Sleep apnea detection method and system based on microwave radar and microphone

Abstract

The invention relates to a sleep apnea detection method and system based on a microwave radar and a microphone, and the method comprises the following steps: a radar data processing step: S11, transmitting and receiving a radar signal through the microwave radar, and searching the position of a human body; s12, the microwave radar is controlled to rotate and / or translate, and a respiration detection area is found; s13, receiving a radar signal returned by respiration detection, and extracting and recording a respiration waveform from the radar signal; an audio data processing step: S21, capturing audio data through a microphone, identifying whether the audio data accords with the snore characteristics, if so, determining that the audio data contains the snore, and recording the snore; and a judgment step: if the sound feature corresponding to the sleep apnea exists in the snore and the respiration feature corresponding to the sleep apnea exists in the respiration waveform, judging that the sleep apnea is detected.

Description

Technical Field

[0001] The present invention relates to the field of OSAHS detection, and in particular to a sleep apnea detection method and system based on microwave radar and microphone. Background Art

[0002] Sleep Apnea Hypopnea Syndrome (SAHS) refers to a clinical syndrome in which various reasons lead to repeated apnea and / or hypopnea, hypercapnia, and sleep interruption during sleep, resulting in a series of pathophysiological changes in the body.

[0003] In the prior art, for example, there is an existing application with publication number CN111657870A, entitled A sleep apnea judgment system based on DS evidence theory, which combines snoring and breathing conditions identified by radar to determine whether there is sleep apnea. However, this application cannot know the position of the human chest cavity. Breathing corresponds to the tiny ups and downs of the human body. Since the human body may turn over and move during sleep, the position of the human body in the radar scanning area is unknown. The movement of human breathing is very weak. Not all areas within the radar scanning range can extract such weak signals. Before measuring the respiratory rate, the direction and orientation of the microwave radar need to be manually adjusted so that the human body is in an area with better signals within the radar scanning range. Only then can the subsequent respiratory rate detection algorithm capture the respiratory rate data from the echo data. It is easy to misjudge sleep apnea due to the inability to detect the respiratory signal.

[0004] The purpose of the present invention is to design a sleep apnea detection method and system based on microwave radar and microphone to solve the above problems in the prior art. Summary of the invention

[0005] In view of the problems existing in the above-mentioned prior art, the present invention provides a sleep apnea detection method and system based on microwave radar and microphone, which can effectively solve at least one problem existing in the above-mentioned prior art.

[0006] The technical solution of the present invention is:

[0007] Therefore, the present invention provides the following effects and / or advantages:

[0008] The present application combines a microwave radar and a microphone. The microwave radar automatically finds the breathing detection area of ​​the human body and records the radar echo to calculate the breathing waveform. At the same time, the microphone is used to capture the snoring of the human body to technically determine the breathing situation. If the snoring contains the sound characteristics corresponding to sleep apnea, and the breathing characteristics corresponding to sleep apnea are contained in the breathing waveform, it is determined that sleep apnea is detected.

[0009] The microwave radar of the present application can rotate. The microwave radar searches for a respiratory detection area, thereby controlling the microwave radar to rotate and / or translate the center of its scanning area toward the respiratory detection area. The center of the scanning area has the best signal-to-noise ratio and the best response to respiratory signs, thereby improving the detection accuracy of the microwave radar in detecting respiratory waveforms and reducing the misjudgment of apnea caused by a weak respiratory waveform that cannot be captured by the microwave radar.

[0010] The present application can control the rotation and / or translation of the microwave radar to find the breathing detection area, so that the microwave radar used for breathing can be placed in multiple locations indoors, and there is no need to manually adjust the scanning area of ​​the microwave radar before use. The microwave radar can find the breathing detection area and perform vital signs detection by itself.

[0011] The present application calculates the breathing waveform at multiple positions within the area of ​​the micro-motion state, and through the breathing conditions at multiple positions, it is possible to determine whether the scanning center of the microwave radar is facing the part of the human body that reacts to breathing. This can be used to accurately adjust the scanning range of the microwave radar, thereby improving the accuracy of breathing detection.

[0012] The present application can accurately determine the sleep breathing condition of the human body from the snoring by identifying the characteristics of the audio data that match the snoring, analyzing the sound characteristics of the snoring, and determining whether the snoring contains the sound characteristics corresponding to sleep apnea. The sound characteristics can be combined with the breathing waveform captured by the radar data to obtain a more accurate apnea condition.

[0013] The present application analyzes the sleeping posture of the human body and corrects the audio data according to the sleeping posture, so that the audio data under different sleeping postures can be corrected to the audio of the human body lying flat, thereby reducing the pressure on the oral cavity, chest cavity, etc. caused by different sleeping postures of the human body, which causes changes in snoring and thus interferes with the sleep apnea analysis process, thereby further improving the accuracy.

[0014] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description and the drawings.

[0015] It is to be understood that both the foregoing general description and the following detailed description of the present invention are exemplary and explanatory and are intended to provide further explanation of the invention as claimed. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a flow chart provided for an embodiment of the present invention.

[0017] Figure 2This is a structural diagram of a vital signs detection device.

[0018] Figure 3 Schematic diagram for finding the human body position for microwave radar.

[0019] Figure 4 Schematic diagram of finding the micromotion area for microwave radar, where the shaded area represents the micromotion area.

[0020] Figure 5 This is a schematic diagram showing that the center of the scanning area of ​​the microwave radar is located at the center of the micro-motion area after rotation.

[0021] Figure 6 This is a schematic diagram of dividing the micro-motion state area into four square areas arranged from top to bottom when measuring the respiratory frequency.

[0022] Figure 7 A schematic diagram of adjusting the scanning area of ​​a microwave radar according to breathing frequency.

[0023] Figure 8 Schematic diagram of the steps for calculating vital signs. DETAILED DESCRIPTION

[0024] In order to facilitate understanding by those skilled in the art, the present invention is now described in further detail by way of examples:

[0025] refer to Figure 1 , a sleep apnea detection method based on microwave radar and microphone, comprising the following steps:

[0026] Radar data processing steps include:

[0027] S11, transmitting and receiving radar signals through microwave radar to find the position of the human body;

[0028] In this embodiment, a radar signal is transmitted by a microwave radar and a radar signal returned by a human body in the room is received. The microwave radar can be placed in the room, for example, on a bedside table, a wall, a wardrobe, etc., without limitation. The microwave radar can scan multiple positions in the room. At this time, the position of the radar or the position of the human body is unknown. The method of searching for a human body by a microwave radar is an existing technology. For example, when the human body is in a sitting or lying state, the human body does not walk or make large body movements, and the preliminary position of the human body is obtained according to the characteristics of the echo data on the microwave radar in this state.

[0029] In this step, the microwave radar can rotate or translate to search for a human body in the room.

[0030] This step is as follows Figure 3 shown.

[0031] S12, controlling the microwave radar to rotate and / or translate to find a respiratory rate detection area;

[0032] In step S11, the microwave radar finds the position of the human body, and it is necessary to further find the respiratory rate detection area. This step can be based on the situation that the human body is in a resting state such as sitting or lying down, the human body does not move in a large range, the human body's breathing will produce a small ups and downs, the human body's heartbeat will produce a small shake, etc., and according to the performance of this situation in the radar echo data, the respiratory rate detection area can be found.

[0033] Because the signal-to-noise ratio of each area within the scanning range of the microwave radar is different, and the position of the human body is unknown before step S1, if the microwave radar does not look for a respiratory rate detection area but directly extracts the respiratory rate through the microwave radar data returned by the human body's breathing or heartbeat, its data will be affected by external interference and different radar signal-to-noise ratios, and the calculated respiratory rate data will have problems such as insufficient accuracy. Therefore, this step searches for a respiratory rate detection area.

[0034] S13, receiving a radar signal returned by the respiratory frequency detection area, and extracting the respiratory frequency from the radar signal.

[0035] In step S12, the human body respiratory rate detection area is found. At this time, only the radar signal of the respiratory rate detection area is extracted to calculate the respiratory rate, which can greatly improve the accuracy of the respiratory rate calculation result. The method of respiratory rate detection can refer to the existing application with publication number CN118452869A and the name of respiratory rate monitoring method, device, equipment and medium based on millimeter wave radar.

[0036] Alternatively, first construct the ADC data collected by each frame of microwave radar within a period of time as a matrix to be processed, and use the covariance matrix decomposition method to remove the clutter that is significantly different from the useful signal; then use the wavelet denoising method to remove the non-stationary noise, so as to reconstruct the original ADC data. Then, by referring to Figure 2 The provided computational process implements vital signs analysis.

[0037] In step S11, searching for the human body position includes:

[0038] Identify the human body and obtain its position through point cloud data.

[0039] In this step, the microwave radar is used to collect ADC data, perform one-dimensional FFT calculation, two-dimensional FFT calculation, select the target position, and then map the horizontal angle and pitch angle between the human body and the microwave radar and the target to the direct coordinate system, so that the human body can be identified, and the microwave radar is rotated or translated through the horizontal angle and pitch angle between the human body and the microwave radar, so that the human body is located in the middle of the microwave radar scanning area.

[0040] When the human body is in a resting state, the point cloud data corresponding to the reflected radar signal is relatively small. By utilizing the amount of point cloud data in this state, it can be determined that there is a human body at that position when the point cloud data is less than a preset point cloud data amount threshold.

[0041] The audio data processing steps include:

[0042] S21, capturing audio data through a microphone, identifying whether the audio data meets the characteristics of snoring, and if so, determining that the audio data contains snoring, and recording the snoring;

[0043] In this embodiment, identifying whether the audio data meets the characteristics of snoring can be achieved through existing technologies, such as an existing application with publication number CN111657870A, entitled A snoring apnea determination system based on DS evidence theory.

[0044] The judgment steps include:

[0045] S31, analyzing the sound characteristics of the snoring sound to determine whether the snoring sound contains a sound characteristic corresponding to sleep apnea;

[0046] S32, analyzing the respiratory waveform to determine whether there is a respiratory feature corresponding to sleep apnea in the respiratory waveform;

[0047] S33: If the snoring sound contains a sound feature corresponding to sleep apnea, and the respiratory waveform contains a respiratory feature corresponding to sleep apnea, it is determined that sleep apnea is detected.

[0048] This embodiment combines a microwave radar and a microphone. The microwave radar identifies the sleeping breathing frequency and thus can detect apnea. At the same time, the sound data recorded by the microphone identifies apnea. Only when both occur multiple times is it determined that sleep apnea exists.

[0049] The specific method of the judgment step will be described in detail later.

[0050] Further, step S12 includes:

[0051] S121, determining whether the human body is in a micro-motion state, and if so, rotating and / or translating the microwave radar toward a breathing detection area;

[0052] In this step, the radar signal can be calculated by the micro-motion algorithm to determine whether the human body is in a micro-motion state. The parts of the human body that are in a micro-motion state are generally the chest and abdomen. This area is used to identify the best area for respiratory rate detection. The micro-motion algorithm can be used to extract the parts of the micro-motion state, such as Figure 4 As shown, in step S11, the microwave radar finds the position of the human body. At this time, the microwave radar can be rotated or translated to face the human body. In this step, by extracting the micro-motion area, the microwave radar can be rotated or translated to face the center of the scanning area toward the abdominal part of the human body, as shown in FIG. Figure 5 shown.

[0053] Since the accuracy of the micro-motion algorithm is average, in this step, when the microwave radar rotates or translates toward the micro-motion area, the chest or abdomen of the human body may not be in the optimal scanning area of ​​the microwave radar.

[0054] S122, calculating the respiratory frequency at multiple locations within the respiratory detection area;

[0055] In this step, the heartbeat is used as an example to illustrate that the human chest cavity produces a slight vibration under the action of the heartbeat, thereby generating a breathing detection area.

[0056] Theoretically, the frequencies corresponding to the tiny vibrations generated by the human heartbeat at various locations on the chest should be the same, and correspondingly, the heartbeat data calculated from the radar signal reflected back by the tiny vibrations generated by the heartbeat should be the same at multiple locations within the breathing detection area. The breathing frequencies are calculated at multiple locations within the breathing detection area obtained in step S121, so as to be used in the subsequent steps to determine whether the best signal-to-noise ratio scanning area of ​​the microwave radar is directly corresponding to the human chest.

[0057] S123, if the breathing frequencies at multiple positions are similar, the breathing detection area is determined to be the breathing frequency detection area; if the breathing frequencies at multiple positions are different, the scanning range of the microwave radar is rotated and / or translated and the process returns to step S121.

[0058] If the respiratory rate calculations at multiple locations are the same or similar, it means that the best detection area of ​​the microwave radar is right for the slight respiratory rate movement of the human chest or abdomen. The best detection area of ​​the microwave radar can be considered to be the center of the radar detection area.

[0059] Due to different sleeping postures of the human body, the radar signal of the micro-motion area captured by the microwave radar may be too weak, interfered, or have calculation errors, or the micro-motion signal may be too weak to be captured when it is transmitted from the center of the chest or abdomen to the edge of the human body, resulting in different calculated respiratory rates at multiple locations. Therefore, this step needs to be performed.

[0060] Preferably, for the respiratory frequency in the respiratory frequency, since the abdomen of the human body will float up and down during breathing, at this time, the respiratory detection area can be divided into N rectangles arranged up and down, for example, divided into 4 rectangles, such as Figure 7 As shown, in this area, the breathing frequency is detected in each rectangular area. If the center of the abdomen is located at the center of N rectangles, the micro-movements generated by the ups and downs of the abdomen in each rectangular area are captured by the microwave radar and the breathing frequency is calculated to be very close.

[0061] At the same time, when measuring the respiratory frequency, since the abdomen is located in the middle of the human body, in order to facilitate capturing the respiratory frequency, in this step, the area of ​​the respiratory detection area close to the human legs can be first divided into N rectangles and then the respiratory frequency can be calculated.

[0062] The legs or head of a human body can be identified by existing technologies. The longitudinal direction refers to the length direction of the human body.

[0063] Therefore, through this step, it is found whether the center of the scanning range of the microwave radar is located at the center where the respiratory data can be extracted. At this time, the radar data obtained by the microwave radar can correctly reflect the respiratory frequency. If the respiratory frequencies at multiple positions are different, the microwave radar can be rotated or moved to make the scanning area of ​​the microwave radar translate or rotate, thereby changing the center point of the scanning area of ​​the microwave radar.

[0064] Furthermore, before step S121, execute:

[0065] S120, record and update the radar signal in real time, and enter step S121 when the number of point clouds of the radar signal is less than the preset number threshold, and use the human body position when the number of point clouds of the radar signal is less than the preset number threshold as the micro-motion position.

[0066] Further, step S122 includes:

[0067] The radar signal of the breathing detection area is equally divided longitudinally and the phases are calculated respectively;

[0068] Multiple phases are calculated through a respiratory frequency algorithm to obtain multiple respiratory frequencies.

[0069] In this embodiment, if the human body is stationary, the number of point clouds of the radar signal will become very small. By presetting a quantity threshold, it can be determined that the human body is in a stationary state when the number is reduced to below the preset quantity threshold.

[0070] Further, step S122 includes:

[0071] The breathing detection area is divided into a plurality of sub-areas at equal intervals, and the breathing frequencies of the plurality of sub-areas are calculated.

[0072] Further, dividing the breathing detection area into a plurality of sub-areas at equal intervals, and calculating the breathing frequencies of the plurality of sub-areas comprises:

[0073] The radar signal in the breathing detection area is equally divided and the phases are calculated respectively;

[0074] Multiple phases are calculated through a respiratory frequency algorithm to obtain multiple respiratory frequencies.

[0075] In this step, you can refer to Figure 6 Or 7, thereby dividing the breathing detection area into equal intervals.

[0076] Specifically, the phase data of the micro-motion area can be extracted first, and the phase data can be phase unwrapped to generate unwrapped data; then the unwrapped data can be preprocessed, and the preprocessed unwrapped data can be subjected to FFT operation to generate the respiratory frequency and track it. For details, reference can be made to the existing application with publication number CN118452869A and the name of the respiratory frequency monitoring method, device, equipment and medium based on millimeter wave radar.

[0077] Further, if the breathing frequencies of the multiple positions are different, the scanning range of the microwave radar is rotated and / or translated according to the breathing frequencies of the multiple sub-areas and the process returns to step S121.

[0078] Further, rotating and / or translating the microwave radar according to the breathing frequencies of the plurality of sub-areas comprises:

[0079] Calculate the mean respiratory rate of multiple sub-areas;

[0080] The microwave radar is rotated and / or translated toward the sub-region where the breathing frequency is close to the breathing frequency mean.

[0081] In this step, if Figure 7 As shown, assuming that the respiratory rates from multiple positions on the abdomen calculated in the above steps are 15, 20, 21, and 20 (times / minute), respectively, the calculated average is 19, and the sub-areas with respiratory rates of 20, 21, and 20 are areas where the respiratory rates are close to the respiratory rate average. At this time, the scanning range of the microwave radar needs to be adjusted downward.

[0082] Further, the breathing rates at multiple locations are determined to be similar by the following steps:

[0083] The standard deviation of the respiratory frequencies at the multiple locations is calculated. If the standard deviation is less than a preset standard deviation threshold, the respiratory frequencies at the multiple locations are similar.

[0084] Further, analyzing the respiratory waveform to determine whether there is a respiratory feature corresponding to sleep apnea in the respiratory waveform includes:

[0085] Calculate the Lyapunov index of the respiratory waveform. If the Lyapunov index of the respiratory waveform is abnormal, proceed to the next step.

[0086] The number of times the respiratory waveform deviates from the respiratory baseline and the number of times the respiratory waveform has a respiratory count of 0 are counted; if the sum of the number of times the respiratory waveform deviates from the respiratory baseline and the number of times the respiratory waveform has a respiratory count of 0 is greater than a preset number threshold, it is determined that the respiratory waveform contains a respiratory feature corresponding to sleep apnea.

[0087] This step is one of the core steps of this embodiment. The Lyapunov index is used to represent the numerical characteristics of the average exponential divergence rate of adjacent trajectories in the phase space. If the Lyapunov index of the respiratory waveform is abnormal, it can be considered that the respiratory waveform is abnormal. If the Lyapunov index is greater than 0, it means that the adjacent trajectories of the respiratory waveform will have an average exponential divergence over time, which indicates that the system exhibits chaotic behavior. On the contrary, if the Lyapunov index is less than 0, the respiratory waveform will tend to be stable and insensitive to the initial state of the respiratory waveform, that is, the mapping is insensitive to the initial value; when the index is equal to 0, the respiratory waveform is in a stable state. Therefore, by calculating and analyzing the Lyapunov index, it is possible to intuitively determine whether the respiratory waveform is in a chaotic state.

[0088] The breathing number of 0 can be obtained from the breathing waveform. When the breathing waveform is briefly paused, it can be determined that the breathing number is 0. The deviation of the breathing waveform from the breathing reference indicates that the breathing is abnormal.

[0089] The number of both is counted, and when the total number of both exceeds a preset number threshold, it can be considered that the breathing characteristics corresponding to sleep apnea exist.

[0090] Specifically, in the above, the respiratory waveform is extracted and calculated from the radar signal. The respiratory waveform of a normal human body fluctuates up and down with an amplitude of 0. In this embodiment, the amplitude of 0 is set as the benchmark, the respiratory waveform is averaged, and a benchmark deviation threshold is set. If the mean of the respiratory waveform is greater than the benchmark deviation threshold, it is determined to be abnormal breathing.

[0091] Further, the audio data is identified to have the characteristics of snoring through the following steps:

[0092] Extract the frequency characteristics of the audio data, calculate the frequency weight, and if the frequency weight meets the preset weight threshold, proceed to the next step;

[0093] Analyze the volume of the audio data, and if the volume meets a preset volume threshold, proceed to the next step;

[0094] The time domain features of the audio data are analyzed, and if the time domain features meet a preset time domain threshold, it is determined that snoring is recognized.

[0095] Specifically, the identification of snoring includes the comprehensive processing of the frequency domain and time domain of audio data. With a certain amount of captured audio data, if the time domain characteristics and volume requirements of snoring are met, the frequency characteristics of the data will be extracted and saved, and a frequency feature library will be gradually formed. The frequency characteristics are dynamically weighted. If the frequency weight meets the preset weight threshold, these frequency characteristics are valid characteristics. When the detected frequency characteristics of the input data meet the requirements of the valid frequency characteristics in the library, and the time domain and volume also meet the requirements, it will be identified as snoring.

[0096] Further, analyzing the sound features of the snoring sound and determining whether there is a sound feature corresponding to sleep apnea in the snoring sound includes:

[0097] If the time domain period of the snoring sound is greater than a preset time domain period of the snoring sound, and the frequency of the snoring sound is lower than a preset frequency threshold, then calculating the frequency domain data energy of the snoring sound;

[0098] If the frequency domain data energy is greater than a preset energy threshold, filtering the snoring sound through a Mel filter bank and calculating the data energy of the filtered snoring sound;

[0099] Calculate the Mel-cepstral coefficients of the data energy, classify the snoring sound according to the Mel-cepstral coefficients through a classifier, count the number of apnea low-pass indexes, and if the number of apnea low-pass indexes is greater than a preset apnea low-pass index number threshold, determine that the snoring sound contains sound features corresponding to sleep apnea.

[0100] Specifically, the frequency domain data is passed through a Mel filter bank. The design of the Mel filter bank is based on the fact that the human auditory system perceives different frequencies nonlinearly. After passing through the Mel filter bank, the energy distribution on the Mel frequency scale is obtained. The output of each filter is logarithmic, and then discrete cosine transform is performed to obtain the Mel cepstral coefficients.

[0101] Further, in S11, after finding the position of the human body, executing: analyzing the sleeping posture of the human body;

[0102] In S21, after the audio data is captured by the microphone, the following is performed: the audio data is corrected according to the sleeping posture, so that the audio data in the sleeping posture is corrected to the audio data of the human body lying flat.

[0103] In this step, the radar is used to identify the sleeping posture of the human body, which is an existing technology. It can identify postures such as lying flat, lying on the side, and lying prone. Due to different sleeping postures of the human body, different pressures will be exerted on the chest and oral cavity. For example, when the human body lies flat, the pressure on the chest and oral cavity is small, while when the human body sleeps on the prone, the pressure on the chest and oral cavity is large. The different pressures on the chest and oral cavity will cause changes in the frequency and tone of snoring, which will interfere with the subsequent process of identifying whether the audio data conforms to the characteristics of snoring and judging whether there is a sound feature corresponding to sleep apnea in the snoring. Therefore, this step is added for processing. In this embodiment, different human sleeping postures, radar data, and snoring can be input into a neural network for training, and a neural network associated with human sleeping postures, radar data, and snoring can be obtained. The neural network is used to correct snoring, thereby eliminating the influence of different sleeping postures on snoring, and finally correcting the snoring of the human body in different sleeping postures to the snoring of the human body in a lying state. Then, the snoring in the lying state is used for the subsequent analysis and judgment steps, so that more accurate sound features and sleep apnea can be obtained.

[0104] Further providing a sleep apnea detection system based on microwave radar and microphone, referring to Figure 8 , including a microwave radar 1 and a microphone 2. The microwave radar 1 and / or the microphone 2 can be rotatably arranged. When the sleep apnea detection system is working, the sleep apnea detection method based on the microwave radar and the microphone is implemented.

[0105] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0106] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1A device that provides the functions specified in a block or multiple blocks.

[0107] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0108] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0109] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms should not be understood as necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, unless they are contradictory.

Claims

1. A sleep apnea detection method based on microwave radar and microphone, characterized in that: The following steps are involved: Radar data processing steps include: S11, transmitting and receiving radar signals through microwave radar to find the position of the human body; S12, controlling the microwave radar to rotate and / or translate to find a breathing detection area; S13, receiving the radar signal returned by the breathing detection, extracting and recording the breathing waveform from the radar signal; The audio data processing steps include: S21, capturing audio data through a microphone, identifying whether the audio data meets the characteristics of snoring, and if so, determining that the audio data contains snoring, and recording the snoring; The judgment steps include: S31, analyzing the sound characteristics of the snoring sound to determine whether the snoring sound contains a sound characteristic corresponding to sleep apnea; S32, analyzing the respiratory waveform to determine whether there is a respiratory feature corresponding to sleep apnea in the respiratory waveform; S33: If the snoring sound contains a sound feature corresponding to sleep apnea, and the respiratory waveform contains a respiratory feature corresponding to sleep apnea, it is determined that sleep apnea is detected.

2. The sleep apnea detection method based on microwave radar and microphone according to claim 1, characterized in that: Step S12 includes: S121, determining whether the human body is in a micro-motion state, and if so, rotating and / or translating the microwave radar toward a breathing detection area; S122, calculating the respiratory frequency at multiple locations within the respiratory detection area; S123, if the breathing frequencies at multiple positions are similar, the breathing detection area is determined to be the breathing frequency detection area; if the breathing frequencies at multiple positions are different, the scanning range of the microwave radar is rotated and / or translated and the process returns to step S121.

3. The sleep apnea detection method based on microwave radar and microphone according to claim 2, characterized in that: Before step S121, execute: S120, record and update the radar signal in real time, and enter step S121 when the number of point clouds of the radar signal is less than the preset number threshold, and use the human body position when the number of point clouds of the radar signal is less than the preset number threshold as the micro-motion position.

4. The sleep apnea detection method based on microwave radar and microphone according to claim 3, characterized in that: Step S122 includes: The radar signal of the breathing detection area is equally divided longitudinally and the phases are calculated respectively; Multiple phases are calculated through a respiratory frequency algorithm to obtain multiple respiratory frequencies.

5. The sleep apnea detection method based on microwave radar and microphone according to claim 4, characterized in that: If the respiratory rates at multiple locations are different, perform: Calculate the mean respiratory rate of multiple sub-areas; The microwave radar is rotated and / or translated toward the sub-region where the breathing frequency is close to the breathing frequency mean.

6. The sleep apnea detection method based on microwave radar and microphone according to claim 1, characterized in that: Analyzing the respiratory waveform to determine whether there is a respiratory feature corresponding to sleep apnea in the respiratory waveform includes: Calculate the Lyapunov index of the respiratory waveform. If the Lyapunov index of the respiratory waveform is abnormal, proceed to the next step. The number of times the respiratory waveform deviates from the respiratory baseline and the number of times the respiratory waveform has a respiratory count of 0 are counted; if the sum of the number of times the respiratory waveform deviates from the respiratory baseline and the number of times the respiratory waveform has a respiratory count of 0 is greater than a preset number threshold, it is determined that the respiratory waveform contains a respiratory feature corresponding to sleep apnea.

7. The sleep apnea detection method based on microwave radar and microphone according to claim 1, characterized in that: Identify the characteristics of the audio data that match snoring by following these steps: Extract the frequency characteristics of the audio data, calculate the frequency weight, and if the frequency weight meets the preset weight threshold, proceed to the next step; Analyze the volume of the audio data, and if the volume meets a preset volume threshold, proceed to the next step; The time domain features of the audio data are analyzed, and if the time domain features meet a preset time domain threshold, it is determined that snoring is recognized.

8. The sleep apnea detection method based on microwave radar and microphone according to claim 1, characterized in that: Analyzing the sound features of the snoring sound and determining whether there is a sound feature corresponding to sleep apnea in the snoring sound includes: If the time domain period of the snoring sound is greater than a preset time domain period of the snoring sound, and the frequency of the snoring sound is lower than a preset frequency threshold, then calculating the frequency domain data energy of the snoring sound; If the frequency domain data energy is greater than a preset energy threshold, filtering the snoring sound through a Mel filter bank and calculating the data energy of the filtered snoring sound; Calculate the Mel-cepstral coefficients of the data energy, classify the snoring sound according to the Mel-cepstral coefficients through a classifier, count the number of apnea low-pass indexes, and if the number of apnea low-pass indexes is greater than a preset apnea low-pass index number threshold, determine that the snoring sound contains sound features corresponding to sleep apnea.

9. A sleep apnea detection method based on microwave radar and microphone according to claim 1 or 8, characterized in that: In S11, after finding the position of the human body, the following steps are performed: analyzing the sleeping posture of the human body; In S21, after the audio data is captured by the microphone, the following is performed: the audio data is corrected according to the sleeping posture, so that the audio data in the sleeping posture is corrected to the audio data of the human body lying flat.

10. A sleep apnea detection system based on microwave radar and microphone, characterized in that: The sleep apnea detection system comprises a microwave radar and a microphone. The microwave radar and / or the microphone can be rotatably arranged. When the sleep apnea detection system is in operation, the sleep apnea detection method based on a microwave radar and a microphone according to any one of claims 1 to 9 is implemented.

Images