TW201306799A - Method and device of identification with characteristics of sleep apnea, cough and asthma - Google Patents

Abstract

This invention relates to the method and device of identification with characteristics of sleep apnea, cough and asthma. About the method, it includes (1) preparing step, (2) body detecting step, (3) signal processing step, (4) signal analyzing step, and (5) finishing step. This device includes a voice detector, an abdomen detector, and a wrist detector for detecting a lung sound signal, an abdomen breath signal, and a wrist movement signal respectively. This device also includes three fuzzy logic systems so as to analyze the characteristics of sleep apnea, cough and asthma based on these three detected signals. Then, it can judge which event is occurred or not. Thus, this invention can precisely analyze the events times and happening time of sleep apnea, cough and asthma. In addition, its wireless detecting design can effectuate a long period of detection time under a static or dynamic status.

Description

Method and device for identifying features of sleep apnea, cough and asthma

The invention relates to a method and a device for identifying the characteristics of sleep breathing, cough and asthma, in particular to a method and device for identifying characteristics of sleep breathing, coughing and asthma, and the utility model has the characteristics of fuzzy logic to identify asthma. The events and sleep-dissipation events and coughing events are quite accurate, and the number and timing of asthmatic events and sleep-disordering events and coughing events can be identified, and the wireless sensing design can perform long-term detection in both static and dynamic conditions.

The signs of respiratory tract during sleep are indicators of many disease control, such as asthma, sleep breathing or coughing. Among them, sleep breathing is suspended according to different standard requirements, there are many products for inspection and monitoring. The traditional detection methods and devices are divided into four levels, which respectively produce the following problems:

[1] The scope of application is limited. The first level is the full-channel standard examination, which is used in hospitals; the evaluation of the effectiveness of the treatment of sleep-disorder (such as positive pressure breathing apparatus, surgery or braces, etc.) is most ideal for long-term observation. Although it is most accurate to go to the hospital for confirmation, it still needs to be scheduled to wait; it is not possible for people with mobility problems or those who do not have time to go to the hospital for immediate examination, and is restricted to hospital use.

[2] Inconvenient to use. The second level of sleep is for all channels in non-medical personnel; the third level is for chest and abdomen, respiratory airflow and blood oxygen concentration as screening and treatment evaluation; monitoring in the second and third levels It is an alternative method, but it may be necessary to carefully pull the relevant sensing line when the activity is active, which is quite inconvenient for long-term use.

[3] Detection is less accurate. The fourth level of sleep is the detection of blood oxygen concentration. It is a simple screening of blood oxygen status and further inference of respiratory conditions. There is no sleep detection signal in the third stage sleep examination. The error rate is relatively high in estimating sleep breathing. As for the traditional device for detecting asthma or detecting cough (for example, the "Spot Physiological Monitoring Device, System and Network" of the Republic of China Invention Patent Publication No. 200704391), there is no recognition of sleep function, and it is impossible to accurately identify whether asthma or cough is present. The problem of sleep. The portable structure (for example, the "portable asthma lung sound monitoring system" of the Republic of China invention patent No. I294280) is also incapable of distinguishing whether the detected signal occurs in waking or sleeping, and the detection is less accurate. In addition, there are also devices that use the personally recorded lung sound database for identification. However, each person has different performances in different diseases, and this device also produces errors.

[4] Wired design is inconvenient to use. Conventional devices for detecting asthma or detecting cough (for example, "SMD-type physiological monitoring devices, systems, and networks" in the Republic of China Invention Patent Publication No. 200704391) still support sensing in a wired manner, so that the test subject is In action or lying on a hospital bed, it increases the chance of pulling during natural activities, and the cable design is inconvenient to use.

In view of this, it is necessary to develop a technique that can solve the above disadvantages.

The object of the present invention is to provide a method and a device for identifying characteristics of sleep breathing, coughing and wheezing, which have the characteristics of using fuzzy logic to identify an asthmatic event, a sleep breathing abort event and a coughing event, and can identify an asthmatic event. And the number and timing of sleep apnea events and coughing events, and the advantages of wireless sensing design for long-term detection both statically and dynamically. In particular, the problems to be solved by the present invention include: limited application range, inconvenient use, inaccurate detection, and inconvenient use of wired designs.

The technical means for solving the above problems is to provide a method for identifying characteristics of sleep breathing, coughing and wheezing, and a device thereof, the method comprising the following steps:

One. Preparation steps

two. Human body detection step;

three. Signal processing step;

four. Signal identification step; and

Fives. Complete the steps.

The device part comprises: a sound sensing device for detecting a patient to be detected and generating a lung sound signal and transmitting the signal to the detection and identification device; and an abdominal sensing device for detecting Measuring the abdominal breathing process of the test subject, and generating a abdominal respiratory motion signal, and transmitting to the detection and identification device; a wrist sensing device for detecting the test subject to generate a wrist The activity signal is transmitted to the detection and identification device; a detection and identification device has a first fuzzy logic system, a second fuzzy logic system, a third fuzzy logic system and an estimation identification device; The identification device is configured to receive the lung sound signal, the abdominal respiratory motion signal and the wrist activity signal; first perform fast Fourier transform on the lung sound signal to obtain a frequency, and signal the lung sound signal and the abdominal respiratory motion signal Cutting processing; thereby, the detection identification device cooperates with the lung sound signal, the abdominal respiratory motion signal and the wrist activity signal to perform the following three identification operations:

[a] Asthma identification: the detection and recognition device is identified by the first fuzzy logic system; and the asthma attack is characterized by wheezing; when the frequency of the lung sound signal reaches a frequency standard value, the system is between 200 Hz and 600 Hz And the frequency standard value continues for a first identification time, which is between 250 milliseconds and 500 milliseconds; and is recognized as wheezing, and is defined as an asthmatic event of the test subject;

[b] sleep breathing abort identification: the detection and recognition device is identified by the second fuzzy logic system; when the adjacent lung sound signal interval continues for a second identification time; the system is between 10 seconds and 90 seconds And the adjacent abdominal respiratory motion signal continues to be separated by the second identification time; and the wrist activity signal determines that the test subject is in a sleep state; and is defined as a sleep breathing suspension event of the test subject;

[c] Cough recognition: the detection and recognition device is identified by the third fuzzy logic system; when the expiratory phase of each lung sound signal continues for an expiratory phase time, the system is between 0.3 seconds and 1 second. And the adjacent lung sound signals are spaced apart from each other by a third identification time, which is between 3 seconds and 6 seconds; and is defined as a coughing event of the test subject;

Finally, the estimation and identification device compares the defuzzification value and identifies the probability of occurrence according to the sequence of the asthmatic event, the sleep breathing abort event, and the cough event in the overlapping time of the first, the second, and the third identification time. The largest event, and identify it as the event of the person to be tested.

The above objects and advantages of the present invention will be readily understood from the following detailed description of the preferred embodiments illustrated herein.

The invention will be described in detail in the following examples in conjunction with the drawings:

The present invention is a method and apparatus for identifying characteristics of sleep apnea, cough and wheezing. Referring to the first figure, the method portion includes the following steps:

One. Preparation Step 11: Referring to the second figure, a sound sensing device 20, an abdominal sensing device 30, a wrist sensing device 40, and a detection and recognition device 50 are prepared in advance. The detection and recognition device 50 has a first a fuzzy logic system 51, a second fuzzy logic system 52, a third fuzzy logic system 53 and an estimation and identification device 54;

two. The human body detecting device 12 is configured to detect a test subject 91 and generate a lung sound signal 20A and transmit the same to the detection and recognition device 50. The abdominal sensing device 30 is used for instant detection. The abdominal respiratory motion of the test subject 91 is measured, and a abdominal respiratory motion signal 30A is generated and transmitted to the detection and recognition device 50. The wrist sensing device 40 is configured to detect the tester 91 in real time. a wrist activity signal 40A, and transmitted to the detection and identification device 50;

three. Signal processing step 13: The detection and recognition device 50 first performs a Fast Fourier Transform (FFT) for the lung sound signal 20A to obtain a frequency (refer to the fourth figure), and the lung sound signal 20A and the Abdominal respiratory motion signal 30A performs signal segmentation processing;

four. The signal recognition step 14: the detection and recognition device 50 performs the following three identification operations using the lung sound signal 20A, the abdominal respiratory motion signal 30A and the wrist activity signal 40A (see the third figure):

[a] Asthma recognition: the detection and recognition device 50 is identified by the first fuzzy logic system 51; and the asthma attack is characterized by wheezing; when the frequency of the lung sound signal 20A reaches a frequency standard value, the system is interposed Between 200 Hz and 600 Hz (refer to the fourth figure (indicated by "medium" and other frequencies), and the frequency standard value continues for a first identification time T1, which is between 250 milliseconds and 500 milliseconds (see the fifth figure, It is expressed as "long".); it is recognized as wheezing, and is defined as the asthmatic event of the test subject 91;

[b] sleep breathing abort identification: the detection and recognition device 50 is identified by the second fuzzy logic system 52; when the adjacent lung sound signal 20A is separated by a second identification time T2; the system is between 10 seconds~ Between 90 seconds (refer to the sixth figure, indicated by "long"); and the adjacent abdominal respiratory motion signal 30A continues at the same interval for the second identification time T2 (refer to the seventh figure, indicated by "long".) And determining, by the wrist activity signal 40A, the test subject 91 is in a sleep state; then defining that the test subject 91 has a sleep breathing suspension event;

[c] Cough recognition: the detection and recognition device 50 recognizes the third fuzzy logic system 53; when the expiratory phase of each lung sound signal 20A continues for an expiratory phase time T31 (see eighth, eighth, eighth) B, eighth C, and eighth D), the system is between 0.3 seconds and 1 second; and the adjacent lung sound signals 20A are spaced apart from each other by a third identification time T32, which is between 3 seconds and 6 seconds. Between seconds; it is defined as the occurrence of a coughing event in the test subject 91;

Finally, the estimation and identification device 54 performs deblurring in the overlapping time of the first, second, and third identification times T1, T2, and T32 according to an asthmatic event, a sleep breathing abort event, and a coughing event. The numerical value is used to identify the event with the highest probability of occurrence, and it is determined that it is the event of the tester 91; Complete step 15: Complete identification of asthmatic events, sleep breathing abortion events, and coughing events.

In practice, the frequency standard value of the lung sound signal 20A can be 400 Hz as the optimum value.

The first identification time T1 can be between 100 milliseconds and 500 milliseconds.

The second identification time T2 can be between 10 seconds and 120 seconds.

The expiratory phase time T31 can be between 0.12 seconds and 2.5 seconds.

The third identification time T32 can be between 2 seconds and 10 seconds.

The sound sensing device 20 includes a first fixing portion 21, a first housing 22, an acoustic sensing component 23, a first memory 24, a first microprocessor 25, and a first radio frequency. The first fixing portion 21 and the first housing portion 22 are configured to be coupled to each other (the technical means can be connected to each other), and the first fixing portion 21 is used for Fixing the first housing 22 to the chest of the test subject 91 (in principle, it is flat on the skin of the chest, close to the position of the lung for long-term detection); the sound sensing component 23 ( For example, the "Micro-microphone diaphragm film and its manufacturing method" of the Republic of China invention patent No. I268115 is used to sense the lung sound signal 20A of the test subject 91; the first memory The body 24 is configured to transmit the lung audio signal 20A to the first microprocessor 25, and the first microprocessor 25 is configured to control the first radio frequency component 26 to transmit the lung audio signal 20A; A battery 27 supplies the power required by the aforementioned associated components.

The abdominal sensing device 30 includes a second fixing portion 31, a second housing 32, a first acceleration sensor 33, a second memory 34, a second microprocessor 35, and a second The radio frequency component 36 and the second battery 37; the second fixing portion 31 and the second casing 32 are combined with each other (the technical means can be connected to each other), and the second fixing portion 31 For fixing the second casing 32 to the abdomen of the test subject 91 (in principle, the skin close to the abdomen for long-term detection); the first acceleration sensor 33 is used to sense the waiting The abdomen of the tester 91 is in an undulating state of the breathing process, and a voltage change is generated, and the voltage change can be converted into the abdominal respiratory motion signal 30A; the second memory 34 is used to memorize and transmit the abdominal respiratory motion signal 30A to the first The second microprocessor 35 is configured to control the second radio frequency component 36 to transmit the abdominal respiratory motion signal 30A; the second battery 37 is to supply the power required by the related components.

The wrist sensing device 40 includes a wrist fixing strap 41, a third housing 42, a second acceleration sensor 43, a third memory 44, a third microprocessor 45, and a first a third radio frequency component 46, a third battery 47, a button group 48 and a screen 49; the wrist fixing strap 41 is used to fix the third housing 42 to the wrist of the test subject 91; the second acceleration The sensor 43 is configured to sense a change in the muscle tension of the wrist of the test subject 91 (there is a voltage change, and when the set voltage is decreased, the muscle tension becomes small → the test subject 91 sleeps. The voltage increases. When the muscle tension becomes large, the test subject 91 is awake. The voltage change is converted into the wrist activity signal 40A, and the third memory 44 is used to memorize the wrist activity signal 40A, and a multiplexer 441 is provided. The received lung sound signal 20A, the abdominal respiratory motion signal 30A, and the wrist activity signal 40A are multiplexed to the third microprocessor 45, and the third microprocessor 45 is configured to control the third wireless RF component. 46, the lung sound signal 20A, the abdominal respiratory motion signal 30A and the wrist activity signal 40A are transmitted The third battery 47 is supplied with power required by the aforementioned related components; the button group 48 includes a selection button 481, a confirmation button 482 and a power button 483; the selection button 481 is used to select a function to be performed; the confirmation button 482 is used to confirm the selected function; the power button 483 is used to open and close the wrist sensing device 40; the screen 49 is used to display the signal of the wrist sensing device 40 or related detection results.

The first and the second fixing portions 21 and 31 are used for flattening on the surface of the skin, and are not easily detached, and can be easily replaced when sweating or other factors cause detachment.

The detection and recognition device 50 can be disposed in the wrist sensing device 40 (used to be carried around for 24 hours), and the sound sensing device 20, the abdominal sensing device 30 and the wrist sensing device 40 can be permeable. The first, the second, and the third radio frequency components 26, 36, and 46 constitute a wireless sensing network system, and the third radio frequency component 46 is configured as a cluster head to configure a network transmission architecture; External devices (such as computers or embedded processors) can download data; they can also be directly connected to external computers or embedded processors for RAW Data analysis.

Referring to the second and third figures, the detection and identification device 50 further includes a signal processing module 50A for receiving the lung sound signal 20A, the abdominal respiratory motion signal 30A and the wrist from the multiplexer 441. The signal processing module 50A is provided with a fast Fourier transform unit 501 for performing Fourier transform (FFT) on the lung sound signal 20A (in this case, a time domain signal) (for discrete time aperiodic The calculation of the signal becomes a frequency domain signal 20A(f); its processing is based on the following formula: The calculation is performed (known techniques are not described here); wherein: x[n] is the signal after sampling, Ω is the unit (diameter); the discrete-time Fourier transform is the frequency distribution of the interval -π<Ω≦π.

A non-linear energy operation unit 502 is configured for the lung sound signal 20A (general time domain signal), the abdominal respiratory motion signal 30A (general time domain signal), and the wrist activity signal 40A (general time domain) The signal and the frequency domain signal 20A(f) perform signal cutting, and then transmit to the first, second and third fuzzy logic systems 51, 52 and 53; the processing method is as follows: input signal x ( n )= A cos( ωn + + w ( n ) performs the operation; where: w ( n ) represents the channel effect of white Gauss; A is the amplitude of the input signal; ω is the frequency of the input signal; as for the nonlinear energy operation, the following formula can be applied: Ψ[x( n)]=x(n-1) ² -x(n)x(n-2) is operated.

A data storage unit 55 (see FIGS. 2 and 22) for storing the signals and results processed by the estimation and identification device 54.

In the signal identification step 14, when the [a] asthma recognition is performed, the wrist activity signal 40A can be further matched (as the muscle tension of each person is different, and the point is based on the actual measurement reference value) The test subject 91 is sleep/awake (same reason, the cough recognition can also be determined by the push to determine that the test subject 91 is sleep/awake).

Here we first explain fuzzy logic. There are many phenomena in nature that cannot be defined by a clear binary method with or without. Fuzzy logic provides a solution; for example, height is 180 cm "high", greater than 180 is high, and less than 160 is low, and the degree of attribution can be set to 0 to 1. If the description is "very high", the degree of attribution greater than 180 cm is 1,160 cm 0; for the description of "short", the degree of attribution of 165 cm may be 0.7; this fuzzification depends on the member function. When the input value is blurred, the logical solution is inferred and finally defuzzified to get the best solution.

Taking the first fuzzy logic system 51 as an example of asthma recognition, the frequency of the lung sound signal 20A of the asthma event is set to 400 Hz, and the frequency of the normal lung sound signal 20A is 100 to 200 Hz (based on 100 Hz); For asthma recognition, the attribution of 400Hz is 1,100Hz, and the attribution degree is 0. Referring to the ninth figure, the frequencies of the six lung sound signals 20A shown in the figure all reach the frequency standard value (400Hz) (the attribution is 1), and the frequency standard value continues for the first identification time T1 (250 milliseconds), defined as an asthma event, the estimation identification device 54 overlaps the first, second, and third identification times T1, T2, and T32 During the time, it is determined that the subject 91 has six asthmatic events.

When the second fuzzy logic system 52 (the setting process is the same as the first fuzzy logic system 51, and will not be described later), the sleep breathing suspension identification is performed, refer to the tenth A and tenth B pictures (the lung sound signal 20A and the respectively Abdominal respiratory motion signal 30A is the waveform after cutting process), the figure shows the waveform above the respiratory airflow pause for more than 10 seconds (the second identification time T2 in the figure is 28 seconds) (reference for sleep breathing suspension event) a feature is defined as a sleep apnea event; the estimation and identification device 54 determines that the test subject 91 has a sleep breath during the overlap time of the first, second, and third identification times T1, T2, and T32. Suspend the incident.

When the third fuzzy logic system 53 (the setting process is the same as the first fuzzy logic system 51, and will not be described later), the cough identification is performed, refer to the 11th and 11th B pictures (the lung sound signal 20A and the respectively Abdominal respiratory motion signal 30A is the waveform after cutting.) The picture shows that there are four exhalation phases in accordance with the expiratory phase time T31 (between 0.3 seconds and 1 second), and the interval is up to the third identification time T32. (interval between 3 seconds and 6 seconds); defined as a coughing event; the estimation and identification device 54 discriminates between the first, second, and third identification times T1, T2, and T32 The subject 91 had four coughing events.

When the lung sound signal 20A, the abdominal respiratory motion signal 30A, and the wrist activity signal 40A are recognized by the first, the second, and the third fuzzy logic systems 51, 52, and 53 , the asthmatic event and the sleep breathing stop are not met. In the case of an event or a coughing event, the estimation and identification device 54 determines that it is normal or noise. For example, as shown in FIG. 12A and FIG. 12B, the lung sound signal 20A and the abdomen of a certain subject 91 respectively. The waveform of the respiratory motion signal 30A after cutting, the frequency of the waveform shown in the figure is less than the frequency standard value (400 Hz) of the "anti-asthmatic event", the adjacent lung sound signal 20A and the adjacent abdominal respiratory motion signal 30A The duration of the interval did not reach the second identification time T2 of the sleep apnea event (10 seconds or more); the expiratory phase time T31 and the third identification time T32 of the expiratory phase of each lung sound signal 20A did not reach the "cough event" The standard value of the estimation means 54 determines that the subject 91 is in a normal breathing state.

The device of the present invention includes (see the second and third figures): a sound sensing device 20 for detecting a test subject 91 and generating a lung sound signal 20A, and transmitting to the detection The identification device 50 is an abdominal sensing device 30 for detecting the abdominal breathing process of the test subject 91, and generating a abdominal respiratory motion signal 30A, and transmitting to the detection and recognition device 50; The sensing device 40 is configured to detect the subject 91 and generate a wrist activity signal 40A and transmit the motion signal to the detection device 50. The detection device 50 has a first fuzzy logic system. 51. A second fuzzy logic system 52, a third fuzzy logic system 53 and an estimation and identification device 54. The detection and identification device 50 is configured to receive the lung sound signal 20A, the abdominal respiratory motion signal 30A and the wrist The activity signal 40A; firstly, the lung sound signal 20A is subjected to Fast Fourier Transform (FFT) to obtain the frequency (refer to the fourth figure), and the lung sound signal 20A and the abdominal respiratory motion signal 30A are signaled. Segmentation processing; borrowing To fit the 50 lung sound identification signal to the detecting means 20A, the signals 30A and abdominal respiratory movement of the arm 40A for the following three activities signal identification operation (see FIG third):

[a] Asthma recognition: the detection and recognition device 50 is identified by the first fuzzy logic system 51; and the asthma attack is characterized by wheezing; when the frequency of the lung sound signal 20A reaches a frequency standard value, the system is interposed Between 200 Hz and 600 Hz (refer to the fourth figure (indicated by "medium" and other frequencies), and the frequency standard value continues for a first identification time T1, which is between 250 milliseconds and 500 milliseconds (see the fifth figure, It is expressed as "long".); it is recognized as wheezing, and is defined as the asthmatic event of the test subject 91;

[b] sleep breathing abort identification: the detection and recognition device 50 is identified by the second fuzzy logic system 52; when the adjacent lung sound signal 20A is separated by a second identification time T2; the system is between 10 seconds~ Between 90 seconds (refer to the sixth figure, indicated by "long"); and the adjacent abdominal respiratory motion signal 30A continues at the same interval for the second identification time T2 (refer to the seventh figure, indicated by "long".) And determining, by the wrist activity signal 40A, the test subject 91 is in a sleep state; then defining that the test subject 91 has a sleep breathing suspension event;

[c] Cough recognition: the detection and recognition device 50 recognizes the third fuzzy logic system 53; when the expiratory phase of each lung sound signal 20A continues for an expiratory phase time T31 (see eighth, eighth, eighth) B, eighth C, and eighth D), the system is between 0.3 seconds and 1 second; and the adjacent lung sound signals 20A are spaced apart from each other by a third identification time T32, which is between 3 seconds and 6 seconds. Between seconds; it is defined as the occurrence of a coughing event in the test subject 91;

Finally, the estimation and identification device 54 performs deblurring in the overlapping time of the first, second, and third identification times T1, T2, and T32 according to an asthmatic event, a sleep breathing abort event, and a coughing event. The numerical value is used to identify the event with the highest probability of occurrence, and it is determined that it is an event occurring by the test subject 91.

In practice, the frequency standard value of the lung sound signal 20A can be 400 Hz as the optimum value.

The first identification time T1 can be between 100 milliseconds and 500 milliseconds.

The second identification time T2 can be between 10 seconds and 120 seconds.

The expiratory phase time T31 can be between 0.12 seconds and 2.5 seconds.

The third identification time T32 can be between 2 seconds and 10 seconds.

The sound sensing device 20 includes a first fixing portion 21, a first housing 22, an acoustic sensing component 23, a first memory 24, a first microprocessor 25, and a first radio frequency. The first fixing portion 21 and the first housing portion 22 are configured to be coupled to each other (the technical means can be connected to each other), and the first fixing portion 21 is used for Fixing the first housing 22 to the chest of the test subject 91 (in principle, it is flat on the skin of the chest, close to the position of the lung for long-term detection); the sound sensing component 23 ( For example, the "Micro-microphone diaphragm film and its manufacturing method" of the Republic of China invention patent No. I268115 is used to sense the lung sound signal 20A of the test subject 91; the first memory The body 24 is configured to transmit the lung audio signal 20A to the first microprocessor 25, and the first microprocessor 25 is configured to control the first radio frequency component 26 to transmit the lung audio signal 20A; A battery 27 supplies the power required by the aforementioned associated components.

The abdominal sensing device 30 includes a second fixing portion 31, a second housing 32, a first acceleration sensor 33, a second memory 34, a second microprocessor 35, and a second The radio frequency component 36 and the second battery 37; the second fixing portion 31 and the second casing 32 are combined with each other (the technical means can be connected to each other), and the second fixing portion 31 For fixing the second casing 32 to the abdomen of the test subject 91 (in principle, the skin close to the abdomen for long-term detection); the first acceleration sensor 33 is used to sense the waiting The abdomen of the tester 91 is in an undulating state of the breathing process, and a voltage change is generated, and the voltage change can be converted into the abdominal respiratory motion signal 30A; the second memory 34 is used to memorize and transmit the abdominal respiratory motion signal 30A to the first The second microprocessor 35 is configured to control the second radio frequency component 36 to transmit the abdominal respiratory motion signal 30A; the second battery 37 is to supply the power required by the related components.

The wrist sensing device 40 includes a wrist fixing strap 41, a third housing 42, a second acceleration sensor 43, a third memory 44, a third microprocessor 45, and a first a third radio frequency component 46, a third battery 47, a button group 48 and a screen 49; the wrist fixing strap 41 is used to fix the third housing 42 to the wrist of the test subject 91; the second acceleration The sensor 43 is configured to sense a change in the muscle tension of the wrist of the test subject 91 (there is a voltage change, and when the set voltage is decreased, the muscle tension becomes small → the test subject 91 sleeps. The voltage increases. When the muscle tension becomes large, the test subject 91 is awake. The voltage change is converted into the wrist activity signal 40A, and the third memory 44 is used to memorize the wrist activity signal 40A, and a multiplexer 441 is provided. The received lung sound signal 20A, the abdominal respiratory motion signal 30A, and the wrist activity signal 40A are multiplexed to the third microprocessor 45, and the third microprocessor 45 is configured to control the third wireless RF component. 46, the lung sound signal 20A, the abdominal respiratory motion signal 30A and the wrist activity signal 40A are transmitted The third battery 47 is supplied with power required by the aforementioned related components; the button group 48 includes a selection button 481, a confirmation button 482 and a power button 483; the selection button 481 is used to select a function to be performed; the confirmation button 482 is used to confirm the selected function; the power button 483 is used to open and close the wrist sensing device 40; the screen 49 is used to display the signal of the wrist sensing device 40 or related detection results.

The first and the second fixing portions 21 and 31 are used for flattening on the surface of the skin, and are not easily detached, and can be easily replaced when sweating or other factors cause detachment.

The detection and recognition device 50 can be disposed in the wrist sensing device 40 (used to be carried around for 24 hours), and the sound sensing device 20, the abdominal sensing device 30 and the wrist sensing device 40 can be permeable. The first, the second, and the third radio frequency components 26, 36, and 46 constitute a wireless sensing network system, and the third radio frequency component 46 is configured as a cluster head to configure a network transmission architecture; External devices (such as computers or embedded processors) can download data; they can also be directly connected to external computers or embedded processors for RAW Data analysis.

Referring to the second and third figures, the detection and identification device 50 further includes a signal processing module 50A for receiving the lung sound signal 20A, the abdominal respiratory motion signal 30A and the wrist from the multiplexer 441. The signal processing module 50A is provided with a fast Fourier transform unit 501 for performing Fourier transform (FFT) on the lung sound signal 20A (in this case, a time domain signal) (for discrete time aperiodic The calculation of the signal becomes a frequency domain signal 20A(f); its processing is based on the following formula: The calculation is performed (known techniques are not described here); wherein: x[n] is the signal after sampling, Ω is the unit (diameter); the discrete-time Fourier transform is the frequency distribution of the interval -π<Ω≦π.

A non-linear energy operation unit 502 is configured for the lung sound signal 20A (general time domain signal), the abdominal respiratory motion signal 30A (general time domain signal), and the wrist activity signal 40A (general time domain) The signal and the frequency domain signal 20A(f) perform signal cutting, and then transmit to the first, second and third fuzzy logic systems 51, 52 and 53; the processing method is as follows: input signal x ( n )= A cos( ωn + + w ( n ) performs the operation; where: w ( n ) represents the channel effect of white Gauss; A is the amplitude of the input signal; ω is the frequency of the input signal; as for the nonlinear energy operation, the following formula can be applied: Ψ[x( n)]=x(n-1) ² -x(n)x(n-2) is operated.

A data storage unit 55 (see FIGS. 2 and 22) for storing the signals and results processed by the estimation and identification device 54.

As for the part of the fuzzy computing application, the laundry time design of the known washing machine is taken as an example:

For the clothing to be cleaned, the "sludge degree" and the "degree of oil stain" are used as input conditions, and the cleaning time is output. The fuzzy functions are shown in the thirteenth, fourteenth and fifteenth figures, respectively. One:

"Clothing sludge: medium; oil: medium, cleaning time: long time."

"Clothing sludge: medium, oil stain: large, cleaning time: long time."

"Clothing sludge: high, oil: medium, cleaning time: long time."

"Clothing sludge: high, oily: large, cleaning time: very long time."

Its fuzzy inference:

Suppose the sludge is 120 and the oil is 140.

Then, the sludge ownership degree is "in" = 0.8, and "high" = 0.2.

Oil pollution attribution "in" = 0.6, "large" = 0.4.

By min operation:

Min (medium, middle) = min (0.8, 0.6) = 0.6; min (medium, large) = min (0.8, 0.4) = 0.4.

Min (high, medium) = min (0.2, 0.6) = 0.2; min (high, large) = min (0.2, 0.4) = 0.2.

The value of the min operation result is further subjected to the max operation:

Max (0.6, 0.4, 0.2) = 0.6 (long); Max (0.2) = 0.2 (long).

Then proceed (for example, to solve the blurring by the center of gravity method) and blur it to (13+23)/2=18 (that is, the washing time).

When applied to the present invention:

[a] Take the identification of asthmatic events as an example: refer to the sixteenth figure, which is a waveform of 120 minutes in the total detection time (assumed to be 24 hours a day), taking the first lung sound signal 20A as an example, the frequency is 200Hz (refer to the fourth figure), and the lung sound signal 20A is cut to maintain the first identification time T1 is 250ms (refer to the fifth figure), the min-max operation is as follows:

By min operation:

MIN (very short, high) = (0, 0) = 0; MIN (very short, medium) = (0, 0) = 0; MIN (very short, low) = (1, 0) = 0; MIN ( Short, high) = (0, 0) = 0; MIN (short, medium) = (0, 0) = 0; MIN (short, low) = (1, 0) = 0; MIN (medium, high) = (0,0)=0;MIN(medium,middle)=(0,0)=0;MIN(medium,low)=(1,0)=0;MIN(long,high)=(0,1) =0; MIN (long, medium) = (0, 1) = 0; MIN (long, low) = (1, 1) = 1.

The first fuzzy logic system 51 recognizes that the set frequency of the asthmatic event is "medium"; and the matching requirement that the first identification time T1 is "long" is "high recognition", and the rest are "low recognition". And MIN (long, medium) = (0, 1) = 0, the recognition degree "high" in the seventeenth figure is 0, and the maximum value of the recognition "low" is 1, the degree of attribution in the seventeenth figure 1 is the last to form the asthma recognition line segment L1, the defuzzification value is 25%, and 20% can be set (50% can also be set as the criterion for judging). If it is set to 20%, 25% is judged as Asthma, or set to 50%, 25% of the judgment does not meet the asthma standard, all depending on actual needs.

[b] For example, to identify the sleep breathing abort event: Refer to Figure 18, and take the waveform of 120 minutes in the total detection time (assumed to be 24 hours a day), and set the first lung sound signal 20A after cutting. The interval maintains the second identification time T2 (refer to the sixth figure, for example, 50 seconds), and the abdominal respiratory motion signal 30A also maintains the second identification time T2 after the cutting interval (refer to the seventh figure, for example, 50 seconds), and its min The -max operation is shown in Table 3 below:

By min operation:

MIN (very long, very long) = (0, 0) = 0; MIN (very long, long) = (0, 0) = 0; MIN (very long, short) = (0, 0) = 0; MIN (long, very long) = (1, 1) = 1; MIN (long, long) = (1, 1) = 1; MIN (long, short) = (1, 1) = 1; MIN (short, very Long) = (0, 0) = 0; MIN (short, long) = (0, 0) = 0; MIN (short, short) = (0, 0) = 0.

The second fuzzy logic system 52 recognizes that the sleep apnea event is set to set the second identification time T2 interval of the interval after the lung sound signal 20A is cut to be "long", and the interval of the abdominal respiratory motion signal 30A is cut. When the second identification time T2 is "long", the recognition of the sleep breathing abort event is high; the remaining combination recognition is low.

And MIN (long, long) = (1, 1) = 1, in the nineteenth figure, the recognition "high" is 1, the recognition "low" is 0, and finally constitutes the sleep breathing abort identification line segment L2, defuzzification The value is 75%.

[3] Take the identification of coughing event as an example: Refer to the 20th figure and take the waveform of 120 minutes in the total detection time (assumed to be 24 hours a day). Take the first lung sound signal 20A as an example. After the 20A cutting, the first identification time T1 is maintained for 200ms (refer to the fifth figure), and the adjacent lung sound signal 20A is maintained at the second identification time T2 (refer to the sixth figure, for example, 4 seconds), and its min-max The operation is as follows in Table 4:

By min operation:

MIN (short, very short) = (1, 0) = 0; MIN (long, very short) = (0, 0) = 0; MIN (very long, very short) = (0, 0) = 0; MIN (short, short) = (1, 0) = 0; MIN (long, short) = (0, 0) = 0; MIN (very long, short) = (0, 0) = 0; MIN (short, medium ) = (1, 0.3) = 0.3; MIN (long, medium) = (0, 0.3) = 0; MIN (very long, medium) = (0, 0.3) = 0; MIN (short, long) = (1 , 0.5) = 0.5; MIN (long, long) = (0, 0.5) = 0; MIN (very long, long) = (0, 0.5) = 0.

Since the third fuzzy logic system 53 recognizes that the coughing event is set to maintain the first identification time T1 as "medium" and the second identification time T2 is "short" to "high recognition", the remaining combinations are all "identification". Low", while MIN (short, medium) = (1, 0.3) = 0.3 (high recognition), as the recognition is low (short, long) = (1, 0.5) = 0.5. In the twenty-first figure; the max is low (18,32), the high degree of recognition is (60,100), and the defuzzification method is (60+100)/2=80%.

The operation process of the estimation and identification device 54 is: three signals input by the multiplexer 441 (refer to the third figure) (including the lung sound signal 20A, the abdominal respiratory motion signal 30A, and the wrist activity signal 40A), The first and second fuzzy logic systems 51, 52, and 53 obtain deblurring values, that is, enter the estimation and identification device 54 to perform threshold determination (see FIG. 22), time series analysis, and Repeatability comparison. The threshold can be set according to the strictness, for example, 50% of the recognition can enter the analysis; the time series analysis is a fixed time (such as 120 seconds) of the fixed time (such as every 1 second) is the calculation unit, if different events occur within the unit Then do a repetitive comparison. For example, suppose the signal generated by t1-t2 time (120 minutes) is 25%, and the threshold (assumed to be 50%) is judged as a rejection signal. The signal generated at 120 minutes of t2-t3 is 87.5%, and the threshold is assumed to be 80%. After passing the threshold, the time series analysis is the same unit time, that is, the repeat comparison is estimated as a "sleep breathing stop event". However, if the values are the same, it is judged by asthma > sleep breathing suspension > cough. After completion, it can be displayed on the screen 49 or stored in the data storage unit 55.

Referring to the twenty-third figure, the estimation and identification process of the estimation and identification device 54 may include the following steps after the start:

Step A (61): Input the defuzzified lung sound signal, abdominal respiratory motion signal and wrist activity signal.

Step B (62): Is the discrimination signal greater than the threshold? If the result of the discrimination is "Yes" (take the first fuzzy logic system to identify the asthmatic event as an example, assume that the threshold is 20%, and if the signal obtained by identifying the asthma event is 25%, the threshold can be passed.), the next step is performed. If the result of the discrimination is "No" (assumed to be 18%), the signal is rejected.

Step C (63): Perform a time series analysis on the signal, and repeat the comparison with the events occurring in the unit time. If the values are the same, the asthma is determined by asthma > sleep breathing suspension > cough.

Step D (64): After the discrimination is completed, the signal is stored or displayed for reference.

The advantages and effects of the present invention can be summarized as follows:

[1] The use of features for fuzzy logic to identify asthma events, sleep breathing abortion events and coughing events is quite accurate. The invention pre-measures a set of various maximum features for a phenomenon of an asthmatic event (wrying sound), a sleep breathing abort event and a coughing event when the human body sleeps or wakes up, respectively, for presetting the identification of an asthmatic event and a sleep breathing, respectively. The three fuzzy logic systems of the suspension event and the coughing event, when the lung sound signal, the abdominal respiratory motion signal and the wrist activity signal of any test subject are respectively input into the three fuzzy logic systems, the estimation and identification device can be matched. Accurately discriminating whether the subject is asleep or awake, an asthmatic event, a sleep-disordering event, and a coughing event occur.

[2] The number and timing of asthmatic events, sleep-disordering events, and coughing events can be identified. The invention can not only accurately identify the asthma event, the sleep breathing abort event and the cough event by three fuzzy logic systems, but also can determine the number and time of occurrence by the estimation and identification device; the use for controlling disease identification is more accurate.

[3] The wireless sensing design allows for long-term detection both statically and dynamically. The invention is designed for wireless sensing, which is convenient for the person to be tested to carry (dynamic) detection with the body (long-time detection), and does not affect the action in the daily work and rest, if used for the person to be tested lying on the bed (static), Then let the person to be tested can easily turn over without touching the sensor line. Therefore, the wireless design is also conducive to long-term detection under static and dynamic conditions.

The present invention has been described in detail with reference to the preferred embodiments of the present invention, without departing from the spirit and scope of the invention.

11. . . Preparation step

12. . . Human detection step

13. . . Signal processing step

14. . . Signal identification step

15. . . Complete the steps

20. . . Sound sensing device

20A. . . Lung sound signal

20A(f). . . Frequency domain signal

twenty one. . . First fixed part

twenty two. . . First housing

twenty three. . . Sound sensing component

twenty four. . . First memory

25. . . First microprocessor

26. . . First radio frequency component

27. . . First battery

30. . . Abdominal sensing device

30A. . . Abdominal respiratory motion signal

31. . . Second fixed part

32. . . Second housing

33. . . First acceleration sensor

34. . . Second memory

35. . . Second microprocessor

36. . . Second radio frequency component

37. . . Second battery

40. . . Wrist sensing device

40A. . . Wrist activity signal

41. . . Wrist strap

42. . . Third housing

43. . . Second acceleration sensor

44. . . Third memory

441. . . Multiplexer

45. . . Third microprocessor

46. . . Third wireless RF component

47. . . Third battery

48. . . Button set

481. . . Selection button

482. . . Enter

483. . . Power button

49. . . Screen

50. . . Detection and identification device

50A. . . Signal processing module

501. . . Fourier transform unit

502. . . Nonlinear energy unit

51. . . First fuzzy logic system

52. . . Second fuzzy logic system

53. . . Third fuzzy logic system

54. . . Estimation identification device

55. . . Data storage department

61. . . Step A

62. . . Step B

63. . . Step C

64. . . Step D

91. . . Subject to be tested

T1. . . First identification time

T2. . . Second identification time

T31. . . Expiratory phase time

T32. . . Third identification time

L1. . . Asthma identification line segment

L2. . . Sleep breathing abort identification line segment

The first figure is a flow chart of the method of the present invention

The second figure is a schematic view of the device of the present invention

The third figure is a system block diagram of the device of the present invention.

The fourth figure is a fuzzy attribution function diagram of the signal cutting frequency after the fast Fourier transform of the lung sound signal of the present invention

The fifth figure is the event retention time (ms) fuzzy attribution function graph of the lung sound signal of the present invention after signal cutting

The sixth figure is the event retention time (s) fuzzy attribution function graph of the lung sound signal of the present invention after signal cutting

The seventh figure is the fuzzy time-dependent function of the event maintenance time after the abdominal respiratory motion signal of the present invention is cut by the signal

The eighth, eighth, eighth, eighth, and eighth D drawings are respectively waveforms of the lung sound signal and the identification time after the signal cutting process of the present invention are consistent with the waveform identified as a cough event.

The ninth graph is a waveform diagram of the six asthmatic events in the present invention.

The tenth A and Xth B diagrams are waveform diagrams of a sleep breathing abort event in the discriminating test subject of the present invention, respectively.

The eleventh A and eleventh B diagrams are waveform diagrams for discriminating four times of coughing events in the present invention.

The twelfth A and twelfth B diagrams are waveform diagrams for determining the normal breathing of the subject to be tested according to the present invention, respectively.

The thirteenth, fourteenth and fifteenth drawings are schematic diagrams of the fuzzy logic function of the degree of sludge, the degree of oil stain and the cleaning time of the present invention, respectively.

The sixteenth and seventeenth figures are schematic diagrams showing the amplitude and attribution of the lung sound signal for the identification of asthmatic events according to the present invention.

The eighteenth and nineteenth figures are schematic diagrams showing the amplitude and attribution of the abdominal respiratory motion signal applied to the sleep breathing stop event of the present invention, respectively.

The twentieth and twenty-first figures are schematic diagrams showing the amplitude and attribution of the lung sound signal applied to the cough event of the present invention, respectively.

The twenty-second figure is a system block diagram of the estimation and identification device of the present invention

The twenty-third figure is the action flow chart of the twenty-second figure

11. . . Preparation step

12. . . Human detection step

13. . . Signal processing step

14. . . Signal identification step

15. . . Complete the steps

Claims (8)

A method for identifying characteristics of sleep breathing, coughing, and wheezing includes the following steps: The preparation step is: preparing a sound sensing device, an abdominal sensing device, a wrist sensing device and a detecting and identifying device, wherein the detecting and identifying device has a first fuzzy logic system, a second fuzzy logic system, a third fuzzy logic system and an estimation and identification device; The human body detecting step is configured to detect a person to be tested and generate a lung sound signal and transmit the signal to the detecting and identifying device; the abdominal sensing device is configured to detect the test subject immediately The abdominal breathing process generates a abdominal respiratory motion signal and transmits it to the detection and identification device; the wrist sensing device is configured to instantly detect the test subject to generate a wrist activity signal and transmit the motion signal to the detective Measuring and identifying device; three. Signal processing step: the detection and recognition device first performs fast Fourier transform on the lung sound signal to obtain a frequency, and performs signal cutting processing on the lung sound signal and the abdominal respiratory motion signal; The signal recognition step: the detection and recognition device performs the following three identification operations using the lung sound signal, the abdominal respiratory motion signal, and the wrist activity signal: [a] asthma recognition: the detection recognition device uses the first fuzzy logic system Identification is performed; and the asthma attack is characterized by wheezing; when the frequency of the lung sound signal reaches a frequency standard value, the system is between 200 Hz and 600 Hz, and the frequency standard value continues for a first identification time. Between 250 milliseconds and 500 milliseconds; it is recognized as wheezing and is defined as an asthmatic event in the subject; [b] sleep breathing abort identification: the detection and recognition device is identified by the second fuzzy logic system; The adjacent lung sound signal interval continues for a second identification time; the system is between 10 seconds and 90 seconds; and the adjacent abdominal respiratory motion signal is equally spaced for the second identification time; and the wrist is The activity signal determines that the subject is in a sleep state; it is defined as a sleep apnea event of the subject; [c] cough recognition: the detection and recognition device is identified by the third fuzzy logic system; The expiratory phase of each lung sound signal continues for an expiratory phase time, which is between 0.3 seconds and 1 second; and the adjacent lung sound signals are spaced apart from each other by a third identification time, which is between 3 seconds. Between ~6 seconds; it is defined as the occurrence of a coughing event in the subject; finally, the estimation and identification device is in the order of the asthmatic event, the sleep breathing abort event, and the coughing event, in the first, the second and the third During the overlap time of the identification time, the value is defuzzified to identify the event with the highest probability of occurrence, and the event is determined to be the event of the test subject; Complete the steps: complete the identification of asthmatic events, sleep breathing suspension events and coughing events. A method for identifying characteristics of sleep apnea, cough, and wheezing as described in claim 1, wherein: the frequency standard value of the lung sound signal is 400 Hz as an optimum value; and the first identification time is between 100 Between milliseconds and 500 milliseconds; the second identification time is between 10 seconds and 120 seconds; the exhalation phase time is between 0.12 seconds and 2.5 seconds; the third identification time is between 2 seconds~ Between 10 seconds. The method for identifying the characteristics of sleep apnea, cough and wheezing as described in claim 1, wherein the sound sensing device comprises: a first fixing portion, a first housing, and an acoustic sensing component a first memory, a first microprocessor, a first radio frequency component, and a first battery; the first fixing portion and the first housing are combined with each other, and the first fixing portion is used for Fixing the first housing to the chest of the test subject; the sound sensing component is configured to sense the lung sound signal of the test subject; the first memory is used to memorize and transmit the lung sound signal to The first microprocessor is configured to control the first radio frequency component to transmit the lung audio signal; the first battery system supplies power required by the related component; the abdominal sensing device includes a second fixing portion, a second housing, a first acceleration sensor, a second memory, a second microprocessor, a second radio frequency component, and a second battery; the second fixing The second housing and the second housing are combined with each other, and The second fixing portion is configured to fix the second housing to the abdomen of the subject; the first acceleration sensor is configured to sense the undulating state of the abdomen of the subject in the breathing process, and generate a voltage change. The voltage change can be converted into the abdominal respiratory motion signal; the second memory is used to memorize and transmit the abdominal respiratory motion signal to the second microprocessor, and the second microprocessor is configured to control the second wireless RF component to be The abdominal respiratory motion signal is transmitted; the second battery is supplied with power required by the related component; the wrist sensing device includes: a wrist fixing strap, a third housing, and a second acceleration sensor a third memory, a third microprocessor, a third radio frequency component, a third battery, a button set, and a screen; the wrist fixing strap is used to fix the third housing to be tested The second acceleration sensor is configured to sense a change in the muscle tension of the wrist of the subject; the voltage change is converted into the wrist activity signal, and the third memory is used to memorize the wrist activity signal. And set up a multiplex Transmitting the received lung sound signal, the abdominal respiratory motion signal and the wrist activity signal to the third microprocessor, the third microprocessor is configured to control the third wireless RF component, and the lung sound signal The abdominal respiratory motion signal and the wrist activity signal are transmitted; the third battery is supplied with power required by the related component; the button group includes a selection button, a confirmation button and a power button; the selection button is used for Selecting the function to be performed; the confirmation button is used to confirm the selected function; the power button is used to activate and close the wrist sensing device; the screen is used to display the signal of the wrist sensing device or the related detection result . The method for identifying characteristics of sleep apnea, cough, and wheezing as described in claim 1, wherein the detection and recognition device further includes: a signal processing module for receiving the lung from the multiplexer The audio signal, the abdominal respiratory motion signal and the wrist activity signal; the signal processing module is provided with: a fast Fourier transform unit for transforming the lung sound signal into a frequency domain signal by Fourier transform; a nonlinear energy operation The unit is configured to perform signal cutting on the lung sound signal, the abdominal respiratory motion signal, the wrist activity signal and the frequency domain signal, and then transmit the signal to the first, second and third fuzzy logic systems for calculation a data storage unit for storing the signals and results processed by the estimation and identification device. A device for recognizing the characteristics of sleep breathing, coughing, and wheezing includes: a sound sensing device for detecting a patient to be detected to generate a lung sound signal and transmitting the signal to the detecting and identifying device; The abdominal sensing device is configured to instantly detect the abdominal undulation process of the test subject, and generate a abdominal respiratory motion signal and transmit the same to the detection and identification device; a wrist sensing device is used for instant detection Measuring the subject to generate a wrist activity signal and transmitting the signal to the detection and identification device; a detection and identification device having a first fuzzy logic system, a second fuzzy logic system, and a third fuzzy logic system And an estimation device for receiving the lung sound signal, the abdominal respiratory motion signal and the wrist activity signal; performing fast Fourier transform on the lung sound signal to obtain a frequency, and the lung The signal signal and the abdominal respiratory motion signal are subjected to signal cutting processing; thereby, the detection and identification device cooperates with the lung sound signal, the abdominal respiratory motion signal and the wrist activity signal Three types of identification operations are listed: [a] Asthma identification: the detection and recognition device is identified by the first fuzzy logic system; and the asthma attack is characterized by wheezing; when the frequency of the lung sound signal reaches a frequency standard value, the system Between 200 Hz and 600 Hz, and the standard value of the frequency continues for a first identification time, which is between 250 milliseconds and 500 milliseconds; it is recognized as wheezing and is defined as an asthmatic event of the test subject; b] sleep breathing abort identification: the detection and recognition device is identified by the second fuzzy logic system; when the adjacent lung sound signal interval continues for a second identification time; the system is between 10 seconds and 90 seconds; And the adjacent abdominal respiratory motion signal continues to be separated by the second identification time; and the wrist activity signal determines that the test subject is in a sleep state; and is defined as a sleep breathing suspension event of the test subject; [c] Cough recognition: the detection and recognition device is identified by the third fuzzy logic system; when the expiratory phase of each lung sound signal continues for an expiratory phase time, the system is between 0.3 seconds and 1 second; and the Adjacent lung audio The third identification time is separated from each other by 3 seconds to 6 seconds; it is defined as the occurrence of a coughing event in the test subject; finally, the estimated identification device is based on an asthmatic event, a sleep breathing stop event, and a coughing event. In the sequence, during the overlapping time of the first, the second, and the third identification time, the deblurring value is compared, and the event with the highest probability of occurrence is identified, and the event is determined to be an event of the test subject. The device for identifying the characteristics of sleep apnea, cough and asthma as described in claim 5, wherein: the frequency standard value of the lung sound signal is 400 Hz as an optimum value; the first identification time is between 100 Between milliseconds and 500 milliseconds; the second identification time is between 10 seconds and 120 seconds; the exhalation phase time is between 0.12 seconds and 2.5 seconds; the third identification time is between 2 seconds~ Between 10 seconds. The device for identifying the characteristics of sleep apnea, cough and asthma as described in claim 5, wherein the sound sensing device comprises: a first fixing portion, a first housing, and an acoustic sensing component a first memory, a first microprocessor, a first radio frequency component, and a first battery; the first fixing portion and the first housing are combined with each other, and the first fixing portion is used for Fixing the first housing to the chest of the test subject; the sound sensing component is configured to sense the lung sound signal of the test subject; the first memory is used to memorize and transmit the lung sound signal to The first microprocessor is configured to control the first radio frequency component to transmit the lung audio signal; the first battery system supplies power required by the related component; the abdominal sensing device includes a second fixing portion, a second housing, a first acceleration sensor, a second memory, a second microprocessor, a second radio frequency component, and a second battery; the second fixing The second housing and the second housing are combined with each other, and The second fixing portion is configured to fix the second housing to the abdomen of the subject; the first acceleration sensor is configured to sense the undulating state of the abdomen of the subject in the breathing process, and generate a voltage change. The voltage change can be converted into the abdominal respiratory motion signal; the second memory is used to memorize and transmit the abdominal respiratory motion signal to the second microprocessor, and the second microprocessor is configured to control the second wireless RF component to be The abdominal respiratory motion signal is transmitted; the second battery is supplied with power required by the related component; the wrist sensing device includes: a wrist fixing strap, a third housing, and a second acceleration sensor a third memory, a third microprocessor, a third radio frequency component, a third battery, a button set, and a screen; the wrist fixing strap is used to fix the third housing to be tested The second acceleration sensor is configured to sense a change in the muscle tension of the wrist of the subject; the voltage change is converted into the wrist activity signal, and the third memory is used to memorize the wrist activity signal. And set up a multiplex Transmitting the received lung sound signal, the abdominal respiratory motion signal and the wrist activity signal to the third microprocessor, the third microprocessor is configured to control the third wireless RF component, and the lung sound signal The abdominal respiratory motion signal and the wrist activity signal are transmitted; the third battery is supplied with power required by the related component; the button group includes a selection button, a confirmation button and a power button; the selection button is used for Selecting the function to be performed; the confirmation button is used to confirm the selected function; the power button is used to activate and close the wrist sensing device; the screen is used to display the signal of the wrist sensing device or the related detection result . The device for identifying the characteristics of sleep apnea, cough and wheezing as described in claim 5, wherein the detection and recognition device further comprises: a signal processing module for receiving the lung from the multiplexer The audio signal, the abdominal respiratory motion signal and the wrist activity signal; the signal processing module is provided with: a fast Fourier transform unit for transforming the lung sound signal into a frequency domain signal by Fourier transform; a nonlinear energy operation The unit is configured to perform signal cutting on the lung sound signal, the abdominal respiratory motion signal, the wrist activity signal and the frequency domain signal, and then transmit the signal to the first, second and third fuzzy logic systems for calculation a data storage unit for storing the signals and results processed by the estimation and identification device.