Methods and Systems for Monitoring Vital Signs and Sleep Based on Millimeter-Wave Radar

Abstract

This invention discloses a method and system for monitoring vital signs and sleep based on millimeter-wave radar. This method utilizes millimeter-wave radar to achieve non-contact sleep monitoring. First, it uses millimeter-wave radar to collect radar echo signals obtained from chest wall vibrations during sleep. Then, it determines the location of the monitored object and extracts the target's vital sign signals. Next, it accurately detects the first and second amplitude body movements of the target. It also detects whether the target experiences apnea in real time and issues an alarm if the event lasts for a prolonged period. It detects three sleeping positions: supine, lateral, and prone. Finally, it generates sleep state monitoring data and combines the previously monitored information to create a user sleep quality analysis report, providing health recommendations. This method is simple, accurate, cost-effective, and suitable for various application scenarios such as homes, hospitals, and nursing homes, providing users with 24 / 7 health monitoring services.

Description

Technical Field

[0001] This invention relates to the fields of biomedical engineering and wireless sensing equipment technology, and in particular to a method and system for non-contact monitoring of human vital signs, sleep posture, sleep stages, and sleep apnea alarms using millimeter-wave radar. Background Technology

[0002] In recent years, with the advancement of millimeter-wave radar technology, its application in non-contact human body monitoring has received widespread attention. Due to its high-frequency operation, millimeter-wave radar has the ability to penetrate non-metallic materials and can detect weak vital signs such as breathing and heartbeat within a certain distance. The advantage of this technology is that it does not require direct contact with the human body, thus avoiding the inconvenience and discomfort associated with traditional monitoring methods. Traditional monitoring devices, such as electrocardiograms (ECGs) and pulse oximeters, while providing accurate vital sign data, may cause discomfort to users during prolonged wear, especially affecting sleep quality during nighttime sleep monitoring.

[0003] Meanwhile, with increasing health awareness and growing demand for high-quality sleep, non-contact health monitoring technologies have gradually become a research hotspot. Millimeter-wave radar, as an emerging non-contact monitoring tool, can not only monitor basic vital signs such as respiratory rate and heart rate, but also further identify sleep postures, divide sleep stages, and monitor sleep disorder symptoms such as sleep apnea. Compared to other non-contact technologies, such as infrared sensors and ultrasonic sensors, millimeter-wave radar has higher resolution and stronger anti-interference capabilities, enabling it to operate stably in complex environments.

[0004] Therefore, there is a need for a method and system for monitoring vital signs and sleep based on millimeter-wave radar. Summary of the Invention

[0005] In view of this, the purpose of the present invention is to provide a method and system for monitoring vital signs and sleep based on millimeter-wave radar. This method can not only improve the accuracy of vital sign monitoring, but also accurately identify sleep posture, classify sleep stages, and effectively monitor the occurrence of sleep apnea.

[0006] To achieve the above objectives, the present invention provides the following technical solution: The present invention provides a method for monitoring vital signs and sleep based on millimeter-wave radar, comprising the following steps: The radar echo signal obtained by collecting chest wall vibration during sleep using millimeter-wave radar. The location of the monitoring target is determined by a bed area positioning algorithm; The adaptive noise extraction algorithm effectively eliminates interference and extracts the target's vital signs signals. The first and second amplitude body movements of the target are accurately detected using a multi-feature body movement detection algorithm. Extract breathing and heartbeat signals, use a multi-feature apnea judgment algorithm to detect whether the target has experienced an apnea event in real time, and issue an alarm message if the event occurs for more than a threshold time. The multi-feature sleeping posture judgment algorithm effectively detects different sleeping postures; The sleep monitoring algorithm monitors and analyzes different sleeping positions, and combines the monitoring information to generate a user's sleep quality analysis report.

[0007] Furthermore, the location of the monitored target is determined according to the following steps: Determine the installation location and bed parameters of the millimeter-wave radar, and determine the monitoring area of ​​the bed; The system detects whether any targets exceeding a threshold appear within the window monitoring area. If a target appears, it continues to detect whether any vital signs are present. If vital signs are detected, it is determined that a monitoring target exists in the bed monitoring area.

[0008] Furthermore, the multi-feature motion detection algorithm is performed according to the following steps: Calculate the second-order derivative of the distance time, the amplitude of the distance image spectrum, and the distance gate diffusion information based on the distance gate data of the target location; When it is detected that: the second derivative information of distance time is less than the first body movement threshold, the amplitude information of distance image spectrum is less than the second body movement threshold, and the distance gate diffusion information is less than the third body movement threshold, it is judged that there is no body movement and the target is in a resting state. When it is detected that: the second derivative information of distance time is less than the first body motion threshold, the amplitude information of distance image spectrum is greater than the second body motion threshold and less than the fourth body motion threshold, and the distance gate diffusion information is greater than the third body motion threshold, it is judged to be the second amplitude body motion, and the target is in the second amplitude body motion state; When the following are detected: the second derivative of distance time is greater than the first body movement threshold, the amplitude of distance image spectrum is greater than the fourth body movement threshold, and the distance gate diffusion information is greater than the third body movement threshold, it is determined to be the first amplitude body movement. The first amplitude body movement information is then input as a sleep feature to the sleep monitoring module, and the data on the time of body movement occurrence is separated from the data entering the vital signs monitoring.

[0009] Furthermore, the extraction of respiratory and heartbeat signals is performed according to the following steps: Phase information is extracted from resting data in which no body movement has occurred, and the original phase information is obtained by phase unwinding; The original phase information is subjected to Fourier transform to obtain phase change period information. The period information in the Fourier transformed phase period information that is higher than the set signal-to-noise ratio is identified as the target respiratory signal period. The original phase information is subjected to IIR filtering to extract the heartbeat signal. The heartbeat signal is then subjected to spectrum detection to calculate the Fourier transform of the heartbeat signal and find the periodic information of the detected heartbeat signal. It calculates real-time heart rate and respiration.

[0010] Furthermore, the multi-feature sleep apnea detection algorithm is performed according to the following steps: When apnea occurs, the intensity, standard score, and range of the respiratory signal are used simultaneously as characteristics to determine apnea, as detailed below: The intensity feature of the respiratory signal is valid when the intensity of the respiratory signal is less than the first respiratory feature threshold. The standard score feature of the respiratory signal is valid when the standard score of the respiratory signal is less than the second respiratory feature threshold. The range feature of the respiratory signal is effective when the range of the respiratory signal is less than the third respiratory feature threshold. When the three characteristics are valid for a period of time that exceeds the set apnea judgment time, an apnea event is confirmed to have occurred.

[0011] Furthermore, the multi-feature sleeping posture determination algorithm is performed according to the following steps: Sleeping posture is determined based on one-dimensional distance image spectrum, two-dimensional distance velocity spectrum, heart rate change, and respiratory rate change; the sleeping posture includes three types: supine, lateral, and prone.

[0012] Furthermore, the sleep monitoring algorithm is performed according to the following steps: A sleep state analysis report is obtained by monitoring and analyzing different sleeping positions. The sleep state includes the waking period, light sleep period, deep sleep period, and deep REM sleep period.

[0013] Furthermore, the awake period, light sleep period, deep sleep period, and deep REM sleep period are determined according to the following steps: When the recording window The number of body movements within a time period is greater than the threshold. Then they are in a state of wakefulness; When the recording window The number of body movements within a time period is less than the threshold. , but greater than They are in a light sleep stage; When the recording window The number of body movements within a time period is less than the ninth threshold. But greater than the threshold At this time, they enter deep sleep or deep REM sleep.

[0014] The present invention provides a vital signs and sleep monitoring system based on millimeter-wave radar, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the above-described method.

[0015] The beneficial effects of this invention are as follows: This invention provides a method and system for monitoring vital signs and sleep based on millimeter-wave radar. The method utilizes millimeter-wave radar to achieve non-contact sleep monitoring. First, it uses millimeter-wave radar to collect radar echo signals obtained from chest wall vibrations during sleep. Then, it uses a bed-area positioning algorithm to determine the location of the monitored object. Next, it uses an adaptive noise extraction algorithm to effectively eliminate interference and extract the target's vital sign signals. Finally, it uses a multi-feature body movement detection algorithm to accurately detect the first and second amplitude body movements of the target. Then, it uses a multi-feature apnea judgment algorithm to detect whether the target has experienced apnea events in real time and to issue alarm information if the event lasts for a long time. Finally, it uses a multi-feature sleeping posture judgment algorithm to effectively detect supine, lateral, and prone sleeping postures. Finally, it uses a sleep monitoring algorithm to monitor four sleep states: wakefulness, light sleep, deep sleep, and REM sleep. The results are combined with the previously monitored information to generate a user sleep quality analysis report and provide health recommendations.

[0016] This invention also provides a system or WeChat mini-program that can generate detailed sleep reports by monitoring vital signs throughout the night, effectively detecting sleep apnea events, and is suitable for long-term monitoring and analysis of personal sleep conditions in a home environment.

[0017] Furthermore, this invention incorporates advanced signal processing technologies and algorithms, resulting in more accurate and reliable monitoring results. This method is simple, accurate, cost-effective, and suitable for various application scenarios such as homes, hospitals, and nursing homes, providing users with 24 / 7 health monitoring services.

[0018] Other advantages, objectives, and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination, or may be learned from practice of the invention. The objectives and other advantages of the invention can be realized and obtained through the following description. Attached Figure Description

[0019] To make the objectives, technical solutions, and beneficial effects of this invention clearer, the following figures are provided for illustration: Figure 1 This is an actual installation diagram of this embodiment.

[0020] Figure 2 This is a diagram of the bed area positioning algorithm in this embodiment.

[0021] Figure 3 This is a diagram of the adaptive noise extraction algorithm in this embodiment.

[0022] Figure 4 This is a diagram of the multi-feature motion detection algorithm in this embodiment.

[0023] Figure 5 This is a diagram of the multi-feature sleep apnea detection algorithm in this embodiment.

[0024] Figure 6 This is a diagram of the multi-feature sleeping posture determination algorithm in this embodiment.

[0025] Figure 7 This is a diagram of the sleep monitoring algorithm in this embodiment.

[0026] Figure 8 This document presents a flowchart of a vital sign monitoring method based on millimeter-wave radar and a sleep monitoring and analysis process.

[0027] Figure 9 This is a schematic diagram illustrating the effect of judging based on the body movement threshold.

[0028] Figure 10 This is a schematic diagram illustrating the effect of judging based on the time-related respiratory feature threshold.

[0029] Figure 11 This is a schematic diagram illustrating the effect of sleep posture assessment. Detailed Implementation

[0030] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0031] Example 1

[0032] This embodiment provides a high-precision human sleep breathing monitoring method based on millimeter-wave radar. This method can accurately monitor human vital signs during sleep without direct contact with the human body. It also innovatively utilizes radar spectrum data and innovative signal processing algorithms to accurately identify sleep apnea and sleeping posture. Furthermore, it enables sleep monitoring, and after sleep ends, users can view their daily sleep status report and health analysis through a WeChat mini-program.

[0033] The vital signs monitoring method and sleep monitoring method based on millimeter-wave radar provided in this embodiment mainly include the following steps: Step 1: Use millimeter-wave radar to collect radar echo signals obtained from chest wall vibration during sleep; determine the location of the monitored object using a bed area positioning algorithm; Using a millimeter-wave radar stand positioned high above the bed, the radar continuously emits electromagnetic waves towards the bed area after power-on. The system receives and processes the echoes from the bed area. Upon receiving the echo signals, the device performs a Fourier transform on the echo information to extract distance image data for the bed area. A bed area positioning algorithm is then used to determine whether a person is in the bed. If the person is not in the bed, vital sign monitoring is not performed. If the person is detected in the bed, vital sign monitoring and sleep monitoring are performed.

[0034] Step 2: After locating the human body in the bed area in Step 1 and determining that it is in bed, the adaptive noise extraction algorithm effectively eliminates interference and extracts the target's vital signs signals; This embodiment uses an adaptive noise extraction algorithm to extract the background noise of the current scene, then calculates the signal-to-noise ratio with the distance image data, and performs moving target detection on the distance image to extract the distance gate data of the target with high signal-to-noise ratio.

[0035] Step 3: Apply a multi-feature body motion detection algorithm to the target distance gate data obtained in Step 2. The multi-feature body motion detection algorithm accurately detects the first and second amplitude body movements of the target. The features primarily include distance-time second derivative information, distance-image spectrum amplitude information, and distance-gate diffusion information. Multi-feature detection identifies whether the target has experienced the first amplitude body movement (turning over) and the second amplitude body movement (coughing, using a mobile phone, etc.). If the first amplitude body movement is detected, the time of the movement needs to be separated from the data entering the vital signs monitoring.

[0036] In this embodiment, the first amplitude body movement is set to a large amplitude body movement, and the second amplitude body movement is set to a small amplitude body movement. The size of the body movement amplitude can be determined according to the actual situation.

[0037] Step 4: Extract respiratory and heart rate signals. Extract phase information from the resting data (where no body movement occurred) in Step 3. Perform a Fourier transform on the phase to obtain its periodicity. Similarly, use the adaptive noise extraction algorithm from Step 1 to extract the respiratory signal with the highest intensity within the respiratory frequency range. Use an IIR filter to extract the heart rate signal. Perform frequency detection on the extracted respiratory and heart rate signals, calculate the real-time heart rate and respiration using formulas, and report abnormalities such as slow breathing, rapid breathing, and heart rate variability in the sleep report.

[0038] Step 5: Real-time detection of whether the target has experienced a breathing apnea event using a multi-feature breathing apnea judgment algorithm, and prompting an alarm message if the event lasts for a long time (i.e., exceeds the preset time); In this embodiment, a multi-feature apnea detection algorithm is used to detect whether apnea has occurred in real time. When apnea occurs, in addition to the weakening of the frequency peak of the corresponding respiratory band, the intensity, standard score, and range of the respiratory signal are simultaneously used as features for apnea detection. Multi-feature fusion signal detection greatly improves the detection accuracy and real-time performance.

[0039] Step 6: Real-time detection of target sleep posture using a multi-feature sleep posture judgment algorithm, effectively detecting supine, lateral, and prone sleeping postures; This embodiment performs statistical processing on the one-dimensional distance image spectrum, two-dimensional distance velocity spectrum, heart rate change and respiratory rate change under different sleeping positions as sleep posture feature information, and performs joint judgment on the three feature information to detect the current target sleep posture.

[0040] Step 7: Monitor four sleep states—wakefulness, light sleep, deep sleep, and REM sleep—using sleep monitoring algorithms, and combine the previously monitored information to generate a user sleep quality analysis report and provide health advice; In this embodiment, the data obtained through the above steps regarding bed status, heart rate, respiration, body movement, respiratory status (pause, bradykinesia, tachykinesia), and sleep posture will be used in this step for target sleep monitoring and analysis. The sleep monitoring section uses the aforementioned characteristic data, employing a combination of body movement recording and cardiopulmonary coupling, to assess and analyze the target individual's wakefulness duration, light sleep duration, deep sleep duration, and REM deep sleep duration. It also generates a sleep quality analysis report based on the monitored heart rate, respiration, body movement, respiratory status, and sleep posture.

[0041] Furthermore, step 1 includes the following sub-steps: Step 1.1: Analyze the echo received by the millimeter-wave radar. With local oscillator signal The intermediate frequency signal is obtained by mixing. The formula is as follows:

[0042] in, This indicates the echo received by the millimeter-wave radar; This indicates the signal strength of the received echo; This indicates the signal frequency of the local oscillator signal; This indicates the frequency offset of the echo signal relative to the local oscillator signal; Represents a time variable; Indicates the phase of the echo signal; Indicates the local oscillator signal; The local oscillator signal strength is represented by this value. Indicates intermediate frequency signal; Step 1.2: Perform a Fourier transform on the intermediate frequency signal in the range image dimension to obtain a one-dimensional range image spectrum.

[0043] Step 1.3: The bed area localization algorithm determines whether a person is on the bed: First, the millimeter-wave radar adaptively updates the bed monitoring range based on the installation height and bed width input through the user interface. If no input is provided, a default installation height and bed width are set. The CACFAR algorithm is then used to detect if any targets exceeding a threshold are present. If a suspected target is detected, the system continues to check for signs of life. If no such sign is detected, the system only suspects someone is present but does not classify them as being on the bed, thus avoiding the detection of false human targets and improving anti-interference performance.

[0044] Furthermore, the adaptive noise extraction algorithm in step 2 is described in detail as follows: An adaptive noise extraction algorithm is applied to the one-dimensional distance image spectrum obtained in step 1 to determine the distance gate of the human target: First, the distance image signals of the bed area and non-bed area are processed. After moving target processing, targets with moving information will be screened out. The static signals that are screened out contain clutter and background noise. The background noise can be filtered out by smoothing filtering. When the strength of the distance gate signal of the suspected target is greater than the set signal-to-noise ratio, it is confirmed as a target, and the corresponding distance gate signal is moved to the next step of processing.

[0045] Furthermore, the multi-feature motion detection algorithm in step 3 is described in detail below: A multi-feature body movement detection algorithm is applied to the target's distance gate data: three features are calculated for the target's distance gate data: distance-time second derivative information, distance image spectrum amplitude information, and distance gate diffusion information. The first amplitude body movement mainly refers to relatively violent body movements, such as turning over or kicking; the second amplitude body movement mainly refers to slight body movements, such as slight head movements or small movements of the hands or feet.

[0046] When it is detected that: the second derivative information of distance time is less than the body movement threshold 1, the amplitude information of distance image spectrum is less than the body movement threshold 2, and the distance gate diffusion information is less than the body movement threshold 3, it is judged that there is no body movement and the target is in a resting state. When the following are detected: the second derivative information of distance time is less than the body motion threshold 1, the amplitude information of distance image spectrum is greater than the body motion threshold 2 and less than the body motion threshold 4, and the distance gate diffusion information is greater than the body motion threshold 3, it is judged to be the second amplitude body motion, and the target is in the second amplitude body motion state. When the following are detected: the second derivative of distance time is greater than the body movement threshold 1, the amplitude of distance image spectrum is greater than the body movement threshold 4, and the distance gate diffusion information is greater than the body movement threshold 3, it is determined to be the first amplitude body movement. The target is in the first amplitude body movement state. At this time, the target's sleep stage may change. The first amplitude body movement information will be input as a sleep feature to the sleep monitoring module, and the data of the body movement occurrence time needs to be separated from the data entering the vital signs monitoring.

[0047] In this embodiment, the first body movement threshold 1, the second body movement threshold 2, the third body movement threshold 3, and the fourth body movement threshold 4 are determined as follows: After data collection and analysis from 10 test subjects (7 males, 3 females, 5 overweight, and 5 of normal build), it was found that the distance-time second derivative information value decreased significantly during body movement compared to when no body movement occurred. Therefore, the first body movement threshold 1 was determined to be -5 times the distance-time second derivative information value under steady-state conditions. During body movement, the magnitude of the distance image spectrum amplitude information was significantly larger than when no body movement occurred. However, the difference between the second and first amplitude body movements was different. Therefore, the second body movement threshold 2 was determined to be 5 times the distance image spectrum amplitude information value under steady-state conditions, and the third body movement threshold 3 was determined to be 10 times the distance image spectrum amplitude information value under steady-state conditions. To better distinguish between the second and first amplitude body movements, distance gate diffusion information was also introduced, such as... Figure 9 As shown, Figure 9 This diagram illustrates the effect of judging based on body movement thresholds. The fourth body movement threshold, 4, is determined based on the difference in actual displacement changes caused by small and large movements of a normal-sized human body. However, due to the different body types of the test subjects, it can be seen that relying on any single indicator may not be completely accurate. Therefore, the introduction of multiple indicators for joint judgment greatly improves accuracy and reduces the false judgment rate.

[0048] Furthermore, the frequency detection process for the extracted respiratory and heartbeat signals in step 4 is performed according to the following sub-steps: Step 4.1: Extract phase information from the resting data in Step 3 that did not involve body movement, and obtain the original phase information by unwinding the phase.

[0049] Step 4.2: Perform a Fourier transform on the original phase information to obtain the phase change period information. Use an adaptive noise extraction algorithm to smooth out the background noise, and identify the period information in the Fourier transformed phase period information that has a higher than the set signal-to-noise ratio as the target respiratory signal period.

[0050] Step 4.3: Perform IIR filtering on the original phase information to extract the heartbeat signal. Then, perform spectral analysis on the heartbeat signal to calculate its Fourier transform and identify the periodicity of the detected heartbeat signal. Step 4.4: Calculate the real-time heart rate and respiration using the following formula:

[0051] Slow breathing, rapid breathing, and heart rate variability are reported as abnormalities in the sleep report.

[0052] Furthermore, the multi-feature sleep apnea detection algorithm in step 5 is described in detail below: When apnea occurs, in addition to the weakening of the peak frequency of the corresponding respiratory band, the intensity, standard score, and range of the respiratory signal are also used as features to determine apnea.

[0053] Apnea detection was performed on the intensity, standard fraction, and range of the raw phase respiratory signal obtained in step 4. The respiratory signal intensity feature is valid when the intensity of the respiratory signal is less than the first respiratory feature threshold of 1. The standard score of the respiratory signal is valid when its score is less than the second respiratory feature threshold of 2. The respiratory signal range feature is valid when the range of the respiratory signal is less than the third respiratory feature threshold 3.

[0054] When the three characteristics are valid for a period of time that exceeds the set apnea judgment time, an apnea event is confirmed to have occurred.

[0055] Multi-feature fusion signal detection greatly improves detection accuracy and real-time performance.

[0056] The first respiratory feature threshold 1, the second respiratory feature threshold 2, the third respiratory feature threshold 3, and the set apnea judgment time provided in this embodiment are determined as follows: Data was collected from 10 test subjects (7 males, 3 females, 5 overweight, and 5 of normal weight). The experiment involved simulating respiratory arrest by holding one's breath for 30 seconds after a period of normal breathing, followed by another 30-second breath-holding period after resuming normal breathing. A total of three respiratory arrests were performed per experiment, lasting approximately 6 minutes. During respiratory arrest, if... Figure 10 As shown, Figure 10This diagram illustrates the effect of judging breathing based on a time-dependent characteristic threshold; signal strength, range, and standard score all show significant troughs. Therefore, statistical analysis of data from 10 test subjects revealed a consistent trend, demonstrating that a uniform threshold can distinguish between normal breathing and apnea. It also shows that relying on any single indicator is not always entirely accurate; therefore, introducing multiple indicators for joint judgment significantly improves accuracy and reduces the false positive rate. The apnea judgment time is set based on the fact that in sleep disorders, if an adult experiences 5 or more apnea events (each lasting at least 10 seconds) within an hour, accompanied by daytime sleepiness, further investigation may be necessary to determine if sleep apnea syndrome is present. Therefore, the apnea judgment time is set at 10 seconds.

[0057] Furthermore, the multi-feature sleeping posture determination algorithm in step 6 is described in detail below: Sleep posture assessment mainly focuses on three sleeping positions: supine, lateral, and prone. The radar echo signal of the human body differs in different sleeping positions because sleeping posture affects the body's radar wave reflection characteristics.

[0058] When lying on your back, the chest and abdomen face the radar, the body is relatively flat, and because the body's major organs (such as the lungs and heart) are directly facing the radar, a strong echo signal is generated. The rise and fall of the chest caused by breathing and heartbeat will form a relatively obvious periodic change in the signal.

[0059] When lying on your side, compared to lying on your back, the movement of one side of the body (especially the side facing the radar) will produce more significant signal changes. The signal strength and pattern will also vary depending on the part of the body facing the radar (e.g., when lying on your left side, the left chest faces the radar).

[0060] When lying prone, the radar waves mainly come into contact with the back. Although the back movements caused by breathing can still be detected, the signal is not as strong as when there is frontal contact.

[0061] Therefore, after extensive experimental testing, one-dimensional distance image spectrum, two-dimensional distance velocity spectrum, heart rate change, and respiratory rate change were selected as sleep posture feature information.

[0062] This invention also provides a WeChat mini-program that allows target users to view generated sleep reports daily. Through monitoring vital signs throughout the night, it can generate detailed sleep reports, effectively detect sleep apnea events, and is suitable for long-term monitoring and analysis of personal sleep conditions in a home environment.

[0063] Example 2

[0064] To more clearly illustrate the objectives, technical solutions, and advantages of the embodiments of this application, the following will provide a detailed and systematic description of the relevant technical solutions in conjunction with the accompanying drawings provided in the embodiments of this application. Efforts will be made to ensure the accuracy and completeness of the description in order to fully understand the technical details and application value of this embodiment.

[0065] like Figures 1 to 8 As shown, the present invention employs a method for non-contact sleep monitoring using millimeter-wave radar, the steps of which are as follows: Step 1: Power on the 77GHz frequency-modulated continuous millimeter-wave radar and suspend it approximately 1.2 meters above the head of the bed. The radar continuously receives the following echoes:

[0066] in, Echo signal strength The local oscillator frequency of the transmitted signal. It is the original phase of the transmitted signal. It is the Doppler frequency change caused by the motion information of an object, collected by the echo. After receiving the echo, it will be compared with the local oscillator signal. ,in The frequency of the local oscillator signal is used for mixing to obtain the intermediate frequency signal.

[0067] The obtained intermediate frequency signal is subjected to a one-dimensional range image Fourier transform, and then a bed area localization algorithm is used to determine whether a person is on the bed. The representation of the bed area on the range image is calculated by the following steps: the radar distance to the bed surface height is... In this example, it equals 1.2 meters; the radar range resolution is... In this example, the radar range resolution is 0.025 meters; the bed length is... In this example, the bed length is 1.8 meters; the number of points in the one-dimensional distance image Fourier transform is... In this example, let the number of points in the Fourier transform of the one-dimensional range image be 512; then, for the radar range image, the area of ​​the bed is: ; in, It is the detected distance to the gate position; This indicates the radar's distance from the bed surface; Indicates radar range resolution; Indicates the length of the bed; For a one-dimensional range image, perform on-the-bed detection; the range gate for detecting the target is... If the target is located within the bed area, a suspected target is detected. The target is then extracted. The data is examined to see if there are frequency components of vital signs signals in the spectrum data; if so, it is considered that the person is in bed.

[0068] Step 2: Once a human target is detected on the bed, the target with motion information will be filtered out after processing the one-dimensional range image. The static signal that is filtered out contains clutter and background noise. The background noise can be filtered out by smoothing filtering. When the range gate signal strength of the suspected target is greater than the set signal-to-noise ratio If the signal is positive, it is confirmed as the target. At this point, the range gate signal corresponding to the maximum signal-to-noise ratio is extracted. It refers to the reflected signal from the target's chest cavity.

[0069] Step 3: [Regarding...] Develop a multi-feature motion detection algorithm and calculate the second-order distance-time derivative information. Distance image spectral amplitude information Distance gate diffusion information Three features. A joint decision on the effectiveness of the three features is made using four thresholds: , , , .in, Indicates the first body motion threshold; Indicates the second body motion threshold; Indicates the third body's dynamic threshold; This represents the fourth body motion threshold.

[0070] First, calculate the distance-time second derivative information. The second derivative is the derivative of the first derivative of a function, and it reflects the concavity / convexity of the signal. In this example, we need to calculate the reflected signal from the target's chest cavity. Calculate the second derivative of distance and time, and calculate the first derivative as follows: Calculating the second derivative is simply taking the derivative of the first derivative again; therefore, the formula is: , in, It is the length of the observation window.

[0071] In this example, it is necessary to detect the reflected signal at the target's chest cavity. Calculate the spectral amplitude information of the range image The calculation formula is as follows: , In this example, it is necessary to detect the reflected signal at the target's chest cavity. Calculate distance gate diffusion information The calculation formula is as follows: , This example categorizes bodily movements into first-amplitude and second-amplitude bodily movements based on the distance, intensity, and speed of the movement. First-amplitude bodily movements primarily involve more vigorous body movements, such as rolling over or kicking. Therefore, the characteristics of first-amplitude bodily movements are as follows: When the second derivative information of distance and time Less than the first body movement threshold Distance image spectral amplitude information Less than the second body motion threshold Distance to the gate diffuses information Less than the third body motion threshold At that time, it was determined that there was no physical movement and the target was in a resting state; When the second derivative information of distance and time is detected: Less than the first body movement threshold Distance image spectral amplitude information Greater than the second body motion threshold Less than the fourth body motion threshold Distance to the gate diffuses information Greater than the third body motion threshold At that time, it was determined to be a second-amplitude body movement, and the target was in a second-amplitude body movement state; When the second derivative information of distance and time is detected: Greater than the first body movement threshold Distance image spectral amplitude information Greater than the fourth body motion threshold Distance to the gate diffuses information Greater than the third body motion threshold When the target is in the first amplitude body movement state, it is determined that the target is in the first amplitude body movement state. At this time, the target's sleep stage may change. The first amplitude body movement information will be input into the sleep monitoring module as a sleep feature, and the data of the time of body movement needs to be separated from the data entering the vital signs monitoring.

[0072] Step 4: If no movement event is detected in the target, proceed with... Phase signal extraction For phase signals Perform a Fourier transform; the number of Fourier transform points is... In this example, the Fourier transform points are set to 512, and the periodic information in the phase signal is obtained after the transform. ,in The adaptive noise extraction algorithm is used again to smooth out the background noise. The phase period information after Fourier transform is higher than the set signal-to-noise ratio. The periodic information was confirmed as the target respiratory signal period. Based on the normal distribution of human respiratory rate: A Butterworth bandpass filter was designed, from Extracting respiratory signals In this example, let the currently extracted respiratory signal cycle be... for For phase signals IIR filtering was performed to extract the heartbeat signal. Based on the normal distribution of heart rate in the human body... Designed a The Butterworth bandpass filter amplifies the heartbeat signal contained in the original signal, making it easier to extract the heartbeat signal. After passing through a well-designed filter, the periodic information of the heartbeat signal can be obtained. The adaptive noise extraction algorithm is used again. Smoothing filters remove background noise, and extracts phase period information from the Fourier transform that has a higher signal-to-noise ratio than the set value. The periodic information was confirmed as the target's heartbeat signal period. In this example, let the currently extracted respiratory signal cycle be... for The real-time heart rate and respiration are calculated using a formula: , Slow breathing, rapid breathing, and heart rate variability are reported as abnormalities in the sleep report.

[0073] Step 5: When apnea occurs, in addition to the weakening of the peak frequency of the corresponding respiratory band, the intensity, standard score, and range of the respiratory signal are simultaneously used as features for apnea determination. The original phase signal obtained in Step 4 is then processed... intensity Standard scores Range Perform a sleep apnea test.

[0074] First, calculate the strength. Intensity refers to the absolute amplitude of a signal. In this example, the object for calculating intensity is... Therefore, it can be calculated using the formula: , in, It is the length of the observation window.

[0075] Next, calculate the standard score. A standard score represents the degree of deviation of a raw score from the mean of a set of data. In this example, the object for calculating the standard score is... Therefore, it can be calculated using the formula: , in It is the average of observations within the observation window, i.e., the signal. The mean, The standard deviation within the observation window, i.e., the signal. The standard deviation.

[0076] Next, calculate the range. The range is a fundamental statistic describing the dispersion of data; it represents the difference between the maximum and minimum values ​​in a set of data. In this example, the object for calculating the standard score is... Therefore, it can be calculated using the formula: , There are also three corresponding threshold values ​​set for the three features, depending on the intensity of the respiratory signal. Less than the respiratory characteristic threshold The intensity characteristics of the respiratory signal are valid, and the standard score of the respiratory signal is... Less than the respiratory characteristic threshold The standard fractional characteristic of the respiratory signal is valid, and the range of the respiratory signal is... Less than the respiratory characteristic threshold The range feature of the respiratory signal is valid. When all three features are valid simultaneously, an apnea event is confirmed. Multi-feature fusion signal detection greatly improves detection accuracy and real-time performance.

[0077] Step 6: Sleeping posture determination mainly involves judging the three sleeping postures: supine, side-lying, and prone. The radar echo signal of the human body will be different in different sleeping postures because the sleeping posture affects the body's radar wave reflection characteristics.

[0078] When lying on your back, the chest and abdomen face the radar, the body is relatively flat, and because the body's major organs (such as the lungs and heart) are directly facing the radar, a strong echo signal is generated. The rise and fall of the chest caused by breathing and heartbeat will form a relatively obvious periodic change in the signal.

[0079] When lying on your side, compared to lying on your back, the movement of one side of the body (especially the side facing the radar) will produce more significant signal changes. The signal strength and pattern will also vary depending on the part of the body facing the radar (e.g., when lying on your left side, the left chest faces the radar).

[0080] When lying prone, the radar waves mainly come into contact with the back. Although the back movements caused by breathing can still be detected, the signal is not as strong as when there is frontal contact.

[0081] Therefore, after extensive experimental testing, we selected the one-dimensional distance image spectrum as the basis for our selection. Two-dimensional distance-velocity spectrum Heart rate changes Changes in respiratory rate This refers to sleep posture characteristics. These four types of characteristics need to be obtained from the results of the previous steps.

[0082] First, the one-dimensional distance image spectrum. The distance profile is obtained by performing a one-dimensional range profile Fourier transform on the intermediate frequency signal obtained in step 1. The one-dimensional range profile spectrum is then analyzed. Statistical analysis was performed to determine the one-dimensional distance image spectrum when the characteristic conditions of one of the three sleeping positions were met. The feature information is valid.

[0083] Secondly, the two-dimensional distance-velocity spectrum. The one-dimensional range profile spectrum is obtained by performing a one-dimensional range profile Fourier transform on the intermediate frequency signal obtained in step 1. After performing another Doppler-dimensional Fourier transform, the two-dimensional distance-velocity spectrum is obtained. For the two-dimensional range-velocity spectrum Statistical analysis was conducted to determine the two-dimensional distance-velocity spectrum when the characteristic conditions of one of the three sleeping positions were met. The feature information is valid.

[0084] Finally, heart rate changes and changes in respiratory rate This refers to the rate of change in heart rate and respiratory rate. The heart rate and respiratory rate results obtained in step 4 are stored in a repository. Statistical analysis is performed on this data repository at regular intervals to obtain information on the changes in heart rate and respiratory rate within the observation window. When the characteristic conditions of one of the three sleeping positions are met, the heart rate change... Changes in respiratory rate The feature information is valid.

[0085] When all three feature information conditions simultaneously meet the feature conditions of a certain sleeping position, the current sleeping position of the person being tested can be confirmed, such as... Figure 11 The data shown in (a) and (b) are used as examples for illustration. Figure 11 This is a schematic diagram illustrating the effect of sleep posture assessment.

[0086] When lying supine, the one-dimensional distance image spectrum The number of peak points is relatively small and the diffusion distance is short; the velocity spectrum information is stable and balanced; and the chest cavity is in a most comfortable state without any pressure, at which point the heart rate changes... and changes in respiratory rate The change is relatively slow, and the statistical information shows that the change trend is slow and the degree of change is low. The above conditions are jointly judged by a threshold structure (which includes the number of peak points, diffusion distance, speed stability and change rate mentioned above). When the overall threshold structure Thod5 is satisfied, it is supine.

[0087] One-dimensional distance image spectrum when lying on one's side The number of peak points is relatively large and the diffusion distance is large; the velocity spectrum information is unstable and drastic; at this time, the heart rate changes. and changes in respiratory rate The change is relatively rapid, and the statistical information shows that the change trend is unstable and the degree of change is large. The above conditions are jointly judged by a threshold structure (which includes the number of peak points, diffusion distance, speed stability and change rate mentioned above). When the overall threshold structure Thod6 is satisfied, it is considered to be lying on one's side.

[0088] When prone, the one-dimensional distance image spectrum The number of peak points is high and the diffusion distance is large; the velocity spectrum information is stable and balanced, but the overall signal strength is weak; the chest cavity is under compression, at which time the heart rate changes... and changes in respiratory rate The change is relatively rapid, and the statistical information shows a rapid trend and a large degree of change; when the overall condition meets the threshold structure Thod7, it is considered to be prone.

[0089] The threshold structures Thod5, Thod6, and Thod7 in this embodiment are determined as follows: Data was collected from 10 test subjects (7 males, 3 females, 5 overweight, and 5 of normal weight). Specific thresholds were determined as follows: Figure 11 As shown, based on the data collection of 10 testers for 3 minutes of single posture, only the data after the middle minute of stabilization was taken for statistical analysis to obtain the value of the threshold structure used in sleeping posture judgment, which is accurate and applicable to 90% of the population.

[0090] Step 7: The data obtained from the previous steps regarding bed condition, heart rate, respiration, body movement, respiratory status (pause, bradykinesia, tachykinesia), and sleep posture will be used in this step for target sleep monitoring and analysis. The sleep monitoring section analyzes the above characteristic data to determine the target individual's awake time, light sleep time, and deep sleep time. The system analyzes and reports the sleep quality of target individuals by monitoring their heart rate, respiration, body movement, respiratory status, and sleeping posture.

[0091] This method primarily categorizes sleep states into four stages: wakefulness, light sleep, deep sleep, and REM sleep. Furthermore, it mainly employs a combination of motion recording and cardiopulmonary coupling to assess sleep states.

[0092] Motion recording is a non-invasive monitoring method primarily used to assess an individual's activity level and sleep-wake patterns. Motion recorders capture an individual's movements, thereby inferring their sleep status. Prolonged periods of low activity levels usually indicate that the individual is asleep. Frequent movements or higher activity levels suggest that the individual may be awake. Occasional slight movements during sleep (such as turning over) can also be captured, and this information helps assess sleep quality.

[0093] In this example, the physical activity result obtained in step 3 will be input into the sleep monitoring module, and the recording window length is [missing information]. Using thresholds , , Used as the threshold for sleep staging, defined when the recording window... The frequency of moderate to strong body movements. State machine transitions between three sleep states: wakefulness, light sleep, and deep sleep. The deep sleep state includes deep sleep phase and deep REM sleep, such as... Figure 7 As shown: When the recording window The number of body movements within a time period exceeds the eighth threshold. If the number of physical movements is high, it indicates that the target person is actively moving and is in a conscious state. When the recording window The number of body movements within a time period is less than the eighth threshold. However, it is greater than the ninth threshold. If the number of body movements decreases slightly, it indicates that the target person is basically in the stage of falling asleep, in a light sleep period; When the recording window The number of body movements within a time period is less than the ninth threshold. But greater than the tenth threshold This indicates that the target person is basically in a state of rest and sleep, with little movement, and has entered a deep sleep or REM sleep stage.

[0094] The eighth threshold in this embodiment Ninth threshold The tenth threshold The determination is as follows: Based on existing professional literature and research reports, the eighth threshold can be basically determined. Ninth threshold The tenth threshold This device will detect body movements less than the tenth threshold per minute. The frequency of body movements is defined as "low activity." For example, some literature defines "low activity" as fewer than 3 body movements per minute. A prolonged period of low activity can be considered a deep sleep state. A frequency exceeding the eighth threshold indicates a deeper sleep state. The frequency of activity is defined as "high activity." For example, more than 5 body movements per minute can be defined as "high activity." If one remains in a state of high activity for an extended period, it can be determined that one is awake. A state between these two can be considered light sleep. Data was collected from 10 test subjects (7 men, 3 women, 5 overweight, and 5 of normal weight) to validate this method, demonstrating its ability to achieve relatively accurate analysis of light and deep sleep.

[0095] Cardiopulmonary coupling is based on the natural interaction between the heart and respiratory systems. The interaction between the heart and respiratory systems changes at different stages of sleep. These changes can be reflected in the coupling strength between electrocardiogram (ECG) and respiratory signals. Studies have shown that during deep sleep, breathing becomes more regular, heart rate and blood pressure decrease, and the cardiopulmonary coupling strength is generally higher. During light sleep and REM sleep, breathing and heart rate are more irregular, and the cardiopulmonary coupling strength is lower. In this example, although millimeter-wave radar cannot directly obtain ECG signals, the heart rate signal, specifically the RR interval within the ECG signal, can be obtained from the detected heart rate signal. Therefore, a similar analysis can be performed, coupling the heart rate signal obtained from the echo signal with the respiratory signal to obtain the coupling strength. The respiratory signal obtained in step 4... With heart rate signals Coupling to obtain cardiopulmonary coupling strength .

[0096] First calculate and The autocorrelation function can be obtained from the formula. and Autocorrelation function:

[0097]

[0098] in and These are respiratory signals and heart rate signals The conjugate function of .

[0099] The power spectral density function of the signal can then be obtained by performing a Fourier transform on the autocorrelation function, as shown in the following formula:

[0100]

[0101] As above, calculate and The cross spectral density function is first calculated as follows: and Cross-correlation function:

[0102] The cross-spectral density function of the two signals can then be obtained by performing a Fourier transform on the cross-correlation function, as shown in the following formula.

[0103] therefore and Coupling strength It is calculated using the following formula:

[0104] Combining motion recording with cardiopulmonary coupling can successfully distinguish four sleep states: wakefulness, light sleep, deep sleep, and deep REM sleep.

[0105] Finally, by combining the monitoring information from the previous steps, a sleep report is generated, which includes the sleep status of the tested target and health analysis suggestions. Users can view the report by clicking on it in the mini-program.

[0106] The above-described embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention. The scope of protection of the present invention is defined by the claims.

Claims

1. A method for monitoring vital signs and sleep based on millimeter-wave radar, characterized in that: Includes the following steps: The radar echo signal obtained by collecting chest wall vibration during sleep using millimeter-wave radar. The location of the monitoring target is determined by a bed area positioning algorithm; The adaptive noise extraction algorithm effectively eliminates interference and extracts the target's vital signs signals. The first and second amplitude body movements of the target are accurately detected using a multi-feature body movement detection algorithm. Extract breathing and heartbeat signals, use a multi-feature apnea judgment algorithm to detect whether the target has experienced an apnea event in real time, and issue an alarm message if the event occurs for more than a threshold time. The multi-feature sleep apnea detection algorithm is performed according to the following steps: When apnea occurs, the intensity, standard score, and range of the respiratory signal are used simultaneously as characteristics to determine apnea, as detailed below: The intensity feature of the respiratory signal is valid when the intensity of the respiratory signal is less than the first respiratory feature threshold. The standard score feature of the respiratory signal is valid when the standard score of the respiratory signal is less than the second respiratory feature threshold. The range feature of the respiratory signal is effective when the range of the respiratory signal is less than the third respiratory feature threshold. When the three characteristics are valid for a period of time that exceeds the set apnea judgment time, an apnea event is confirmed to have occurred. The multi-feature sleeping posture judgment algorithm effectively detects different sleeping postures; The sleep monitoring algorithm monitors and analyzes different sleeping positions, and combines the monitoring information to generate a user's sleep quality analysis report; The multi-feature motion detection algorithm is performed according to the following steps: Calculate the second-order derivative of the distance time, the amplitude of the distance image spectrum, and the distance gate diffusion information based on the distance gate data of the target location; When it is detected that: the second derivative information of distance time is less than the first body movement threshold, the amplitude information of distance image spectrum is less than the second body movement threshold, and the distance gate diffusion information is less than the third body movement threshold, it is judged that there is no body movement and the target is in a resting state. When it is detected that: the second derivative information of distance time is less than the first body motion threshold, the amplitude information of distance image spectrum is greater than the second body motion threshold and less than the fourth body motion threshold, and the distance gate diffusion information is greater than the third body motion threshold, it is judged to be the second amplitude body motion, and the target is in the second amplitude body motion state; When the following are detected: the second derivative of distance time is greater than the first body movement threshold, the amplitude of distance image spectrum is greater than the fourth body movement threshold, and the distance gate diffusion information is greater than the third body movement threshold, it is determined to be the first amplitude body movement. The first amplitude body movement information is then input as a sleep feature to the sleep monitoring module, and the data on the time of body movement occurrence is separated from the data entering the vital signs monitoring. The first amplitude body movement is set as a large amplitude body movement, and the second amplitude body movement is set as a small amplitude body movement.

2. The method for monitoring vital signs and sleep based on millimeter-wave radar as described in claim 1, characterized in that: The location of the monitored target is determined according to the following steps: Determine the installation location and bed parameters of the millimeter-wave radar, and determine the monitoring area of ​​the bed; The system detects whether any targets exceeding a threshold appear within the window monitoring area. If a target appears, it continues to detect whether any vital signs are present. If vital signs are detected, it is determined that a monitoring target exists in the bed monitoring area.

3. The method for monitoring vital signs and sleep based on millimeter-wave radar as described in claim 1, characterized in that: The extraction of respiratory and heartbeat signals is performed according to the following steps: Phase information is extracted from resting data in which no body movement has occurred, and the original phase information is obtained by phase unwinding; The original phase information is subjected to Fourier transform to obtain phase change period information. The period information in the Fourier transformed phase period information that is higher than the set signal-to-noise ratio is identified as the target respiratory signal period. The original phase information is subjected to IIR filtering to extract the heartbeat signal. The heartbeat signal is then subjected to spectrum detection to calculate the Fourier transform of the heartbeat signal and find the periodic information of the detected heartbeat signal. It calculates real-time heart rate and respiration.

4. The method for monitoring vital signs and sleep based on millimeter-wave radar as described in claim 1, characterized in that: The multi-feature sleeping posture determination algorithm is performed according to the following steps: Sleeping posture is determined based on one-dimensional distance image spectrum, two-dimensional distance velocity spectrum, heart rate change, and respiratory rate change; the sleeping posture includes three types: supine, lateral, and prone.

5. The method for monitoring vital signs and sleep based on millimeter-wave radar as described in claim 1, characterized in that: The sleep monitoring algorithm is performed according to the following steps: A sleep state analysis report is obtained by monitoring and analyzing different sleeping positions. The sleep state includes the waking period, light sleep period, deep sleep period, and deep REM sleep period.

6. The method for monitoring vital signs and sleep based on millimeter-wave radar as described in claim 5, characterized in that: The states of wakefulness, light sleep, deep sleep, and deep REM sleep are determined according to the following steps: When the recording window The number of body movements within a time period is greater than the threshold. Then they are in a state of wakefulness; When the recording window The number of body movements within a time period is less than the threshold. , but greater than They are in a light sleep stage; When the recording window The number of body movements within a time period is less than the ninth threshold. But greater than the threshold When the cardiopulmonary coupling strength is less than the threshold 3, the patient enters the deep sleep stage or deep REM sleep stage. When the cardiopulmonary coupling strength is greater than the threshold 3, the patient enters the deep sleep stage. The threshold 3 refers to the cardiopulmonary coupling strength value in the cardiopulmonary coupling method.

7. A vital signs and sleep monitoring system based on millimeter-wave radar, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the method described in any one of claims 1 to 6.

Images