[0020]

[0021] In the formula, Z f is the feedback impedance of the charge amplifier, R f is the feedback resistance, X cf is the capacitive reactance of the feedback capacitor, and j is the imaginary unit;

[0022] Step S22, when R f is large enough, it is simplified as: Z f ≈-jX cf ;

[0023] Step S23, when analyzing the input impedance of the charge amplifier, the expression of the input impedance can be derived as follows:

[0024] Z i = R i -jX ci ;

[0025] In the formula, Z i is the input impedance of the charge amplifier, R i is the input resistance, and X ci is the capacitive reactance of the input capacitance.

[0026] Step S24, to calculate the voltage gain of the charge amplifier to evaluate the amplification capability of the circuit to the signal, the approximate calculation formula of the voltage gain is as follows:

[0027]

[0028] In the formula, A is the voltage gain of the charge amplifier.

[0029] Step S25, determine the high-pass cutoff frequency of the band-pass filter to filter out the signals of the required frequency band, the parameters of the band-pass filter are determined by the following, and the high-pass cutoff frequency formula is as follows:

[0030]

[0031] In the formula, f H is the high-pass cutoff frequency of the band-pass filter, R i is the input resistance, and C i is the input capacitance.

[0032] Step S26, determine the low-pass cutoff frequency of the band-pass filter to limit the high-frequency interference signal, and the low-pass cutoff frequency calculation formula is as follows:

[0033]

[0034] In the formula, f L is the low-pass cutoff frequency of the band-pass filter, R f is the feedback resistance, and C f is the feedback capacitance.

[0035] Step S27, calculate the cutoff frequency of the RC passive low-pass filter circuit to clarify its suppression capability to high-frequency signals, and the calculation formula of the cutoff frequency of the two-stage RC passive low-pass filter is as follows:

[0036]

[0037] In the formula, f0 is the cutoff frequency of the RC passive low-pass filter, R lp is the low-pass filter resistance, and C lpLow pass filter capacitor.

[0038] Further, the step S3 comprises the following steps:

[0039] Step S31, the ADC is configured to 12-bit resolution, sampling rate ≥ 100Hz;

[0040] Step S32, DMA transmission makes CPU work in parallel with peripherals.

[0041] Further, the step S4 comprises the following steps:

[0042] Step S41, Kalman filtering is performed in two stages:

[0043] Prediction stage: update the estimation error covariance, the formula is as follows:

[0044] In the formula, is the prior estimation error covariance at time k, A is the state transition matrix, P k-1 is the posterior estimation error covariance at time k-1, A T is the transpose of matrix A, Q is the process noise covariance;

[0045] Update stage: calculate the Kalman gain, the formula is as follows: and update the state estimation;

[0046] In the formula, K k is the Kalman gain at time k, H is the observation matrix, H T is the transpose of matrix H, R is the measurement noise covariance;

[0047] Step S42, dynamically adjust the process noise covariance Q and the measurement noise covariance R to suppress the baseline drift.

[0048] Further, the step S5 comprises the following steps:

[0049] Step S51, the original signal is de-meaned to eliminate the influence of direct current component, which is convenient for subsequent zero-crossing point analysis, and the de-meaning formula is as follows:

[0050]

[0051] In the formula, y(n) is the signal after de-meaning, x(n) is the original respiratory signal, N is the total number of signal sequence, is the summation operation of the original signal from the first point to the Nth point, is the mean of the original signal;

[0052] Step S52, the zero-crossing point spacing sequence d1, d2, …, d M ;

[0053] Step S53, measure the complexity of the signal, by calculating the mean value of the zero-crossing interval sequence as a characteristic parameter, the calculation formula of the mean value mZCI is as follows:

[0054]

[0055] In the formula, mZCI is the mean value of the zero-crossing interval sequence, M is the total number of zero-crossing intervals, d(m) is the mth zero-crossing interval, is the summation operation of all zero-crossing intervals;

[0056] Step S54, further quantifying the complexity of the signal, calculating the normalized standard deviation of the zero-crossing interval sequence, the calculation formula of nsZCI is as follows:

[0057]

[0058] In the formula, nsZCI is the normalized standard deviation of the zero-crossing interval sequence, [d(m)-mZCI] 2 is the square of the difference between the mth zero-crossing interval and the mean value, is the summation of the squares of the differences between all intervals and the mean value;

[0059] Step S55, when the respiratory signal mZCI value is large, the baseline threshold method is used to resist false peak interference.

[0060] Further, the step S6 comprises the following steps:

[0061] Step S61, set the baseline voltage to 1.65V;

[0062] Step S62, define integer variables T1: inspiration time; T2: expiration time;

[0063] Step S63, the timer interrupt frequency is 100Hz(period 10ms), which meets the accuracy requirement of 3s breathing cycle;

[0064] Step S64, the flag assignment rule is:

[0065] Waveform>baseline→Sign1=1 and T1←T1+1;

[0066] Waveform<baseline→Sign2=1 and T2←T2+1;

[0067] Step S65: breathing cycle judgment condition: Sign1=1 and Sign2=1 and waveform crosses baseline;

[0068] Step S66, trigger the execution of zero clearing: T1=0, T2=0, Sign1=0, Sign2=0;

[0069] Step S67, calculate the respiratory frequency according to the recorded respiratory cycle time, get the respiratory frequency per minute, the respiratory frequency calculation formula is as follows:

[0070]

[0071] In the formula, T1+T2 is the total time of a complete respiratory cycle, and 60 is the conversion of cycle time to respiratory frequency per minute.

[0072] Further, the step S7 comprises the following steps:

[0073] Step S71, the OLED realizes waveform dynamic refreshing through a hardware IIC interface;

[0074] Step S72, the data encapsulation format of the Bluetooth transmission contains a timestamp, a respiratory frequency value and a device ID;

[0075] Step S73, the display interface of the upper computer is used for subsequent processing by doctors.

[0076] The present application has the following beneficial effects:

[0077] 1. The piezoelectric film sensor of the present application adopts a high-molecular polymer sensitive material, the electrode plate is perpendicular to the direction of nasal airflow, thereby improving the sensitivity, solving the problem of insufficient sensitivity of the traditional contact sensor, and the method of attaching to the outside of the nasal cavity is more comfortable, reduces the discomfort caused by rigid contact, significantly improves the compliance of long-term night sleep monitoring, and meets the continuous monitoring demand.

[0078] 2. The waveform is smoothed by 0.05-3Hz band-pass filtering, 5Hz low-pass filtering and Kalman filtering, the baseline drift is suppressed by dynamically adjusting the process noise and measurement noise covariance, the baseline drift, 50Hz power frequency interference and motion pseudo-noise are effectively filtered, the signal-to-noise ratio of the respiratory signal is greatly improved, and the problem of difficult signal processing caused by insufficient filtering mechanism of the existing device is solved.

[0079] 3. The present application adopts zero-point analysis method to de-mean and detect the zero-crossing interval, combines with the baseline threshold discrimination method to accumulate the inhalation and exhalation time, the frequency of the timer interrupt is 100Hz to ensure the cycle calculation accuracy, reduce the pseudo-peak interference, improve the respiratory frequency calculation accuracy, and solve the problem of large frequency calculation error caused by the pseudo-peak interference caused by the traditional peak detection method.

[0080] Of course, any product implementing the present application does not necessarily need to achieve all the advantages described above at the same time. BRIEF DESCRIPTION OF DRAWINGS

[0081] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without any creative effort on the basis of these drawings.

[0082] Figure 1 Flowchart of the portable breath monitoring method and device based on piezoelectric film sensor of the present application;

[0083] Figure 2 System block diagram for monitoring breath frequency;

[0084] Figure 3 Nasal cavity breath airflow model diagram (left: expiration, right: inspiration);

[0085] Figure 4 Zero-crossing point analysis diagram;

[0086] Figure 5 Respiratory frequency calculation flowchart;

[0087] Figure 6 Respiratory signal acquisition module block diagram;

[0088] Figure 7 Charge amplifier schematic diagram;

[0089] Figure 8 Charge amplifier frequency response diagram;

[0090] Figure 9 Low-pass filter frequency response diagram;

[0091] Figure 10 ADC and DMA related parameter configuration diagram;

[0092] Figure 11 Main program flowchart;

[0093] Figure 12 STM32F103C8T6 chip configuration diagram;

[0094] Figure 13 Circuit diagram of IP5306-CK charge and discharge module;

[0095] Figure 14 Circuit board 3D rendering diagram (left: top layer, right: top layer);

[0096] Figure 15 System function display diagram;

[0097] Figure 16 Bluetooth APP data recording and display diagram. DETAILED DESCRIPTION

[0098] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the protection scope of the present application.

[0099] Please refer to Figure 1 The present application is a portable breath monitoring method and device based on a piezoelectric film sensor, which comprises the following steps:

[0100] Step S1: A piezoelectric film sensor using a high polymer sensitive material is attached to the outside of the nasal cavity, ensuring that the electrode plate is perpendicular to the airflow direction. The vibration of the respiratory airflow causes the internal lattice polarity of the sensor to displace and generate surface charges, and the electric potential difference between the upper and lower electrode layers outputs an electric signal.

[0101] Step S2: The electric signal is transmitted to the charge amplifier circuit through the shielding line, and 0.05 to 3 Hz band-pass filtering and 5 Hz low-pass filtering are sequentially performed.

[0102] Step S3: The filtered signal is converted into a digital signal using the built-in 12-bit ADC of STM32F103C8T6 at a sampling rate of not less than 100 Hz, and a dedicated buffer area is opened to realize non-blocking transmission in the DMA cyclic transmission mode.

[0103] Step S4: The transmitted digital signal after conversion is smoothed using a Kalman filter, and the optimal estimation is realized by updating the estimation error covariance in the prediction stage and calculating the Kalman gain in the update stage.

[0104] Step S5: Based on the smoothed waveform, the signal is de-meaned and the zero-crossing interval sequence is detected by the zero-point analysis method, and the complexity parameter is calculated.

[0105] Step S6: Set a baseline voltage of 1.65V, and use the baseline threshold discrimination method to accumulate the inhalation and exhalation time according to the parameters obtained by feature extraction. The timer interrupt frequency is 100Hz, and the frequency is calculated when the respiratory cycle is completed.

[0106] Step S7: The calculated respiratory frequency and related waveform parameters are dynamically displayed by hardware IIC driving OLED, and are packaged and transmitted to the host computer through Bluetooth protocol.

[0107] Further, the step S1 comprises the following steps:

[0108] Step S11, the sensor output electric potential difference is proportional to the input charge, satisfying the charge-voltage conversion relationship: Vout a AQ;

[0109] wherein, V out is the potential difference outputted by the sensor, a is the proportional relationship, AQ is the amount of surface charge change generated by the sensor due to the vibration of the respiratory airflow;

[0110] Step S12, the direction of the nasal airflow is perpendicular to the electrode plate to improve the sensitivity (exhalation / inspiration airflow model as shown in Figure 3 ).

[0111] Further, the step S2 includes the following steps:

[0112] Step S21, the charge amplifier circuit realizes impedance matching, which is to analyze the feedback impedance characteristics of the charge amplifier to realize impedance matching, and the calculation formula of the feedback impedance is as follows:

[0113]

[0114] In the formula, Z f is the feedback impedance of the charge amplifier, R f is the feedback resistance, X cf is the capacitive reactance of the feedback capacitor, and j is the imaginary unit;

[0115] Step S22, when analyzing the input impedance of the charge amplifier, the expression of the input impedance can be derived as follows:

[0116] Z i = R i -jX ci ;

[0117] Step S23, when R f is large enough, it is simplified as: Z f ≈-jX cf ;

[0118] In the formula, Z i is the input impedance of the charge amplifier, R i is the input resistance, and X ci is the capacitive reactance of the input capacitor;

[0119] Step S24, to calculate the voltage gain of the charge amplifier to evaluate the amplification ability of the circuit to the signal, the approximate calculation formula of the voltage gain is as follows:

[0120]

[0121] In the formula, A is the voltage gain of the charge amplifier;

[0122] Step S25, determine the high-pass cutoff frequency of the band-pass filter to filter out the signal of the desired frequency band, the band-pass filter parameters are determined as follows, the high-pass cutoff frequency formula is as follows:

[0123]

[0124] In the formula, f H The high-pass cutoff frequency of the band-pass filter, R i is the input resistance, C i is the input capacitance.

[0125] Step S26, determine the low-pass cutoff frequency of the band-pass filter to limit the high-frequency interference signal, the low-pass cutoff frequency calculation formula is as follows:

[0126]

[0127] In the formula, f L is the low-pass cutoff frequency of the band-pass filter, R f is the feedback resistance, C f is the feedback capacitance.

[0128] Step S27, calculate the cutoff frequency of the RC passive low-pass filter circuit to determine its suppression ability to high-frequency signals, the two-stage RC passive low-pass filter cutoff frequency calculation formula is as follows:

[0129]

[0130] In the formula, f0 is the cutoff frequency of the RC passive low-pass filter, is the low-pass filter resistance, is the low-pass filter capacitance.

[0131] Further, the step S3 includes the following steps:

[0132] Step S31, the ADC is configured to have a 12-bit resolution and a sampling rate ≥100Hz;

[0133] Step S32, DMA transmission makes the CPU work in parallel with the peripherals.

[0134] Further, the step S4 includes the following steps:

[0135] Step S41, Kalman filtering is performed in two stages:

[0136] Prediction stage: update the estimation error covariance, the formula is as follows:

[0137] In the formula, is the prior estimation error covariance at time k, A is the state transition matrix, P k-1is the posterior estimation error covariance of time k-1, A T is the transpose of matrix A, Q is the process noise covariance;

[0138] Update phase: calculate the Kalman gain, the formula is as follows: and update the state estimation;

[0139] In the formula, K k is the Kalman gain of time k, H is the observation matrix, H T is the transpose of matrix H, R is the measurement noise covariance;

[0140] Step S42, dynamically adjust the process noise covariance Q and the measurement noise covariance R to suppress the baseline drift.

[0141] Further, the step S5 comprises the following steps:

[0142] Step S51, the original signal is de-meaning processing, eliminate the influence of direct current component, facilitate subsequent zero crossing analysis, de-meaning formula is as follows:

[0143]

[0144] In the formula, y(n) is the signal after de-meaning processing, x(n) is the original respiratory signal, N is the total number of signal sequence, is the summation operation of the original signal from the first point to the Nth point, is the mean of the original signal;

[0145] Step S52, the zero crossing interval sequence d1, d2, …, d M is extracted (such as Figure 4 );

[0146] Step S53, measure the complexity of the signal, calculate the mean of the zero crossing interval sequence as a characteristic parameter, the calculation formula of the mean mZCI is as follows:

[0147]

[0148] In the formula, mZCI is the mean of the zero crossing interval sequence, M is the total number of zero crossing intervals, d(m) is the mth zero crossing interval, is the summation operation of all zero crossing intervals;

[0149] Step S54, further quantize the complex characteristics of the signal, calculate the normalized standard deviation of the zero crossing interval sequence, the calculation formula of nsZCI is as follows:

[0150]

[0151] Wherein, nsZCI is the normalized standard deviation of the zero-crossing interval sequence, [d(m)-mZCI] 2 is the square of the difference between the mth zero-crossing interval and the mean value, is the sum of the squares of the differences between all intervals and the mean value;

[0152] Step S55, when the respiratory signal mZCI value is large, the baseline threshold method is used to resist false peak interference.

[0153] Further, the step S6 comprises the following steps:

[0154] Step S61, the baseline voltage is set to 1.65V;

[0155] Step S62, define integer variables T1: inhalation time; T2: exhalation time;

[0156] Step S63, the timer interrupt frequency is 100Hz (period 10ms), which meets the accuracy requirement of 3s breathing cycle;

[0157] Step S64, the flag assignment rule is:

[0158] Waveform>baseline→Sign1=1 and T1←T1+1;

[0159] Waveform<baseline→Sign2=1 and T2←T2+1;

[0160] Step S65: breathing cycle judgment condition ( Figure 5 ): Sign1=1 and Sign2=1 and waveform crosses baseline;

[0161] Step S66, after triggering, execute zero clearing: T1=0, T2=0, Sign1=0, Sign2=0;

[0162] Step S67, calculate the respiratory frequency according to the recorded breathing cycle time, get the respiratory rate per minute, and the respiratory frequency calculation formula is as follows:

[0163]

[0164] Wherein, T1+T2 is the total time of a complete breathing cycle, and 60 is the conversion of cycle time to respiratory rate per minute.

[0165] Further, the step S7 comprises the following steps:

[0166] Step S71, the OLED realizes dynamic waveform refresh through hardware IIC interface;

[0167] Step S72, the Bluetooth transmission data packaging format contains timestamp, respiratory frequency value, and device ID;

[0168] Step S73, the host computer displays an interface for the doctor to perform subsequent processing.

[0169] When the breathing is normal (15-30 times per minute), as shown in (a), the waveform change and the breathing state can be displayed, at this time, the breathing state flag bit is "1", indicating that the breathing is normal; when the breathing is faster (more than 30 times per minute), as shown in (b), at this time, the breathing state flag bit is "2", indicating that the breathing is too fast; when the breathing is slower (less than 15 times per minute), as shown in (c), at this time, the breathing state flag bit is "0", indicating that the breathing is slower. Figure 15 Figure 15 Figure 15

[0170] At the same time, the data can also be uploaded to the cloud in real time by using the Bluetooth module, realizing data monitoring to respond to early warning information in time, as shown in (d). Figure 16

[0171] One specific application of the embodiment is:

[0172] Step one,

[0173] 1. A polyvinylidene fluoride (PVDF) piezoelectric film sensor (thickness 50 μm) is attached to the outside of the user's nasal cavity to ensure that the electrode plate is perpendicular to the airflow direction. When the respiratory airflow passes through, the sensor generates a surface charge change amount ΔQ, and the output potential difference V out ∝ ΔQ, with a sensitivity of ±0.5 Pa -1 ;

[0174] 2. Charge amplifier: feedback resistance R f = 1 GΩ, feedback capacitance C f = 3.3 μF, and the feedback impedance is calculated as Z f ≈ -jX cf , realizing high impedance matching;

[0175] Band-pass filter: high-pass cutoff frequency f H = 3 Hz; low-pass cutoff frequency f L = 0.05 Hz;

[0176] Second-order RC low-pass filter: cutoff frequency f0= 5 Hz;

[0177] STM32F103C8T6 microcontroller, configured with 12-bit ADC (sampling rate 100 Hz), DMA circular buffer (size 256 bytes) to realize non-blocking transmission;

[0178] 0.96 inch OLED (SSD1306 driver, hardware IIC interface), HC-05 Bluetooth module (transmission protocol: UUID = 0xFFE1); ​​​​

[0179] Step two,

[0180] 1. Sensor output signal is transmitted to the charge amplifier through the shielded line, and after double-stage filtering, it is converted into digital signal by ADC. DMA stores data into buffer at a rate of 100Hz

[0181] 2. Initialization parameters: state transition matrix A = 1, observation matrix H = 1, process noise covariance Q = 0.01, measurement noise covariance R = 0.1;

[0182] Prediction stage: prior error covariance:

[0183] Update stage: Kalman gain: State estimation:

[0184] Dynamic adjustment: when baseline drift > 0.1V, adaptively reduce Q to 0.001 to suppress drift;

[0185] 3. De-meaning: calculate mean value for 1024-point data window Output y(n) = x(n) - μ α ;

[0186] Zero-crossing detection: identify zero-crossing sequence of y(n), calculate:

[0187] Mean

[0188] Normalized standard deviation:

[0189] 4. Baseline threshold method: baseline voltage V base = 1.65V, timer interrupt period 10ms (100Hz);

[0190] Flag bit rule:

[0191] If waveform > V base - Sign1 = 1, inhale time T1 accumulates;

[0192] If waveform < V base - Sign2 = 1, inhale time T2 accumulates;

[0193] Period determination: when Sign1 = 1, Sign2 = 2 and waveform crosses baseline, trigger end of respiratory period, calculate:

[0194] Total period time: T cycle = T1 + T2 = 3.8s;

[0195] Respiratory frequency:

[0196] Step three,

[0197] OLED dynamic display: real-time refresh of breathing waveform (refresh rate 20Hz), superimposed display of frequency value;

[0198] Bluetooth transmission: data is packaged in JSON format;

[0199] Host computer interface: Python development platform, receives data and draws real-time breathing curve, supports doctor's annotation of abnormal events;

[0200] Step four,

[0201] Anti-interference test: in a 50Hz power frequency environment, the signal-to-noise ratio of the original signal is improved to 28dB;

[0202] Clinical comparison: compared with a medical breathing chest strap, the frequency error is <±0.5bpm.

[0203] In the description of the present specification, the description of the terms "one embodiment", "example", "specific example" and the like means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are contained in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above terms does 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.

[0204] The preferred embodiments of the application disclosed above are only used to help explain the application. The preferred embodiments do not describe all the details and limit the application to the specific embodiments described. Obviously, many modifications and changes can be made according to the content of the present specification. The present specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the application, so that those skilled in the art can well understand and utilize the application. The application is limited only by the claims and their entire scope and equivalents.

Claims

1. A portable breath monitoring method and apparatus based on a piezoelectric thin film sensor, characterized by: Comprise the following steps: Step S1: The piezoelectric film sensor of high polymer sensitive material is attached to the outside of the nasal cavity, ensuring that the electrode plate is perpendicular to the airflow direction; the vibration of the respiratory airflow causes the internal lattice of the sensor to generate surface charge by polar displacement, and the electric signal is output through the potential difference between the upper and lower electrode layers; Step S2: The electric signal is transmitted to the charge amplifier circuit through the shielding line, and 0.05 to 3Hz band-pass filtering and 5Hz low-pass filtering are performed in turn; Step S3: The filtered signal is converted into a digital signal using an ADC with a sampling rate of not less than 100Hz, and a DMA cyclic transmission mode is configured to open a dedicated buffer area to realize non-blocking transmission; Step S4: The converted and transmitted digital signal is smoothed using a Kalman filter, and the estimated error covariance is updated in the prediction stage, and the Kalman gain is calculated in the update stage to realize optimal estimation; Step S5: Based on the smoothed waveform, the signal is de-meaned and the zero-crossing interval sequence is detected by the zero-point analysis method, and the complexity parameter is calculated; Step S6: Set the baseline voltage, and use the baseline threshold judgment method to accumulate the inhalation and exhalation time according to the parameters obtained by feature extraction, and the timer interrupt frequency is 100Hz, and the frequency is calculated when the respiratory cycle is completed; Step S7: The calculated respiratory frequency and related waveform parameters are dynamically displayed by hardware IIC driving OLED, and are packaged and transmitted to the upper computer through Bluetooth protocol.

2. The portable breath monitoring method and apparatus based on piezoelectric thin film sensor according to claim 1, wherein, The step S1 comprises the following steps: Step S11, the sensor output potential difference is proportional to the input charge, satisfying the charge-voltage conversion relationship: V out ∝ ΔQ; where V out is the potential difference outputted by the sensor, a is the proportional relationship, and AQ is the amount of change in the surface charge of the sensor due to the vibration of the respiratory airflow. Step S12, the nasal airflow direction is perpendicular to the electrode plate to improve the sensitivity.

3. The portable breath monitoring method and apparatus based on piezoelectric thin film sensor according to claim 1, wherein, The step S2 comprises the following steps: Step S21, the charge amplifier circuit realizes impedance matching, analyzes the feedback impedance characteristics of the charge amplifier, and realizes impedance matching. The calculation formula of the feedback impedance is as follows: where Z f is the feedback impedance of the charge amplifier, R f is the feedback resistance, X cf is the reactance of the feedback capacitance, and j is the imaginary unit. Step S22, when R f sufficiently large simplifies to: Z f ≈ -jX cf ; Step S23, when analyzing the input impedance of the charge amplifier, the expression of the input impedance can be derived as follows: Z i = R i -j X ci ; wherein Z i is the input impedance of the charge amplifier, R i is the input resistance, X ci is the reactance of the input capacitance; Step S24, in order to calculate the voltage gain of the charge amplifier, the amplification capability of the circuit to the signal is evaluated. The approximate calculation formula of the voltage gain is as follows: In the formula, A is the voltage gain of the charge amplifier; Step S25, determine the high-pass cutoff frequency of the band-pass filter to screen out the signals of the required frequency band. The parameters of the band-pass filter are determined as follows. The high-pass cutoff frequency formula is as follows: where f H The high pass cutoff frequency of the band pass filter, R i is the input resistance, C i is the input capacitance; Step S26, determine the low-pass cutoff frequency of the band-pass filter to limit the high-frequency interference signal. The low-pass cutoff frequency calculation formula is as follows: where f L is the low pass cutoff frequency of the band pass filter, R f is the feedback resistance, C f is the feedback capacitance; Step S27, calculate the cutoff frequency of the RC passive low-pass filter circuit to clarify its suppression ability to high-frequency signals. The calculation formula of the cutoff frequency of the two-stage RC passive low-pass filter is as follows: where f0 is the cut-off frequency of the RC passive low-pass filter, R lp is the low-pass filter resistance, and C lp is the low-pass filter capacitance.

4. The portable breath monitoring method and apparatus based on piezoelectric thin film sensor according to claim 1, wherein, The step S3 comprises the following steps: Step S31, the ADC is configured to have a resolution of 12 bits and a sampling rate of ≥100Hz; Step S32, DMA transmission makes CPU and peripherals work in parallel.

5. The portable breath monitoring method and apparatus based on piezoelectric thin film sensor according to claim 1, wherein, The step S4 comprises the following steps: Step S41, Kalman filtering is performed in two stages: Prediction phase: Update the estimated error covariance, as follows: wherein is the a priori estimation error covariance at time k, A is the state transition matrix, P k-1 is the a posteriori estimation error covariance at time k-1, A T is the transpose of matrix A, Q is the process noise covariance; Update phase: Compute Kalman gain, formula as follows: And update the state estimate; In the formula, K k Let H be the Kalman gain at time k, and H be the observation matrix. T Let H be the transpose of matrix H, and R be the measurement noise covariance. Step S42, dynamically adjust the process noise covariance Q and the measurement noise covariance R to suppress the baseline drift.

6. The portable breath monitoring method and apparatus based on piezoelectric thin film sensor according to claim 1, wherein, The step S5 comprises the following steps: Step S51, the original signal is de-meaning processing, eliminate the influence of direct current component, facilitate subsequent zero-crossing analysis, de-meaning formula as follows: In the formula, y(n) is the signal after the de-meaning processing, x(n) is the original respiratory signal, N is the total number of points of the signal sequence, is the summation operation from the 1st point to the Nth point of the original signal, is the mean value of the original signal; Step S52, the zero-crossing interval sequence d1, d2,..., d M ; Step S53, measure the complexity of the signal, by calculating the mean of the zero-crossing interval sequence as a characteristic parameter, the calculation formula of the mean mZCI as follows: wherein mZCI is the mean of the zero-crossing interval sequence, M is the total number of zero-crossing intervals, d(m) is the mth zero-crossing interval, is the summation over all zero-crossing intervals; Step S54, further quantification of the complex characteristics of the signal, calculate the normalized standard deviation of the zero-crossing interval sequence, nsZCI calculation formula as follows: where nsZCI is the normalized standard deviation of the zero-crossing interval sequence, [d(m) - mZCI]2 2 is the square of the difference of the mth zero-crossing interval from the mean, is the sum of the square of the difference of all intervals from the mean. Step S55, respiratory signal mZCI value is larger, using baseline threshold method anti-pseudo-peak interference.

7. The portable breath monitoring method and apparatus based on piezoelectric thin film sensor according to claim 1, wherein, The step S6 includes the following steps: Step S61, the baseline voltage is set to 1.65V; Step S62, define integer variable T1: inspiration time; T2: expiration time; Step S63, timer interrupt frequency 100Hz (period 10ms), meet the accuracy requirement of 3s breathing cycle; Step S64, flag assignment rules: Waveform>baseline→Sign1=1 and T1←T1+1; Waveform<baseline→Sign2=1 and T2←T2+1; Step S65: respiratory cycle determination condition: Sign1=1 and Sign2=1 and waveform across baseline; Step S66, trigger after execution of zero: T1=0, T2=0, Sign1=0, Sign2=0; Step S67, according to the recorded respiratory cycle time to calculate the respiratory rate, get the number of breaths per minute, respiratory rate calculation formula as follows: In the formula, T1+T2 is the total time of a complete respiratory cycle, 60 is the cycle time converted to the number of breaths per minute.

8. The portable breath monitoring method and apparatus based on piezoelectric thin film sensor according to claim 1, wherein, The step S7 includes the following steps: Step S71, OLED through the hardware IIC interface to realize waveform dynamic refresh; Step S72, Bluetooth transmission data encapsulation format contains timestamp, respiratory rate value, device ID; Step S73, the host computer display interface for doctors to carry out subsequent processing.

Images