Exercise prescription health screening system and method based on pulse rest

Abstract

The invention belongs to the field of health screening instruments, and particularly relates to an exercise prescription health screening system based on pulse rest. The system comprises an acquisition module, a processing module and a regulation and control module. The acquisition module comprises a hanging ring, and the hanging ring is made of a silica gel material; a pulse sensor is arranged in the hanging ring, and the pulse sensor is a piezoelectric sensor; a plurality of airflow sensors are fixedly arranged on the right side of the pulse sensor in the hanging ring, and the airflow sensors are MEMS microphones; pressure sensors are evenly distributed on the surface, making contact with the neck, of the hanging ring. A gyroscope for collecting three-axis acceleration is arranged in the hanging ring, and the pulse sensor is used for collecting pulse data of the neck of a user. According to the scheme, the collection module, the processing module and the regulation module cooperate, the limitation of the prior art is broken through, and the problem that exercise prescription health screening is inaccurate due to the fact that physiological signals are easily disturbed by behaviors of a user when collected at the neck part is solved.

Description

TECHNICAL FIELD

[0001] The present scheme belongs to the field of health screening instruments, and particularly relates to a health screening system for exercise prescription based on pulse respiration technique. BACKGROUND

[0002] In the field of exercise health monitoring, existing wearable devices are mostly worn on the wrist to collect pulse and breathing frequency data. However, the wrist is far away from the respiratory organs, and the physiological signals generated by breathing are greatly attenuated when transmitted to the wrist, making it difficult to accurately collect breathing frequency. At the same time, the wrist part has large movement amplitude and complex action, and the collected pulse data is easily disturbed. The deviation in the collection of breathing frequency and pulse data will cause errors in the recognition of exercise state, and the matching degree between the exercise prescription generated based on the data and the real-time physiological state of the user will be significantly reduced, ultimately making it difficult to meet the needs of precise health screening, and restricting the scientificity and effectiveness of exercise health guidance.

[0003] Compared with the wrist, the neck part is closer to the heart and respiratory organs, and has better physiological signal collection conditions: the airflow change during breathing is more significant, and the signal transmitted by the heart is more intense and stable. Wearing the device near the neck can synchronously capture clearer and more accurate breathing frequency and pulse signals than the wrist, laying a good foundation for data collection. However, the blood vessels in the neck are relatively shallow, and the device has higher requirements for fit, which objectively increases the dependence on the cooperation of the user. In addition, the user's speaking, swallowing action and head movement during exercise can all produce strong interference to the neck signals, causing deviation in the recognition of exercise state, and thus leading to inaccurate exercise prescription health screening. SUMMARY

[0004] The purpose of the present scheme is to provide a health screening system for exercise prescription based on pulse respiration technique, to solve the problem that the physiological signals collected at the neck part are easily disturbed by the user's behavior, leading to inaccurate exercise prescription health screening.

[0005] In order to achieve the above purpose, the present scheme provides a health screening system for exercise prescription based on pulse respiration technique, comprising:

[0006] The collection module comprises a hanging ring, which is made of silica gel. A pulse sensor is arranged in the hanging ring, and the pulse sensor is a piezoelectric sensor. A plurality of airflow sensors are fixedly arranged on the right side of the pulse sensor in the hanging ring, and the airflow sensors are MEMS microphones. The surface of the hanging ring in contact with the neck is provided with uniformly distributed pressure sensors. A gyroscope for collecting three-axis acceleration is arranged in the hanging ring. The pulse sensor is used to collect pulse data of the user's neck, the airflow sensor is used to collect airflow data around the neck, and the pressure sensor is used to collect contact pressure data of the user at different positions.

[0007] The processing module acquires contact pressure data and triaxial acceleration. When the triaxial acceleration is not greater than a preset static threshold and the contact pressure data is within a preset wearing range, it acquires pulse data and airflow data. It calculates the ratio of pulse data to airflow data as the pulse-breath ratio by combining the acquisition time. The processing module stores the types of diseases and the corresponding pulse-breath ratio ranges and pulse ranges. The processing module uses pulse data and pulse-breath ratios to match the corresponding disease types.

[0008] The regulation module acquires disease type, pulse-to-breath ratio, and pulse data, and combines them to generate exercise prescription suggestions.

[0009] The processing module also analyzes the stable movement time by combining the contact pressure value and triaxial acceleration with the acquisition location and acquisition time when the triaxial acceleration is greater than the preset static threshold and the contact pressure data is within the preset wearing range. Based on the differences in airflow data at different acquisition locations during the stable movement time, it separates the breathing airflow and movement airflow from the ambient airflow. Then, it combines the breathing airflow, movement airflow, pulse data, and contact pressure data to calculate the exercise intensity and construct an exercise muscle function model. The control module obtains the exercise intensity and exercise muscle function model to generate exercise prescription suggestions.

[0010] The principle and technical effects of this solution are as follows: First, the flexibility and biocompatibility of the silicone loop ensure a stable fit of the device to the neck, reducing signal errors caused by loosening and lowering the risk of skin allergies. Its deformation adaptability avoids compressing blood vessels and affecting the signal. The piezoelectric pulse sensor directly collects neck pulse signals, reducing signal attenuation compared to wrist collection, and supports heart rate variability analysis through real-time waveforms. The MEMS airflow sensor accurately distinguishes between breathing, speaking, and movement airflow. Its ability to perceive minute airflows highlights the differences in airflow collected at different locations, such as a higher proportion of movement airflow and less breathing airflow at the back of the neck, and a lower proportion of movement airflow and more breathing airflow at the jaw. The uniform distribution of pressure sensors provides contact pressure data, and the detection of muscle movement changes is achieved by combining pressure change trends. The three-axis acceleration data of the gyroscope quantifies the motion state and calculates energy consumption through integration. These sensor data not only accurately collect physiological signals individually, but also solve the problems of device loosening and signal interference in existing technologies through spatial layout and material adaptation (such as the combination of the silicone loop and pressure sensors), providing a high-quality data foundation for subsequent analysis.

[0011] Secondly, by combining contact pressure and triaxial acceleration data, interference actions (such as swallowing and head turning) are eliminated by identifying stable motion time, and motion pattern classification is achieved using pressure-acceleration correlation analysis, overcoming the limitations of existing technologies that simply summarize data. When separating airflow data through independent component analysis, it not only accurately distinguishes between breathing and speaking airflow but also optimizes motion scenario judgment based on changes in environmental airflow, solving the problem of separating mixed signals that traditional methods struggle with. The motor function model integrates multi-dimensional data such as respiration, pulse, and pressure, and dynamically updates to adapt to different populations, providing a more comprehensive reflection of physiological state compared to existing single-indicator analyses. These processing steps form a progressive analysis chain: first, selecting effective time windows; then, separating signal types; and finally, constructing a comprehensive model. This not only eliminates core interference but also supports extended functions such as motion type recognition.

[0012] Furthermore, the regulation module generates personalized prescriptions based on exercise intensity, exercise performance models, and stable exercise time, and dynamically adjusts exercise recommendations through real-time data, overcoming the shortcomings of existing technologies that rely solely on a single indicator for prescription generation. The historical data comparison function supports reviewing exercise effects and optimizing long-term plans, extending single-session monitoring to full-cycle health management. The realization of these functions relies on multi-dimensional data collected at the front end and precise models processed in the middle, forming a closed loop: the data collection module provides data, the processing module purifies and analyzes it, and the regulation module generates decisions, ultimately achieving a high degree of matching between exercise prescriptions and real-time physiological states.

[0013] Existing technologies, lacking optimized sensor layout and deep data fusion, can only provide basic monitoring functions. This solution, however, achieves cross-module collaboration through dynamic adaptive fitting of the silicone ring, motion type recognition based on pressure and acceleration data, and real-time adjustment of exercise prescriptions. This collaboration not only solves the core problems of signal interference and prescription matching but also adds value, such as the identification of respiratory-muscle coordination anomalies, intelligent adaptation to exercise scenarios, and in-depth assessment of cardiopulmonary function. The multiple roles of data from each module (e.g., pressure data is used for both device fit monitoring and muscle state analysis) and cross-module collaboration (e.g., airflow data from the acquisition layer supports model construction in the processing layer, thus influencing prescription generation in the control layer) make the overall efficiency of this solution far exceed the sum of the individual effects of each component, truly achieving a breakthrough from single health monitoring to comprehensive sports and health management.

[0014] In summary, this solution, through the collaborative integration of acquisition, processing, and control modules, overcomes the limitations of existing technologies and solves the problem that physiological signals collected from the neck area are easily interfered with by the user's behavior, leading to inaccurate health screening of exercise prescriptions.

[0015] Furthermore, when constructing the exercise performance model, the processing module analyzes the variation patterns of contact pressure values ​​at different acquisition locations and in terms of numerical values. Combined with the acquisition time and variation patterns of triaxial acceleration, it analyzes and obtains the stable exercise time. It acquires airflow data at various locations during the stable exercise time, and separates chest vibration airflow and ambient airflow from the airflow data through independent component analysis. Then, based on the influence of chest vibration on respiration and the numerical differences in airflow data acquired at different locations, it separates respiratory airflow and exercise airflow from the ambient airflow. It combines the exercise airflow with triaxial acceleration to calculate exercise intensity, acquires pulse data during the stable exercise time, and combines the contact pressure data acquired simultaneously with respiratory airflow, chest vibration airflow, and pulse data to construct the exercise performance model.

[0016] Furthermore, the hanging ring is equipped with airflow sensors at the back of the neck and the neck corresponding to the chin. When the processing module processes the airflow data, it first collects airflow data at the back of the neck and the chin respectively, and records the difference in the collection time of the two data. Then, combining the periodic correspondence and synchronous change characteristics between the chest vibration airflow and the respiratory airflow, the respiratory airflow is separated from the airflow data collected at the chin. Subsequently, the difference between the airflow data collected at the back of the neck and the chin is compared to perform a second separation of the separated respiratory airflow. Finally, the respiratory airflow after the second separation is compared with the chest vibration airflow to exclude the respiratory-related airflow components from the airflow data. The remaining part is the motion airflow.

[0017] When users engage in strenuous exercise such as running, the time difference and airflow intensity characteristics of airflow sensors located at the back of the neck and jaw can accurately separate the airflow generated by breathing from that generated by running movements, avoiding interference from exercise airflow on breathing frequency and ensuring the stability of respiratory parameter calculations. When users speak or swallow, by comparing subtle differences in airflow data from two locations, instantaneous airflow interference from these movements can be efficiently identified and eliminated, ensuring consistently high accuracy of respiratory data. In contrast, existing wrist-worn devices, being far from the respiratory organs, suffer from severe respiratory signal attenuation and are easily interfered with by large wrist movements, making it difficult to obtain reliable data. Even existing neck-worn technologies do not employ multi-position airflow sensors and multi-step processing logic, failing to distinguish between breathing, speaking, and exercise airflow, often resulting in signal confusion during complex movements. Therefore, this solution, through targeted design, provides high-precision physiological signal support for exercise prescription generation.

[0018] Furthermore, airflow sensors are installed on both sides of the neck below the ears. The processing module analyzes the head movement direction and curvature based on the differences in airflow data collected by the airflow sensors on both sides. The implementation steps are as follows: First, airflow data is collected from both sides using the airflow sensors below the ears, and the differences in the data between the two sides are recorded. Then, the distribution pattern and trend of pressure data are analyzed by combining the numerical distribution changes of pressure sensors around the neck. The differences in airflow between the two sides are correlated and matched with the changes in pressure distribution to identify the user's head movement direction. The curvature of the head movement is analyzed by combining the degree of airflow difference and the range of pressure distribution changes with triaxial acceleration. The control module obtains the head movement direction and curvature information, and extracts the pulse change characteristics during the head movement process by combining the corresponding collection time. Based on the correlation between the pulse change characteristics and the head movement, an exercise prescription suggestion for head movement is generated.

[0019] In scenarios involving significant head movement, airflow sensors located on either side of the neck below the ears capture subtle airflow changes triggered by head movements. These, combined with pressure sensors recording pressure distribution changes around the neck, are analyzed by a processing module that examines the distribution patterns and trends of the pressure data. The module then correlates the differences in airflow between the two sides with these pressure distribution changes to accurately identify the direction of head movement. By combining triaxial acceleration data and analyzing the degree of airflow difference and the range of pressure distribution changes, the processing module quantifies the magnitude of the head movement's arc. The control module acquires this information, extracts pulse change characteristics during head movement based on the corresponding acquisition time, and generates targeted suggestions based on the correlation between the two, such as prompting adjustments to the tilting motion or slowing down head turning. Traditional wearable devices rely on single sensors and cannot capture detailed head movements through multi-location airflow and pressure analysis. They lack the processing logic to correlate movement characteristics with physiological signals. Even with multiple sensors, data can only be simply aggregated, making it difficult to achieve precise matching of movement direction, arc, and pulse changes, thus failing to provide realistic exercise guidance and hindering the prevention of sports injuries.

[0020] Furthermore, the hanging ring is equipped with an adjustable magnetic clasp. When the user wears the hanging ring, the data acquisition module first collects the contact pressures and records the magnitude and distribution of the values. The processing module determines whether the wearing is complete based on the preset pressure threshold range and uniform distribution standard. If the wearing is complete, a preset start-of-exercise prompt is played. If the wearing is not complete, the magnetic clasp is controlled to loosen, and a preset adjustment prompt is played again. The airflow data difference between the two sides of the neck is acquired, and the triaxial acceleration change is combined to determine whether the user has completed the adjustment. After the user completes the adjustment, the magnetic clasp is controlled to tighten. Then, the user is checked again to determine whether the wearing is complete, until the wearing is complete.

[0021] When a user is exercising, if the loop is not worn correctly, this solution can objectively determine whether the wearing is standard by collecting the contact pressure value and distribution through pressure sensors. When the pressure value does not reach the preset threshold or the distribution is uneven, it is determined that the wearing is not standard, avoiding the problem of improper wearing caused by relying on the wearer's subjective judgment in existing technologies. At this time, the device automatically prompts the user to adjust, and the magnetic buckle automatically performs loosening and tightening operations, repeatedly adjusting until the pressure value reaches the standard and the distribution is even. The entire process does not require manual operation by the user, which not only improves the convenience and comfort of wearing, but also ensures that signal acquisition is not affected by improper wearing.

[0022] This standardized judgment method based on numerical values ​​provides a stable foundation for subsequent data collection, reduces signal errors caused by wearing issues, and makes physiological data such as pulse and airflow more accurate. Existing technologies, even those with adjustable tightness, lack this quantitative data-based judgment and automatic adjustment process. They either rely on the user's subjective feeling for adjustment, which can easily lead to non-standard wearing, or their adjustment mechanisms are too simple to guarantee the stability of data collection, resulting in lower accuracy, data reliability, and user experience.

[0023] Furthermore, the acquisition module is equipped with a unified clock source, and each airflow sensor acquires airflow data based on this unified clock source. The control module determines whether the user is stationary based on changes in the triaxial acceleration values. When the user is stationary, the acquisition module acquires contact pressure data, airflow data, and pulse data as reference data. It compares the acquired reference data with the stored initial calibration data of the device, analyzes the degree of sensor sensitivity attenuation, and automatically generates compensation parameters based on the attenuation. When the user is not stationary, the processing module uses the compensation parameters to adjust the various data acquired by the acquisition module. The processing module is equipped with a scene recognition model. By analyzing the correlation between the differences in pulse data, triaxial acceleration, and airflow data in historical motion data and the motion type, the scene recognition model identifies the user's current motion type and intensity, matches the stored processing parameters, and adjusts the acquisition frequency of each component in the acquisition module according to the processing parameters.

[0024] This solution ensures synchronized data acquisition from all sensors through a unified clock source, eliminating time discrepancies. When the control module determines the user is stationary based on triaxial acceleration, the acquisition module automatically obtains contact pressure, airflow, and pulse data as a baseline, compares them with initial calibration data, accurately identifies the degree of sensor sensitivity degradation, and dynamically generates compensation parameters. When the user moves, the processing module applies these parameters in real time to adjust the acquired data, ensuring a high degree of match between physiological signals and the actual state. This mechanism proactively addresses the performance degradation of sensors after long-term use, maintaining initial accuracy.

[0025] In scenarios emphasizing the coordination of breathing and movement, the scene recognition model accurately identifies the yoga scene by analyzing the smooth fluctuations in triaxial acceleration (reflecting slow movements), the periodic differences in airflow data (reflecting deep breathing), and a stable pulse frequency. It then adjusts the airflow sensor's sampling frequency to a moderate level, avoiding energy waste from high-frequency sampling while still capturing subtle changes in breathing rhythm, providing accurate data support for breathing-guided exercise prescriptions. Conversely, when facing high-intensity interval training (HIIT), the scene recognition model determines the exercise type based on the dramatic fluctuations in triaxial acceleration, the rapid rise and fall of the pulse, and the rapid changes in airflow. It then instantly adjusts the sampling frequency of each sensor to the highest level, ensuring that no physiological signal mutations are missed within a second (such as peak heart rate during sprints or rapid breathing airflow pulses), providing high-density data for risk warnings during high-intensity exercise (such as over-fatigue alerts).

[0026] This scenario-specific dynamic frequency adjustment achieves a balance between energy consumption and accuracy: the lowest acquisition frequency is used in the resting state to extend the device's standby time; when switching to a motion state, the frequency is quickly increased to ensure data timeliness. This solution sets the acquisition frequency differently based on the physiological signal characteristics of different movements, reducing energy consumption in low-intensity scenarios and maintaining accurate data acquisition in high-intensity scenarios, thus balancing accuracy and efficiency.

[0027] Furthermore, the acquisition module also includes a humidity sensor disposed on the surface of the hanging ring. The humidity sensor is used to collect humidity data in real time in the area where the hanging ring contacts the neck. The processing module analyzes the frequency and numerical change relationship between humidity data, contact pressure data and triaxial acceleration, and determines whether the user is in a swimming state based on the change relationship. If the user is in a swimming state, the operating frequency of the airflow sensor is increased. If the user is determined to be in a daily exercise or resting state, the operating frequency of the airflow sensor is decreased, while the operating frequency of the pulse sensor is increased.

[0028] In swimming scenarios, the impact of water flow causes periodic fluctuations in pressure around the neck, and these fluctuations are highly synchronized with the fluctuations in triaxial acceleration (such as during strokes and turns). Meanwhile, the humidity sensor continuously detects high humidity. This collaborative characteristic of multi-sensor data allows the device to accurately distinguish swimming from other high-humidity scenarios (such as heavy sweating), avoiding misjudgments. Increasing the airflow sensor frequency in this situation allows for the capture of brief and rapid airflow pulses during breathing (e.g., a single breath change lasting only 0.5-1 second), ensuring that key data such as breathing rhythm and breathing intervals are not lost, providing accurate evidence for the "coordination of breathing and strokes" recommendations in swimming prescriptions. In non-swimming scenarios (such as daily running and yoga), the humidity detected by the humidity sensor is mostly a slow increase caused by localized sweat evaporation, and the correlation between pressure changes and acceleration is weaker. At this point, reducing the airflow sensor frequency (e.g., from 50Hz during swimming to 10Hz) can reduce the impact of interfering airflow caused by sweat evaporation on the data; at the same time, increasing the pulse sensor frequency can more sensitively capture subtle fluctuations in heart rate during exercise (e.g., the slow decrease in heart rate during yoga meditation), balancing data accuracy and energy consumption control.

[0029] Furthermore, the processing module acquires real-time triaxial acceleration data, adjusts the acquisition frequency of the triaxial acceleration based on the values, and monitors changes in the direction and angle of the acceleration vector. When an abnormal shift in the lateral acceleration vector is detected, it combines the contact pressure distribution data acquired by the pressure sensor to determine whether the neck tilt angle is greater than a preset angle. If it is greater than the preset angle, it generates and plays an exercise suggestion to "adjust head posture to avoid muscle strain." Simultaneously, it performs integral calculations on the triaxial acceleration, accumulates the changes, and combines them with the accumulated changes in pulse data to calculate the increase in the user's exercise intensity. The control module adjusts the exercise prescription suggestion based on the rate of increase in exercise intensity and the increased exercise intensity.

[0030] By dynamically adjusting the triaxial acceleration sampling frequency, the sampling density can be automatically increased when the user's movement changes drastically (such as rapid turns or jumps), and decreased during gentle movements. This ensures monitoring accuracy while reducing unnecessary energy consumption and extending device battery life. For real-time neck posture monitoring, combined with pressure distribution data, a multi-dimensional judgment is formed. Compared to single acceleration monitoring, it can more accurately identify poor postures such as looking down or tilting to the side. Its warning response speed covers various exercise scenarios such as yoga and running, avoiding the risk of injury due to delays. In terms of exercise intensity assessment, the fusion calculation of acceleration integrals and pulse data accurately reflects the user's actual exercise load (e.g., distinguishing the intensity difference between brisk walking and jogging). The control module dynamically adjusts the prescription based on intensity changes; for example, it promptly suggests slowing down when the user's exercise intensity suddenly increases, and recommends increasing the range of motion when the intensity is insufficient, achieving closed-loop optimization from monitoring to guidance.

[0031] Furthermore, the acquisition module collects pulse data in real time, and the processing module first extracts the peak interval of the pulse to calculate the real-time heart rate. Combined with the amplitude of heart rate changes, it helps to judge the current exercise intensity and provides data support for the assessment of exercise intensity. At the same time, the processing module analyzes the amplitude change characteristics of the pulse waveform, obtains the changes in pulse data within a preset time period after the end of exercise based on triaxial acceleration, assesses vascular elasticity based on the changes in pulse data, and incorporates the assessment results into the health screening results, adding a health assessment dimension to the exercise prescription.

[0032] This solution forms a complete closed loop for monitoring pulse changes by collecting pulse data in real time during exercise and during the rest and recovery period after exercise. This significantly improves the depth of health assessment and the accuracy of prescription recommendations. During exercise, the real-time heart rate and its amplitude calculated through peak intervals dynamically reflect the body's tolerance to the current exercise intensity, providing a basis for timely adjustments to the exercise pace. During the preset rest period after exercise, the recovery trend of pulse data (such as the time it takes to drop from the peak exercise level to the resting level and the stability of fluctuations) has higher health reference value. Users with different health conditions show significant differences in pulse recovery during this stage. For example, those with stronger cardiopulmonary function recover faster and with smaller fluctuations, while those with potential cardiovascular burden may recover more slowly and have disordered waveforms.

[0033] This dual-stage data collection, encompassing both exercise and recovery periods, allows the device to comprehensively assess a user's physical reserves and physiological regulation capabilities. If the pulse stabilizes rapidly during recovery, it indicates good physical adaptation, suggesting an appropriate increase in exercise intensity for the next session. Conversely, slow recovery with abnormal waveforms indicates a need to reduce intensity and increase adaptive training. In contrast, existing technologies only collect pulse data during exercise or record resting heart rate, failing to capture real-time endurance during exercise and lacking assessment data during the crucial recovery period, leading to incomplete health status assessments. This solution, through collaborative analysis of data from both time periods, improves the accuracy of health assessments and provides a more comprehensive physiological basis for developing personalized exercise prescriptions. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of the functional module of a health screening system based on pulse diagnosis in an embodiment of the present invention. Detailed Implementation

[0035] The following will describe the concept and technical effects of the present invention clearly and completely with reference to embodiments, so as to fully understand the purpose, features and effects of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are all within the scope of protection of the present invention.

[0036] like Figure 1 As shown, a health screening system for exercise prescription based on pulse diagnosis includes:

[0037] The data acquisition module includes a hanging ring made of silicone; a pulse sensor, which is a piezoelectric sensor, is installed inside the hanging ring; multiple airflow sensors, which are MEMS microphones, are fixed to the right side of the pulse sensor inside the hanging ring; pressure sensors are evenly distributed on the surface of the hanging ring that contacts the neck; a gyroscope for acquiring three-axis acceleration is installed inside the hanging ring; the pulse sensor is used to acquire pulse data from the user's neck; the airflow sensors are used to acquire airflow data around the neck; and the pressure sensors are used to acquire contact pressure data at various points on the user's body.

[0038] The processing module acquires contact pressure data and triaxial acceleration. When the triaxial acceleration is not greater than a preset static threshold and the contact pressure data is within a preset wearing range, it acquires pulse data and airflow data. It calculates the ratio of pulse data to airflow data as the pulse-breath ratio by combining the acquisition time. The processing module stores the types of diseases and the corresponding pulse-breath ratio ranges and pulse ranges. The processing module uses pulse data and pulse-breath ratios to match the corresponding disease types.

[0039] The regulation module acquires disease type, pulse-to-breath ratio, and pulse data, and combines them to generate exercise prescription suggestions.

[0040] The processing module also analyzes the stable movement time by combining the contact pressure value and triaxial acceleration with the acquisition location and acquisition time when the triaxial acceleration is greater than the preset static threshold and the contact pressure data is within the preset wearing range. Based on the differences in airflow data at different acquisition locations during the stable movement time, it separates the breathing airflow and movement airflow from the ambient airflow. Then, it combines the breathing airflow, movement airflow, pulse data, and contact pressure data to calculate the exercise intensity and construct an exercise muscle function model. The control module obtains the exercise intensity and exercise muscle function model to generate exercise prescription suggestions.

[0041] In the process of constructing the exercise muscle function model, the processing module analyzes the variation of contact pressure values ​​at different acquisition locations and in terms of numerical values. Combined with the acquisition time and variation of triaxial acceleration, the stable exercise time is obtained. Airflow data at various locations during the stable exercise time are acquired. Independent component analysis is used to separate the chest vibration airflow and ambient airflow from the airflow data. Based on the influence of chest vibration on respiration and the numerical differences in airflow data acquired at different locations, respiratory airflow and exercise airflow are separated from the ambient airflow. The exercise airflow is combined with triaxial acceleration to calculate the exercise intensity. Pulse data during the stable exercise time is obtained. The contact pressure data acquired simultaneously with respiratory airflow, chest vibration airflow, and pulse data are combined to construct the exercise muscle function model.

[0042] The pendant is equipped with airflow sensors at the back of the neck and the corresponding neck position below the chin. When processing airflow data, the processing module first collects airflow data at the back of the neck and the chin, respectively, and records the difference in the collection time between the two data points. Then, combining the periodic correspondence and synchronous change characteristics between the chest vibration airflow and the respiratory airflow, the respiratory airflow is separated from the airflow data collected at the chin. Subsequently, the difference between the airflow data collected at the back of the neck and the chin is compared to perform a second separation of the separated respiratory airflow. Finally, the respiratory airflow after the second separation is compared with the chest vibration airflow to exclude the respiratory-related airflow components from the airflow data. The remaining part is the motion airflow.

[0043] In one embodiment of this solution, the learner's breathing status can be accurately monitored through airflow sensors at the back of the neck and jaw, along with multi-step processing, to help optimize the exercise rhythm. The specific implementation steps and related models are as follows: The data collected by the airflow sensor at the back of the neck is denoted as Q. h (t), the data collected by the airflow sensor at the lower jaw is recorded as Q. j (t), recording the time difference Δt = t between the two data collection points. j -t h (t j t represents the data acquisition time at the lower jaw. h (Data collection time at the back of the neck). During user exercise training, an airflow intensity difference model D(t) = |Q is constructed. j (t)-Q h The formula (t-Δt)| is used to distinguish between respiratory and kinetic airflow. Here, D(t) is the intensity difference between the two airflow points at time t after time calibration. When the user breathes, the airflow intensity at the jaw increases sharply, and D(t) will show a significant peak. The change in D(t) caused by kinetic airflow (airflow data collected at the back of the neck) is relatively gradual.

[0044] Specifically, airflow sensors are installed on both sides of the neck below the ears. The processing module analyzes the head movement direction and curvature based on the differences in airflow data collected by the airflow sensors on both sides. The implementation steps are as follows:

[0045] First, airflow data from both sides is collected using airflow sensors located on either side below the ears, and the differences in the data between the two sides are recorded.

[0046] By combining the changes in the pressure sensor readings around the neck, we can analyze the distribution pattern and trend of the pressure data.

[0047] By correlating and matching the differences in airflow between the two sides with changes in pressure distribution, the direction of the user's head movement can be identified.

[0048] The magnitude of the head motion arc is analyzed by combining the degree of airflow difference and the range of pressure distribution changes with triaxial acceleration.

[0049] The control module acquires information on the direction and curvature of head movement, and extracts pulse change characteristics during head movement in conjunction with the corresponding acquisition time; based on the correlation between pulse change characteristics and head movement, it generates exercise prescription suggestions for head movement.

[0050] In one embodiment of this solution, the airflow sensors on both sides of the neck below the ears and the related processing flow can be combined with the following model to achieve accurate analysis of head movements: Let the data collected by the airflow sensor on the left side below the ears be Q. z (t), with Q on the right. y (t), the time difference in data acquisition is Δt′=t y -t z , where t y The acquisition time on the right side, t z The data collection time is for the left side. A head motion airflow difference model is constructed: D′(t)=|Q y (t)-Q z (t-Δt′)|, when the head turns to the left, the airflow on the left is compressed and strengthened, and D′(t) shows a negative fluctuation; when turning to the right, D′(t) shows a positive fluctuation, thus distinguishing the direction of movement. Simultaneously, the neck pressure distribution data collected by the pressure sensor is denoted as P(t,s), where s is the coordinate of different neck positions. The overall pressure change amplitude is calculated using the pressure distribution change rate model K(t)=∑|P(t,s)-P(t-1,s)|. The larger the K(t) value, the greater the head movement arc. By correlating D′(t), K(t), and the triaxial acceleration data A(t), a motion arc evaluation model R(t)=β·|D′(t)|+γ·K(t)+δ·|A(t)| (β, γ, and δ are weighting coefficients, determined by fitting head movement experimental data, with values ​​ranging from 0.2 to 0.4), is constructed. The R(t) value is positively correlated with the motion arc.

[0051] The control module extracts pulse data B(t) during head movement and calculates the pulse fluctuation coefficient F = max(B(t)) - min(B(t)) during the movement period. When F exceeds a preset threshold and is positively correlated with R(t), targeted suggestions are generated, such as "The current head rotation is too large, resulting in significant pulse fluctuation; it is recommended to reduce the rotation amplitude." In this model, Δt′ is selected based on the propagation characteristics of airflow on both sides of the neck, consistent with the time calibration logic; D′(t) and D(t) are both expressed as airflow intensity differences, focusing on head movement and respiratory airflow respectively through differential analysis of sensors at different locations.

[0052] The hanging loop is equipped with an adjustable magnetic clasp. When the user wears the loop, the data acquisition module first collects the contact pressure and records the magnitude and distribution of the values. The processing module determines whether the wearing is complete based on the preset pressure threshold range and uniform distribution standard. If the wearing is complete, a preset start-of-exercise prompt is played. If the wearing is not complete, the magnetic clasp is controlled to loosen, and a preset adjustment prompt is played. The module also acquires the airflow data difference between the two sides of the neck and combines it with the triaxial acceleration change (when the airflow data difference tends to stabilize and the triaxial acceleration change amplitude is lower than the preset threshold, it is determined that the user has completed the adjustment) to determine whether the user has completed the adjustment. After the user completes the adjustment, the magnetic clasp is controlled to tighten. Then, the system checks again whether the wearing is complete. If it is still not complete, the above loosening, prompting, adjustment, and tightening process is repeated until the wearing is complete.

[0053] The acquisition module is equipped with a unified clock source, and each airflow sensor acquires airflow data based on the unified clock source. The control module determines whether the user is stationary based on the changes in the triaxial acceleration values. When the user is stationary, the acquisition module acquires contact pressure data, airflow data, and pulse data as reference data, compares the acquired reference data with the stored initial calibration data of the device, analyzes the degree of sensor sensitivity attenuation, and automatically generates compensation parameters based on the degree of attenuation. When the user is not stationary, the processing module uses the compensation parameters to adjust the various data acquired by the acquisition module.

[0054] The processing module includes a scene recognition model. By analyzing the correlation between the differences in pulse data, triaxial acceleration, and airflow data in historical motion data and the type of motion, the scene recognition model identifies the user's current type and intensity of motion, matches the stored processing parameters, and adjusts the acquisition frequency of each component in the acquisition module according to the processing parameters.

[0055] In one embodiment of this solution, the acquisition module synchronously acquires airflow data, pressure sensor data, and triaxial acceleration data from the back of the neck, jaw, and below the ears at a base frequency (e.g., 10Hz), while also recording the pulse waveform. The acquired data is then filtered using a sliding window (window size defaulted to 2 seconds) to remove motion artifacts and environmental noise, and feature vectors (such as airflow fluctuation amplitude, pressure change gradient, and acceleration variance) are extracted.

[0056] The scene recognition model has six built-in templates for typical exercise modes: resting, walking, running, yoga, swimming preparation, and high-intensity interval training (HIIT).

[0057] The Dynamic Time Warping (DTW) algorithm is used to calculate the similarity between the real-time feature vector and each template. When the matching degree of a template exceeds 85% and the duration exceeds 10 seconds, it is determined to be the corresponding motion type.

[0058] Exercise intensity is graded by combining the root mean square value (RMS) of triaxial acceleration with pulse variability (HRV). For example, low intensity: RMS < 0.5g and HRV > 60ms; moderate intensity: 0.5g ≤ RMS < 1.2g and 30ms ≤ HRV ≤ 60ms; high intensity: RMS ≥ 1.2g and HRV < 30ms.

[0059] The processing parameters stored in the acquisition module are shown in Table 1 below:

[0060]

[0061]

[0062] Table 1

[0063] When identified as "Running (Moderate Intensity)," the airflow sensor frequency is automatically increased from the base 10Hz to 20Hz, and the pressure sensor frequency is increased from 5Hz to 10Hz to ensure that rapid breathing changes can be captured. If the exercise intensity is upgraded to high intensity (such as switching from jogging to sprinting), the frequencies of each sensor are further increased to the HIIT mode parameters.

[0064] At the end of each exercise and upon entering a resting state, 30 seconds of baseline data are collected and compared with the initial calibration value. If the deviation of a certain sensor exceeds the threshold (e.g., the sensitivity of the airflow sensor decreases by >10%), compensation parameters are calculated and the amplification factor is dynamically adjusted.

[0065] In the resting state, non-critical sensors (such as triaxial accelerometers) sleep for 1 minute every 5 minutes; during low-intensity exercise, the sensors operate at the frequency specified in the parameter table; during high-intensity exercise, all sensors operate at full load, but data is only cached for the most recent 30 seconds; when a long period of stillness is detected (default is more than 30 minutes), it automatically enters power-saving mode to reduce overall power consumption.

[0066] The acquisition module also includes a humidity sensor on the surface of the hanging ring, which is used to collect humidity data in real time in the area where the hanging ring contacts the neck; the processing module analyzes the frequency and numerical change relationship between humidity data, contact pressure data and triaxial acceleration, and determines whether the user is swimming based on the change relationship; if swimming, the operating frequency of the airflow sensor is increased; if it is determined to be daily exercise or resting, the operating frequency of the airflow sensor is decreased and the operating frequency of the pulse sensor is increased.

[0067] Specifically, the acquisition module also includes a flexible humidity sensor array embedded in the contact surface of the hanging ring. The humidity sensor is fabricated using a micro / nano structured hydrophilic membrane and an ion-conducting polymer composite, and is used to collect real-time humidity gradient distribution data in the contact area between the hanging ring and the neck. The processing module uses a multi-sensor data fusion algorithm to analyze the temporal rate of change of humidity data, the fluctuation frequency characteristics of contact pressure data, and the spatial vector distribution of triaxial acceleration to construct a swimming state feature recognition model. A swimming state is determined when the model simultaneously meets the following conditions:

[0068] Humidity data must consistently exceed 90% RH for 3 seconds with a change rate ≥ 0.5% RH / second;

[0069] The contact pressure data exhibits periodic fluctuations, with a frequency in the range of 0.5-2Hz and an amplitude variation of ≥30%.

[0070] The Z-axis component of the triaxial acceleration accounts for ≥60% and the rate of change of the angle between the XY plane vectors is ≥15° / second.

[0071] If the system determines that the user is swimming, the processing module will increase the sampling frequency of the airflow sensor to 100Hz and activate the waterproof mode (i.e., turn off unnecessary sensors). If the system determines that the user is exercising or resting, the airflow sensor frequency will be reduced to 10Hz, while the sampling frequency of the pulse sensor will be increased to 200Hz, and the heart rate variability (HRV) monitoring mode will be activated.

[0072] The processing module acquires real-time triaxial acceleration data, adjusts the acquisition frequency of the triaxial acceleration based on the values, and monitors changes in the direction and angle of the acceleration vector. When an abnormal shift in the lateral acceleration vector is detected, it combines the contact pressure distribution data acquired by the pressure sensor to determine whether the neck tilt angle is greater than a preset angle. If it is greater than the preset angle, it generates and plays an exercise suggestion to "adjust head posture to avoid muscle strain." Simultaneously, it performs integral calculations on the triaxial acceleration, accumulates the changes, and combines them with the accumulated changes in pulse data to calculate the increase in the user's exercise intensity. The control module adjusts the exercise prescription suggestion based on the rate of increase in exercise intensity and the increased exercise intensity.

[0073] Specifically, the processing module uses the Kalman filter algorithm to denoise the real-time acquired triaxial accelerations and dynamically adjusts the sampling frequency based on the root mean square (RMS) value of the acceleration.

[0074] When RMS < 0.5g, the sampling frequency is set to 10Hz (quiet mode);

[0075] When 0.5g≤RMS<2g, the sampling frequency is set to 50Hz (daily exercise mode);

[0076] When RMS ≥ 2g, the sampling frequency is set to 100Hz (intense exercise mode).

[0077] In one embodiment of this solution, the processing module monitors the quaternion change of the acceleration vector in real time. When the Euler angle offset of the lateral acceleration vector is detected to be >15° in the X-axis direction and the duration is >2 seconds, the contact pressure distribution data of the pressure sensor is called to verify the posture: if the pressure distribution unevenness coefficient in the cervical spine area is >0.6 and the center of gravity offset is >15mm, it is determined that the neck tilt angle exceeds the limit; 3D spatial audio prompts (such as "Please pay attention to neck posture") are played through the bone conduction speaker, and at the same time, the vibration motor outputs pulse vibration with a frequency of 160Hz.

[0078] The processing module performs double integration on the triaxial acceleration, extracts motion features using wavelet transform, and couples this with the frequency domain features of the pulse data (such as changes in the LF / HF ratio). It then uses the Dynamic Time Warping (DTW) algorithm to calculate the exercise intensity enhancement index. The regulation module, based on a reinforcement learning model, dynamically adjusts the recommended exercise prescription strategy according to the first derivative (increase velocity) and second derivative (acceleration) of the intensity enhancement index.

[0079] When the increase rate is greater than 0.3 and the acceleration is greater than 0, a message will be sent saying "The current intensity is increasing too fast. It is recommended to reduce the range of motion by 30%".

[0080] When the speed increase is less than -0.2 and the duration is greater than 5 minutes, the message "Insufficient exercise intensity, it is recommended to increase the resistance by 1 level" is pushed.

[0081] When the intensity index remains within ±10% of the target range for 3 consecutive minutes, a positive feedback message is generated: "Maintain the current state and keep going."

[0082] The acquisition module collects pulse data in real time. The processing module first extracts the pulse peak interval time and calculates the real-time heart rate. Combined with the amplitude of heart rate changes, it helps to judge the current exercise intensity and provides data support for the assessment of exercise intensity. At the same time, the processing module analyzes the amplitude change characteristics of the pulse waveform, obtains the changes in pulse data within a preset time period after the end of exercise based on triaxial acceleration, assesses vascular elasticity based on the changes in pulse data, and incorporates the assessment results into the health screening results, adding a health assessment dimension to the exercise prescription.

[0083] In another specific embodiment of this solution, the pulse data assessment mechanism is further integrated with respiratory airflow and chest vibration airflow data to form a multi-physiological signal collaborative analysis system, which significantly improves the accuracy of health assessment.

[0084] After exercise ends and the body enters the recovery period, when the triaxial acceleration is almost constant, the processing module first distinguishes two types of characteristic airflow using an airflow sensor. Exercise-related airflow (such as residual shortness of breath) gradually disappears with rest, while respiratory airflow (a smooth inspiratory-expiratory cycle) and chest vibration airflow (weak airflow fluctuations generated by heartbeats) become the dominant signals. Since the back of the neck is less affected by limb movement and the conduction path of chest vibration airflow is stable at this location, the processing module can separate the pure chest vibration airflow signal using a filtering algorithm (such as low-frequency filtering below 5Hz) while preserving the complete cyclic characteristics of respiratory airflow.

[0085] In a healthy state, the pulse wave amplitude increases slightly during inhalation and decreases slightly during exhalation; this characteristic of "respiratory sinus arrhythmia" is particularly pronounced during the recovery period. If a weakened correlation between the two is detected (such as irregular fluctuations in pulse amplitude during the respiratory cycle), it suggests possible abnormalities in autonomic nervous system regulation.

[0086] The frequency of chest cavity vibration airflow is consistent with the heart rate, and its amplitude changes can reflect cardiac output. During the recovery period, if the pulse rate has dropped to the resting level, but the amplitude of chest cavity vibration airflow is still significantly lower than the baseline value, it suggests that myocardial contractility has not fully recovered and there may be potential cardiac function burden.

[0087] When the pulse decreases from its peak during exercise to the resting level, the respiratory airflow rate also decreases to a calm state (e.g., from 30 breaths / minute to 15 breaths / minute), and the amplitude of chest cavity vibration airflow gradually stabilizes. The consistency of the recovery trends of these three factors can verify the reliability of the health status assessment results. If the pulse has recovered but breathing is still rapid and chest cavity vibration airflow is disordered, it suggests that there may be respiratory muscle fatigue or cardiopulmonary coordination dysfunction.

[0088] The advantage of this multi-signal fusion mechanism lies in the fact that single pulse recovery data may be affected by factors such as emotions and body position (e.g., a sudden rise in pulse due to standing up), while respiration and chest vibration airflow have stronger physiological stability. For example, individuals with good cardiopulmonary function will exhibit a synergistic characteristic of "stable pulse decline + even and slowed breathing + regular and enhanced chest vibration" during the recovery period, while those with insufficient cardiopulmonary reserve often exhibit an abnormal combination of "slow pulse recovery + shallow, rapid, and disordered breathing + weak fluctuations in chest vibration." Through dynamic correlation analysis of the three, the random errors of a single signal can be effectively eliminated, improving the accuracy of health status assessment compared to using only pulse data, and providing a more precise quantitative basis for setting parameters such as "rest duration during recovery" and "next exercise intensity threshold" in exercise prescriptions.

[0089] The above descriptions are merely embodiments of the present invention, and common knowledge regarding specific structures and characteristics is not elaborated upon here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the structure of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements will not affect the effectiveness of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A health screening system for exercise prescriptions based on pulse diagnosis, characterized in that, include: The data acquisition module includes a hanging ring, which is made of silicone. The pendant ring contains a pulse sensor, which is a piezoelectric sensor. To the right of the pulse sensor, multiple airflow sensors, which are MEMS microphones, are fixedly installed. The surface of the pendant ring that contacts the neck is equipped with uniformly distributed pressure sensors. The pendant ring also contains a gyroscope for collecting three-axis acceleration. The pulse sensor collects pulse data from the user's neck, the airflow sensors collect airflow data around the neck, and the pressure sensors collect contact pressure data at various points on the user's body. The processing module acquires contact pressure data and triaxial acceleration. When the triaxial acceleration is not greater than a preset static threshold and the contact pressure data is within a preset wearing range, it acquires pulse data and airflow data. It calculates the ratio of pulse data to airflow data as the pulse-breath ratio by combining the acquisition time. The processing module stores the types of diseases and the corresponding pulse-breath ratio ranges and pulse ranges. The processing module uses pulse data and pulse-breath ratios to match the corresponding disease types. The regulation module acquires disease type, pulse-to-breath ratio, and pulse data, and combines them to generate exercise prescription suggestions. The processing module also analyzes the stable movement time by combining the contact pressure value and triaxial acceleration with the acquisition location and acquisition time when the triaxial acceleration is greater than the preset static threshold and the contact pressure data is within the preset wearing range. Based on the differences in airflow data at different acquisition locations during the stable movement time, it separates the breathing airflow and movement airflow from the ambient airflow. Then, it combines the breathing airflow, movement airflow, pulse data, and contact pressure data to calculate the exercise intensity and construct an exercise muscle function model. The control module obtains the exercise intensity and exercise muscle function model to generate exercise prescription suggestions.

2. The exercise prescription health screening system based on pulse diagnosis as described in claim 1, characterized in that: When the processing module constructs the motor muscle function model, it analyzes the variation of contact pressure values ​​at the acquisition location and in terms of value, and combines the acquisition time and variation of triaxial acceleration to analyze and obtain the stable motion time. Airflow data at various locations during a stable exercise period were acquired. Independent component analysis was used to separate the chest vibration airflow and ambient airflow from the airflow data. Based on the influence of chest vibration on respiration and the numerical differences in airflow data collected at different locations, respiratory airflow and exercise airflow were separated from the ambient airflow. The exercise airflow was combined with triaxial acceleration to calculate the exercise intensity. Pulse data during the stable exercise period were acquired. Contact pressure data collected simultaneously from respiratory airflow, chest vibration airflow, and pulse data were combined to construct an exercise muscle function model.

3. The exercise prescription health screening system based on pulse diagnosis according to claim 2, characterized in that: The hanging ring is equipped with airflow sensors at the back of the neck and the neck corresponding to the lower jaw. When the processing module processes the airflow data, it first collects the airflow data at the back of the neck and the lower jaw respectively, and records the difference in the collection time of the two data. Then, it combines the periodic correspondence and synchronous change characteristics between the chest vibration airflow and the respiratory airflow to separate the respiratory airflow from the airflow data collected at the lower jaw. Subsequently, the differences in airflow data collected from the back of the neck and jaw were compared, and the separated respiratory airflow was further separated. Finally, the respiratory airflow after secondary separation was compared with the chest vibration airflow, and respiratory-related airflow components were excluded from the airflow data. The remaining part was the motion airflow.

4. The exercise prescription health screening system based on pulse diagnosis according to claim 3, characterized in that: The hanging ring is equipped with airflow sensors on both sides of the neck below the ears. The processing module analyzes the head movement direction and curvature based on the differences in airflow data collected by the airflow sensors on both sides. The implementation steps are as follows: First, airflow data from both sides is collected using airflow sensors located on either side below the ears, and the differences in the data between the two sides are recorded. By combining the changes in the pressure sensor readings around the neck, we can analyze the distribution pattern and trend of the pressure data. By correlating and matching the differences in airflow between the two sides with changes in pressure distribution, the direction of the user's head movement can be identified. The magnitude of the head motion arc is analyzed by combining the degree of airflow difference and the range of pressure distribution changes with triaxial acceleration. The control module acquires information on the direction and curvature of head movement, and extracts pulse change features during head movement in conjunction with the corresponding acquisition time. Based on the correlation between pulse change characteristics and head movements, exercise prescription suggestions are generated for head movements.

5. A health screening system for exercise prescription based on pulse diagnosis according to claim 4, characterized in that: The hanging ring is equipped with an adjustable magnetic clasp. When the user wears the hanging ring, the data acquisition module first collects the contact pressure and records the magnitude and distribution of the values. The processing module determines whether the wearing is complete based on the preset pressure threshold range and uniform distribution standard. If the wearing is complete, a preset start-of-exercise prompt is played. If the wearing is not complete, the magnetic clasp is controlled to loosen, and a preset adjustment prompt is played again. The module also acquires the airflow data difference changes on both sides of the neck and combines it with the triaxial acceleration changes to determine whether the user has completed the adjustment. After the user completes the adjustment, the magnetic clasp is controlled to tighten. Then, the system checks again to determine whether the wearing is complete until the wearing is finished.

6. The exercise prescription health screening system based on pulse diagnosis according to claim 5, characterized in that: The acquisition module is equipped with a unified clock source, and each airflow sensor acquires airflow data based on the unified clock source; the control module determines whether the user is in a stationary state based on the change in the triaxial acceleration value. The acquisition module collects contact pressure data, airflow data, and pulse data as reference data when the user is stationary. It compares the collected reference data with the stored initial calibration data of the device, analyzes the degree of sensor sensitivity attenuation, and automatically generates compensation parameters based on the degree of attenuation. The processing module uses the compensation parameters to adjust the various data collected by the acquisition module when the user is not stationary. The processing module includes a scene recognition model. By analyzing the correlation between the differences in pulse data, triaxial acceleration, and airflow data in historical motion data and the type of motion, the scene recognition model identifies the user's current type and intensity of motion, matches the stored processing parameters, and adjusts the acquisition frequency of each component in the acquisition module according to the processing parameters.

7. A health screening system for exercise prescription based on pulse diagnosis as described in claim 6, characterized in that: The acquisition module also includes a humidity sensor on the surface of the hanging ring, which is used to collect humidity data in real time in the area where the hanging ring contacts the neck; the processing module analyzes the frequency and numerical change relationship between humidity data, contact pressure data and triaxial acceleration, and determines whether the user is swimming based on the change relationship; if swimming, the operating frequency of the airflow sensor is increased; if it is determined to be daily exercise or resting, the operating frequency of the airflow sensor is decreased and the operating frequency of the pulse sensor is increased.

8. A health screening system for exercise prescription based on pulse diagnosis according to claim 6, characterized in that: The processing module acquires real-time triaxial acceleration data, adjusts the acquisition frequency of the triaxial acceleration based on the values, and monitors changes in the direction and angle of the acceleration vector. When an abnormal shift in the lateral acceleration vector is detected, it combines the contact pressure distribution data acquired by the pressure sensor to determine whether the neck tilt angle is greater than a preset angle. If it is greater than the preset angle, it generates and plays an exercise suggestion to "adjust head posture to avoid muscle strain." Simultaneously, it performs integral calculations on the triaxial acceleration, accumulates the changes, and combines the accumulated changes in pulse data to calculate the increase in the user's exercise intensity. The control module adjusts exercise prescription recommendations based on the rate of increase in exercise intensity and the increased exercise intensity.

9. A health screening system for exercise prescription based on pulse diagnosis as described in claim 8, characterized in that: The acquisition module collects pulse data in real time, and the processing module first extracts the pulse peak interval time, calculates the real-time heart rate, and combines the heart rate change amplitude to help judge the current exercise intensity, providing data support for the assessment of exercise intensity. Meanwhile, the processing module analyzes the amplitude change characteristics of the pulse waveform, obtains the changes in pulse data within a preset time period after the end of exercise based on triaxial acceleration, assesses vascular elasticity based on the changes in pulse data, and incorporates the assessment results into the health screening results, increasing the health assessment dimension of the exercise prescription.

10. A health screening system method for exercise prescription based on pulse diagnosis, characterized in that: The system described in any one of claims 1-9 is employed.

Images