A method and system for optimizing resonant frequency breathing training

Abstract

The application discloses a method and system for optimizing resonance frequency breathing training, which comprises the following steps: step S1: obtaining a resonance frequency breathing rate detection range and a sliding detection total time length of a user; step S2: generating a gradual breathing guidance animation in which a breathing frequency gradually decreases from a maximum value in the detection range to a minimum value; step S3: collecting physiological signals of the user; step S4: extracting a heartbeat interval sequence to obtain a heartbeat interval time curve; step S5: sequentially defining a plurality of sliding analysis windows on the heartbeat interval time curve to obtain a comprehensive evaluation value sequence; and step S6: obtaining a breathing frequency corresponding to the target sliding analysis window as the resonance frequency breathing rate of the user. The application can automatically, efficiently and accurately determine the resonance frequency breathing rate of the user, thereby solving the problems of a long time consumption, low precision and dependence on subjective experience of professional physicians of an existing ladder type detection method.

Description

Technical Field

[0001] This invention relates to the field of biofeedback breathing training technology, specifically to a method and system for optimizing resonant frequency breathing training. Background Technology

[0002] A healthy heart exhibits complex nonlinear oscillations. Studies have shown that when an adult's respiratory rate approaches the inherent resonance frequency (RF) of their cardiopulmonary system, the greatest heart rate oscillations occur due to the negative feedback mechanism of the baroreflex. Active resonance frequency breathing training is a biofeedback training method that uses spontaneous breathing to induce heart rate variability (HRV) resonance in the cardiopulmonary system, and it has effects such as regulating autonomic nerve tension and relieving emotional stress.

[0003] Currently, the main process of respiratory feedback training proposed in existing research is as follows: professional physicians observe the subject's HRV index, as well as the amplitude and phase of the respiratory waveform in real time. Based on their professional knowledge and clinical experience, they determine the magnitude of the resonant respiratory frequency and then guide the subject to adjust their breathing pattern. Among them, the traditional method for measuring the resonant respiratory frequency is mainly the "step method," which requires the subject to perform breathing training for a fixed duration at several discrete respiratory frequencies (e.g., 6.5 breaths / minute, 6.0 breaths / minute, 5.5 breaths / minute, 5.0 breaths / minute, etc.). Then, the physiological data and HRV-related indicators collected in each round are calculated separately. Finally, the most ideal frequency is selected from these discrete respiratory frequencies as the individual's resonant frequency, and this frequency guides subsequent training.

[0004] Although resonant breathing feedback training is still in its nascent stage, most existing studies have used the traditional "step-by-step" method described above to determine resonant breathing frequency. However, this existing technique has the following shortcomings: 1. Reliance on the subjective experience of professional physicians hinders widespread adoption. Current resonant respiratory rate detection largely depends on personnel with specialized medical knowledge and corresponding equipment, lacking objective quantitative indicators, resulting in high human and material costs and making it difficult to promote and apply widely.

[0005] 2. The method for determining the resonant respiratory rate urgently needs improvement. The existing stepwise method requires measuring the physiological indicators of the subject at multiple discrete respiratory rates to find the resonant frequency. Because it requires measurement one after another, the whole process is time-consuming, has low accuracy, and a poor user experience.

[0006] 3. Other respiratory parameters (such as the inspiratory-to-expiratory ratio) have not been quantitatively limited, and the control of the inspiratory-to-expiratory ratio has not been incorporated into the breathing training system. Summary of the Invention

[0007] To overcome the shortcomings of the prior art, one of the objectives of this invention is to provide a method for optimizing resonant frequency breathing training. This method uses a sliding detection approach that combines progressive breathing guidance animation with a sliding analysis window to automatically, efficiently, and accurately determine the user's resonant frequency breathing rate. This solves the problems of existing step-by-step detection methods being time-consuming, inaccurate, and dependent on the subjective experience of professional physicians.

[0008] The second objective of this invention is to provide an optimized resonant frequency breathing training system, which automatically, efficiently, and accurately determines the user's resonant frequency breathing rate through a sliding detection method that combines progressive breathing guidance animation with a sliding analysis window, thereby solving the problems of existing step-by-step detection methods being time-consuming, inaccurate, and dependent on the subjective experience of professional physicians.

[0009] To achieve one of the objectives of this invention, the following solution is adopted: A method for optimizing resonant frequency breathing training includes the following steps: Step S1: Obtain the user's resonant frequency respiratory rate detection range and total sliding detection time as preset parameters; Step S2: Based on the preset parameters, generate a progressive breathing guidance animation in which the respiratory rate decreases from the maximum value to the minimum value within the detection range in one breath, and record the instantaneous respiratory rate corresponding to each breathing cycle; Step S3: During the user's breathing training by following the progressive breathing guidance animation, the user's physiological signals are collected synchronously to obtain the original physiological signals, and the user's real breathing signals are also collected. Step S4: Preprocess the raw physiological signal and identify heartbeat feature points, extract the heartbeat interval sequence, and generate a curve of heartbeat interval changing with time based on the heartbeat interval sequence to obtain the heartbeat interval time curve; Step S5: According to the preset sliding analysis window parameters, multiple sliding analysis windows are sequentially defined on the heartbeat interval time curve; for each sliding analysis window, a respiratory synchronization quality score is calculated based on at least one of the frequency deviation, phase deviation, period integrity, and amplitude stability between the real respiratory signal and the progressive respiratory guidance animation; if the respiratory synchronization quality score is lower than a preset threshold, the sliding analysis window is downweighted or removed; the heart rate oscillation amplitude in the remaining sliding analysis windows is combined with the respiratory synchronization quality score to calculate the comprehensive evaluation value of each sliding analysis window, thus obtaining a comprehensive evaluation value sequence; Step S6: Determine the target sliding analysis window corresponding to the maximum value in the comprehensive evaluation value sequence, and backtrack to obtain the breathing frequency corresponding to the target sliding analysis window based on the progressive breathing guidance animation, as the user's resonant frequency breathing rate; Step S7: Execute steps S1 to S6 multiple times within a training cycle to obtain multiple effective resonant frequency respiratory rates. Eliminate invalid results with respiratory synchronization quality scores below the threshold, and perform weighted aggregation on the effective resonant frequency respiratory rates to output the stage resonant frequencies on a weekly or monthly scale for updating subsequent respiratory training parameters.

[0010] Further, in step S4, the raw physiological signal is preprocessed, specifically including: filtering the raw physiological signal using a Butterworth bandpass filter, wherein the Butterworth bandpass filter consists of a 1Hz high-pass filter and a 30Hz low-pass filter.

[0011] Further, in step S4, the intercardiac interval sequence is extracted, specifically including: A heartbeat feature point detection algorithm is used to identify heartbeat feature points in preprocessed physiological signals; Calculate the time interval between adjacent heartbeat feature points to obtain the initial heartbeat interval sequence; The initial heartbeat interval sequence is subjected to outlier removal and stabilization processing. Outliers that do not conform to physiological characteristics are replaced with the harmonic mean of their adjacent elements to obtain the heartbeat interval sequence.

[0012] Further, in step S4, generating the heartbeat interval time curve specifically includes: interpolating the heartbeat interval sequence into a continuous time curve using the spline cubic spline interpolation method to obtain the heartbeat interval time curve.

[0013] Furthermore, in step S5, the window length of the sliding analysis window is 1 minute, and the step size for each slide is 15 seconds, with the step size corresponding to the movement distance of the center point of the sliding analysis window.

[0014] Further, in step S5, calculating the comprehensive evaluation value within each sliding analysis window specifically includes: extracting the heart rate peak and heart rate trough within the current sliding analysis window based on the heart rate interval curve, calculating the stable difference between the heart rate peak and the heart rate trough, and using the stable difference as a component of the comprehensive evaluation value within the current sliding analysis window.

[0015] Furthermore, in step S6, after determining the resonant frequency respiratory rate, the step of matching the inspiratory-to-expiratory ratio is also included: The user's breathing rate is fixed at the resonant frequency breathing rate, and multiple preset inspiratory-to-expiratory ratio modes are switched sequentially, generating corresponding inspiratory-to-expiratory ratio guidance animations. While the user follows the breathing training animation for each inspiratory-to-expiratory ratio, the user's physiological signals are collected synchronously, and the heart rate variability index is calculated based on the collected physiological signals for each inspiratory-to-expiratory ratio mode. The maximum value among the various heart rate variability indicators is selected, and the inspiratory-to-expiratory ratio corresponding to the maximum value is determined as the optimal inspiratory-to-expiratory ratio for the user.

[0016] Furthermore, step S3 also includes synchronously collecting the user's breathing signals; the breathing signals are used for real-time interactive feedback during breathing training to help the user synchronize with the progressive breathing guidance animation.

[0017] Furthermore, in step S3, the synchronously acquired user's respiratory signal is preprocessed, specifically including filtering the respiratory signal using a Butterworth filter with a passband range of 0.03Hz to 0.7Hz.

[0018] To achieve the second objective of this invention, the following solution is adopted: A breathing training system for optimizing resonant frequency includes: The parameter preset module is used to obtain the user's resonant frequency respiratory rate detection range and total sliding detection time as preset parameters; An animation generation module is used to generate a progressive breathing guidance animation based on the preset parameters, in which the breathing rate decreases from the maximum value to the minimum value within the detection range in one breath, and to record the instantaneous breathing rate corresponding to each breathing cycle. The physiological acquisition module is used to synchronously acquire the user's physiological signals to obtain raw physiological signals and acquire the user's real breathing signals during the breathing training process while the user follows the progressive breathing guidance animation. The signal processing module is used to preprocess the raw physiological signal and identify heartbeat feature points, extract the heartbeat interval sequence, and generate a curve of the heartbeat interval changing over time based on the heartbeat interval sequence to obtain the heartbeat interval time curve. The sliding analysis module is used to sequentially delineate multiple sliding analysis windows on the heartbeat interval time curve according to preset sliding analysis window parameters. For each sliding analysis window, a respiratory synchronization quality score is calculated based on at least one of the following: frequency deviation, phase deviation, period integrity, and amplitude stability between the actual respiratory signal and the progressive respiratory guidance animation. If the respiratory synchronization quality score is lower than a preset threshold, the sliding analysis window is downweighted or removed. The heart rate oscillation amplitude within the remaining sliding analysis windows is combined with the respiratory synchronization quality score to calculate the comprehensive evaluation value of each sliding analysis window, resulting in a comprehensive evaluation value sequence. The resonance frequency determination module is used to determine the target sliding analysis window corresponding to the maximum value in the comprehensive evaluation value sequence, and to backtrack to obtain the breathing frequency corresponding to the target sliding analysis window based on the progressive breathing guidance animation, which is used as the user's resonance frequency breathing rate.

[0019] The long-term treatment management module is used to call the parameter preset module, animation generation module, physiological acquisition module, signal processing module, sliding analysis module, and resonance frequency determination module multiple times within a training cycle to obtain multiple effective resonance frequency respiratory rates. Invalid results with respiratory synchronization quality scores below the threshold are eliminated, and the effective resonance frequency respiratory rates are weighted and aggregated to output the stage resonance frequencies on a weekly or monthly scale for updating subsequent respiratory training parameters.

[0020] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention achieves automated detection of resonant frequency respiratory rate, reducing reliance on professional physicians. Specifically, this invention automatically determines the user's resonant frequency respiratory rate by acquiring user-preset parameters, generating progressive breathing guidance animations, simultaneously acquiring electrocardiogram signals, and performing sliding analysis window processing. The entire detection process does not require real-time observation or subjective judgment by professional physicians, significantly reducing labor costs and technical barriers, and facilitating its widespread application among ordinary users.

[0021] 2. This invention significantly shortens the detection time for resonant frequency respiratory rate, improving detection efficiency and accuracy. Compared to the cumbersome process of the traditional "step method," which requires users to perform fixed-duration training and successive measurement calculations at multiple discrete respiratory frequencies, this invention adopts a continuous and progressive sliding detection method for respiratory frequency. Users only need to follow the animation to complete the respiratory training once, and the system can lock the resonant frequency by sequentially scanning through the sliding analysis window, avoiding multiple interruptions and repeated tests, greatly shortening the detection time. At the same time, because the respiratory frequency changes continuously rather than discretely, the detection resolution is higher and the accuracy is better.

[0022] 3. The detection process of this invention is more in line with physiological continuity, improving the user experience. In this invention, the respiratory rate smoothly decreases from its maximum value to its minimum value with each breath, avoiding the discomfort caused by sudden changes in respiratory rate in the step method. The user training process is natural and smooth, and with the help of animation guidance, the user experience is significantly improved.

[0023] 4. This invention provides a precise resonant frequency basis for subsequent personalized breathing training. The resonant frequency respiratory rate determined by this invention is obtained by backtracking from the maximum value of the heart rate variability assessment value, which has an objective quantitative basis and provides a precise frequency benchmark for subsequent inspiratory-to-expiratory ratio matching and long-term breathing training, thus helping to improve the overall effect of breathing training.

[0024] 5. The results of the sliding method of this invention are highly consistent with those of the step method. Experiments show that in 80% of the subjects, the deviation between the resonant frequency measured by the sliding method and the result of the step method is within 0.5 times / minute, verifying the accuracy and reliability of this method.

[0025] 6. The invention, through long-term treatment management, can dynamically adjust the user's resonant frequency training parameters based on multiple test results, adapting to changes in different dates, physical conditions, and training stages, achieving a truly personalized long-term breathing training treatment, and avoiding the invalidation of single test results due to user status drift. Attached Figure Description

[0026] Figure 1 This is a flowchart of the method for optimizing resonant frequency breathing training in an embodiment of the present invention; Figure 2 This is a framework diagram of the method for optimizing resonant frequency breathing training in an embodiment of the present invention; Figure 3 This is a first illustration of the software program function for optimizing the resonant frequency breathing training method in an embodiment of the present invention; Figure 4 This is a second illustration of the software program function for optimizing the resonant frequency breathing training method in an embodiment of the present invention; Figure 5 This is a third illustration of the software program function used to optimize the resonant frequency breathing training method in an embodiment of the present invention; Figure 6 This is a schematic diagram illustrating the principle of the sliding method in an embodiment of the present invention; Figure 7 This is a flowchart of the sliding detection of resonant frequency breathing rate in an embodiment of the present invention; Figure 8 This is a diagram showing the results of a comparative experiment on the resonant frequencies of the step method and the sliding method in an embodiment of the present invention. Figure 9 This is a flowchart illustrating the long-term RF determination and periodic fine-tuning in an embodiment of the present invention; Figure 10 This is a block diagram of a breathing training system for optimizing resonant frequency in an embodiment of the present invention. Detailed Implementation

[0027] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments. It should be noted that, without conflict, the various embodiments or technical features described below can be arbitrarily combined to form new embodiments.

[0028] Example 1 like Figure 1-8 As shown, this embodiment of the invention provides a method for optimizing resonant frequency breathing training, including the following steps: Step S1: Obtain the user's resonant frequency respiratory rate detection range and total sliding detection time as preset parameters.

[0029] Step S2: Based on the preset parameters, generate a progressive breathing guidance animation in which the breathing rate decreases from the maximum value to the minimum value within the detection range one breath at a time, and record the instantaneous breathing rate corresponding to each breathing cycle.

[0030] Step S3: During the user's breathing training by following the progressive breathing guidance animation, the user's physiological signals are collected synchronously to obtain the original physiological signals, and the user's real breathing signals are also collected.

[0031] Step S4: Preprocess the raw physiological signal and identify heartbeat feature points, extract the heartbeat interval sequence, and generate a curve of heartbeat interval changing with time based on the heartbeat interval sequence to obtain the heartbeat interval time curve.

[0032] Step S5: According to the preset sliding analysis window parameters, multiple sliding analysis windows are sequentially defined on the heartbeat interval time curve; for each sliding analysis window, a respiratory synchronization quality score is calculated based on at least one of the frequency deviation, phase deviation, period integrity, and amplitude stability between the real respiratory signal and the progressive respiratory guidance animation. If the respiratory synchronization quality score is lower than a preset threshold, the sliding analysis window is downweighted or removed; the heart rate oscillation amplitude in the remaining sliding analysis windows is combined with the respiratory synchronization quality score to calculate the comprehensive evaluation value of each sliding analysis window, thus obtaining a comprehensive evaluation value sequence.

[0033] Step S6: Determine the target sliding analysis window corresponding to the maximum value in the comprehensive evaluation value sequence, and backtrack to obtain the breathing frequency corresponding to the target sliding analysis window based on the progressive breathing guidance animation, as the user's resonant frequency breathing rate.

[0034] Step S7: Execute steps S1 to S6 multiple times within a training cycle to obtain multiple effective resonant frequency respiratory rates. Eliminate invalid results with respiratory synchronization quality scores below the threshold, and perform weighted aggregation on the effective resonant frequency respiratory rates to output the stage resonant frequencies on a weekly or monthly scale for updating subsequent respiratory training parameters.

[0035] In this invention, the physiological signal refers to any biological signal from which heart rate variability (HRV) indicators can be extracted. Those skilled in the art will recognize that any signal from which an inter-cardiac interval sequence can be extracted can serve as the physiological signal source for this invention, such as photoplethysmography (PPG), cardiac impulse mapping (BCG), magnetocardiography (MCG), and phonocardiography (PCG).

[0036] When photoplethysmography (PPG) is used, pulse wave signals from the fingertips or earlobes can be collected using photoelectric sensors (such as pulse oximeters). Then, peak detection algorithms are used to identify pulse wave peaks as heartbeat feature points. The time interval between adjacent peaks is the pulse interval sequence. Since the pulse interval and RR interval are highly consistent, this sequence can be directly used for subsequent HRV analysis.

[0037] When using cardiac impaction signals, weak mechanical vibration signals caused by the beating of the human heart can be collected by devices such as pressure sensors, accelerometers, or piezoelectric films. Then, template matching or adaptive thresholding algorithms are used to detect feature points such as the J wave in the BCG signal, extract the inter-heartbeat sequence, and then perform HRV assessment.

[0038] Regardless of the physiological signals used, the core is to guide the user to continuously change their breathing rate through progressive breathing guidance animation, while extracting the heart rate interval curve from the collected physiological signals, and then using a sliding analysis window to calculate the HRV assessment value in each window in turn. Finally, the resonant frequency respiratory rate is determined by backtracking based on the maximum value of the HRV assessment value.

[0039] The following is a more detailed description of the breathing training method for optimizing resonant frequency according to an embodiment of the present invention.

[0040] like Figure 2 As shown, this invention enables intelligent detection of resonant frequency and matching of appropriate inspiratory-to-expiratory ratio for breathing training. The design of this invention comprises both hardware and software components. The subject wears ECG electrode patches, and the ECG signal is transmitted via leads to the ADS1292R module for preprocessing, before being transmitted to the system software in the computer. The software can display, process, and analyze the subject's physiological signals in real time and guide the subject's breathing training process through animation. The software has the function of detecting resonant frequency and matching inspiratory-to-expiratory ratio, allowing the subject to adjust their breathing training parameters accordingly.

[0041] The hardware solution of this invention is as follows: The ECG signal preprocessing module used in this invention is the ADS1292R. The ADS1292R is an integrated signal conditioning module for ECG and bio-respiratory signal measurement applications. This device is designed to extract, amplify, and filter weak bioelectrical signals in the presence of noise generated by remote electrode placement. The device has a high common-mode rejection ratio of 120dB, which allows the chip to effectively filter noise and acquire output signals. However, since the ADS1292R does not have a dedicated 50Hz notch filter, it is necessary to avoid environments with strong current when using the ADS1292R for ECG signal measurement. At the same time, the ADS1292R also integrates a respiratory impedance measurement module, which can directly measure respiratory signals. The measured signal is sampled at 500sps by the internal STM32 chip and converted into a digital signal output, which is then connected to an FT232 for USB-to-serial conversion.

[0042] The software design of this invention is described below: This invention develops an intelligent biofeedback breathing training program on the QT platform based on an optimized breathing training scheme. QT is a cross-platform C++ application framework that can run on desktop, embedded, and mobile platforms, and can be well integrated with personalized resonant frequency breathing feedback training schemes to complete a biofeedback training system based on resonant breathing frequency. It allows for the setting of multiple parameters according to user needs, uses animation to guide users in breathing training, and provides real-time feedback on the subject's electrocardiogram signals, respiratory signals, heart rate, respiratory rate, and heart rate variability. After training, the algorithm can detect the user's resonant frequency and inspiratory-to-expiratory ratio.

[0043] In this embodiment, the specific details of the real-time signal feedback are as follows: The software incorporates a self-developed serial communication program for the ADS1292R module, receiving ECG data from the serial port and displaying it in real time. Simultaneously, it filters the ECG signal in the background using a Butterworth bandpass filter (0.05~30Hz). The program detects the R-wave position and amplitude, and calculates the real-time heart rate from the RR interval. The HRV score is expressed as the difference between the maximum and minimum heart rate over a period of time (HRmax-HRmin), reflecting the amplitude of heart rate oscillations, i.e., an indicator of heart rate variability. Respiratory signals can be obtained through synchronous acquisition or fitted from the R-wave amplitude in the ECG signal.

[0044] In this embodiment, the specific details of resonant frequency respiratory rate detection are as follows: Resonance frequency detection can employ discrete stepwise and sliding methods. The former is a traditional technique for resonance frequency respiratory rate detection, while the latter, the sliding method, is a key new approach used in this invention for efficiently extracting resonance frequency respiratory rate.

[0045] (1) Separate stepwise method: Under the premise of a fixed inspiratory-to-expiratory ratio, the program uses animation to guide the subject to perform breathing training according to a set breathing rate stepwise method (such as 6.5, 6.0, 5.5, 5.0, 4.5 breaths / minute). The breathing rate at the maximum value of each HRV is updated to the resonant frequency of the subject.

[0046] (2) Sliding method: Based on the characteristics of the subject, the detection range (maximum respiratory rate and minimum respiratory rate) and the total sliding detection time are set. During the detection process, the respiratory rate guided by the animation will gradually decrease from the maximum respiratory rate to the minimum respiratory rate. The program automatically identifies the subject's resonant frequency according to the algorithm for sliding detection of resonant frequency respiratory rate.

[0047] In this embodiment, the specific details of the inspiratory-to-expiratory ratio matching are as follows: After determining the resonant breathing frequency, subjects undergo breathing training guided by animation under preset inspiratory-to-expiratory ratio (IVR) modes. The system provides real-time feedback on heart rate velocity (HRV) and records the IVR corresponding to the maximum HRV value, updating it to the subject's optimal IVR. Subjects can also choose to store their electrocardiogram (ECG) data locally, and the program's built-in algorithm analyzes the HRV amplitude to confirm the subject's latest optimal IVR.

[0048] The algorithm design of this invention is described below: In detecting resonant frequency respiratory rate, this invention primarily employs a sliding RF detection method. The main process is as follows: By setting a progressively changing respiratory rate as a "visual guide," the duration of a single respiratory cycle is gradually increased (reducing the respiratory rate) during training. The heart rate variability (HRV) is calculated by sequentially scanning through a sliding analysis window within the window, corresponding to the respiratory state. As the analysis window slides sequentially, the scanning range passes through the resonant frequency, undergoing a process of "approaching RF - reaching RF - moving away from RF," causing the heart rate oscillation amplitude (i.e., the HRV assessment value) to first increase and then decrease. Ultimately, the effective resonant frequency respiratory rate corresponding to a stable heart rate oscillation amplitude can be calculated. The detailed algorithm process is described below.

[0049] In this embodiment, the specific details of signal acquisition and preset are as follows: Based on user requirements and characteristics, the approximate range of this resonant frequency respiratory rate detection (RF scan) is preset (e.g., Figure 6 The preset values ​​range from 6.75 beats / min to 4.25 beats / min. After setting the total duration of the measurement, the ECG and respiration signals can be collected synchronously (the collection of respiration signals is not mandatory, but a respiratory wave is introduced to improve interactivity and training feedback).

[0050] In this embodiment, the specific content of the preprocessing is as follows: During the preprocessing stage, based on preset values, the algorithm generates a mapping table between guided respiratory rate and time. This automatically and quantitatively calculates the incremental duration of each respiratory cycle and progressively guides the user to adjust to the specified respiratory rate through animation. To reduce noise such as power line interference, electromyographic interference, and baseline drift, the acquired signals are filtered using Butterworth filters. The ECG signal filter consists of a 1Hz high-pass filter and a 30Hz low-pass filter; the respiratory signal filtering passband range is 0.03Hz to 0.7Hz. The preprocessed signals will be used for real-time interactive feedback display in the software and subsequent HRV analysis.

[0051] In this embodiment, the specific details of generating the RRI analysis sequence are as follows: The preprocessed ECG signal underwent R-peak identification, and the Pan-Tomkins algorithm was used for QRS waveform identification to obtain the time interval between adjacent R-peaks, denoted as the RRI sequence. Elements that did not conform to ECG characteristics and refractory period rules were stabilized through outlier removal and other methods, with their values ​​defined as the harmonic average of adjacent elements, thus ensuring the stability and effectiveness of the RRI sequence. A curve showing the RRI change over time was generated using the spline cubic spline interpolation method, used for both real-time display feedback and heart rate variability analysis.

[0052] In this embodiment, the specific details of the sliding analysis are as follows: Parameters for the sliding analysis window are set, including the window length (e.g., 1 minute) and the step size of the window center point movement during each slide (e.g., 15 seconds). Based on the set sliding parameters, the relationship between heart rate variability and respiratory rate is evaluated segment by segment within each window. The instantaneous heart rate within each analysis window can be obtained from the RRI time curve, and the stable difference between the heart rate peaks and troughs is used as the effective heart rate oscillation amplitude, i.e., the assessment value of heart rate variability. The calculated HRV index for each window is defined at the center of each sliding window. Finally, an HRV-respiratory rate curve is interpolated, and the window corresponding to the maximum HRV amplitude is locked. By tracing back to the respiratory frequency corresponding to that window based on the time sequence relationship, the resonant frequency respiratory rate of the user can be determined.

[0053] To more precisely describe the steps described above, including sliding analysis, respiratory synchronization quality assessment, phase compensation, comprehensive assessment value calculation, and long-term treatment management, the following detailed explanation is provided in conjunction with specific formulas: Specifically, symbol definition and sliding analysis window: Let the first Each sliding analysis window is Its central time is Window length is ,but: in, , This represents the total number of sliding analysis windows defined during a single resonant frequency detection process.

[0054] Let the progressive breathing guidance signal be The true respiratory signal is The system pre-records the "time-instantaneous guided breathing rate" mapping function for the progressive breathing guidance animation: in, Indicates time The corresponding instantaneous respiratory rate.

[0055] In the Within a sliding analysis window, define: in, The guide breathing rate corresponding to the center of this window. This refers to the actual respiratory rate extracted from real respiratory signals.

[0056] The progressive breathing guidance animation synchronously records the instantaneous guided breathing frequency at each moment during generation, forming a mapping function between time and instantaneous guided breathing frequency. Subsequently, when determining the target sliding analysis window, the system backtracks from the mapping function based on the center time of the sliding analysis window or the center time after phase compensation to obtain the corresponding breathing frequency.

[0057] Specifically, respiratory synchronization quality score: 1. Overall score: To avoid misjudgment of resonant frequency due to users not accurately following the guidance animation, it is recommended to define the breathing synchronization quality score as a weighted result of multiple sub-scores: in, For the first The respiratory synchronization quality score of a sliding analysis window; Scoring for frequency synchronization; For phase synchronization scoring; Score the cycle integrity. Amplitude stability score; For the corresponding weights, the following conditions must be met: In a preferred embodiment, it is preferable to: 2. Frequency synchronization score: The frequency synchronization score measures the degree of consistency between the actual respiratory rate and the guided respiratory rate, and can be defined as: in, This represents the maximum permissible frequency deviation. In one embodiment, Desirable times per minute.

[0058] 3. Phase synchronization score: Phase synchronization score is used to measure the overall lag or lead of the actual respiratory signal relative to the guiding respiratory signal, and can be defined as: in, For the first The time offset of the actual respiratory signal relative to the guiding respiratory signal within a sliding analysis window. This is the maximum allowed time offset threshold.

[0059] It can be obtained through cross-correlation: in, For the first A real respiratory signal fragment within a sliding window. For the first A segment of guiding breathing signals within a sliding window.

[0060] 4. Periodic integrity score: Cycle integrity score is used to determine whether a user has completed the required number of breathing cycles within the sliding window: in, For the first The number of complete respiratory cycles identified within each sliding window. The number of respiratory cycles that should occur according to the guideline respiratory rate. To prevent extremely small positive numbers with a denominator of zero.

[0061] 5. Amplitude stability score: Amplitude stability score is used to determine whether the actual breathing amplitude is stable, avoiding the influence of shallow breathing, sudden pauses, or large fluctuations on the judgment: in, For the first Amplitude sequence of real respiratory signals within a sliding window Its coefficient of variation, This is the threshold for the maximum allowable amplitude variation.

[0062] The coefficient of variation can be defined as: The breathing synchronization quality score does not simply determine whether the user has completed breathing training, but comprehensively reflects the frequency consistency, phase consistency, periodic integrity, and amplitude stability between the real breathing signal and the guided breathing signal. Through this score, the system can identify situations where the user is not following, partially following, or is following with overall delay, and accordingly retain, downgrade, or remove items from the sliding analysis window.

[0063] Specifically, phase compensation and time correction: 1. Phase offset calculation: For the A sliding analysis window is used, and the system determines the time offset based on the cross-correlation results between the actual respiratory signal and the guided respiratory signal: when When, it indicates that the actual breathing signal lags behind the guidance signal; when This indicates that the actual breathing signal is ahead of the guidance signal.

[0064] 2. Sliding window center time correction: If the user's actual breathing is lagging behind or leading the guidance signal, but still maintains a periodic synchronization relationship, then the center time of the sliding window will be corrected: in, This is the center time of the sliding analysis window after phase compensation.

[0065] 3. Corrected guided respiratory rate: Based on the center time after phase compensation The corrected respiratory rate is obtained by backtracking from the time-instantaneous guided respiratory rate mapping function: in, For the first The corresponding respiratory rate after time correction for each sliding analysis window.

[0066] 4. Phase deviation grading processing: It is recommended to handle phase deviation in three cases: in, The first time offset threshold, This is the second time offset threshold, and In one embodiment, it is preferable to: When the actual respiratory signal lags or leads the guided respiratory signal only in an overall phase, the system does not directly discard the sliding analysis window. Instead, it performs time correction on the correspondence between the sliding analysis window and the instantaneous guided respiratory frequency based on the phase offset. This avoids misinterpreting slight user delays as changes in resonant frequency due to time mismatch.

[0067] Specifically, the sliding window validity determination and weighting rules are as follows: 5. Validity weight: Based on respiratory synchrony quality score , define the first Effectiveness weights of each sliding analysis window : in, For low quality threshold, This is a high-quality threshold.

[0068] In one embodiment, the following is acceptable: when When, the sliding window is canceled; when When the sliding window participates in subsequent calculations, it is downweighted; when In this case, the sliding window participates in subsequent calculations as a high-quality window.

[0069] 6. Integration with phase processing rules: If a sliding window simultaneously satisfies both high respiratory synchronization quality and small phase deviation, it is preferentially retained; if the respiratory synchronization quality is low or the phase deviation is too large, it is discarded or downweighted. in, The phase processing weights can be defined as follows: in, In one embodiment, Desirable .

[0070] The system does not simply delete all incompletely synchronized windows; instead, it transforms the actual degree of respiratory synchronization and phase shift into effectiveness weights. For windows with slight shifts, the system retains them through time correction; for windows with moderate shifts, the system reduces their contribution to the overall evaluation value; and for windows with severe asynchrony, the system removes them.

[0071] Specifically, heart rate oscillation amplitude and HRV stability indicators: 1. Heart rate oscillation amplitude: Let the first The instantaneous heart rate sequence obtained from the heart rate interval curve within each sliding window is as follows: To reduce the impact of outliers, the amplitude of heart rate oscillations can be represented by percentile differences: in, and They represent the first The 95th and 5th percentiles of the instantaneous heart rate sequence within a sliding window.

[0072] In another embodiment, the amplitude of heart rate oscillations can also be expressed as the difference between the median of the heart rate peaks and the median of the heart rate troughs: 2. RMSSD Metrics: Let the first The heart rate interval sequence within each sliding window is as follows: The root mean square of the difference between adjacent heartbeats Defined as: 3. Proportion of abnormal heartbeat intervals: To further evaluate the quality of heartbeat-related physiological signals, the proportion of abnormal heartbeat intervals within the sliding window can be calculated: in, This represents the number of abnormal heartbeat intervals within the sliding window. This represents the total number of intracardiac intervals within the sliding window.

[0073] The corresponding signal quality score can be defined as: Heart rate oscillation amplitude is used to characterize the intensity of respiratory-induced heart rate fluctuations, RMSSD is used to characterize the stability of intracardiac interval changes within the sliding analysis window, and the proportion of abnormal intracardiac intervals is used to characterize the reliability of intracardiac-related physiological signals within that window. By considering these indicators simultaneously, the impact of abnormal pulsations, missed detections, and motion artifacts on the determination of resonant frequencies can be reduced.

[0074] Specifically, the formula for normalized and comprehensive evaluation values ​​is as follows: 1. Indicator normalization: Because the dimensions of heart rate oscillation amplitude, respiratory synchrony quality score, and RMSSD are different, it is recommended to normalize them before calculating the comprehensive assessment value.

[0075] Normalize the amplitude of heart rate oscillations: Normalize RMSSD: because For scores already in the range of 0 to 1, let: 2. Overall evaluation value: The comprehensive evaluation value can be defined as: in, For the first The overall evaluation value of each sliding analysis window To incorporate the window effectiveness weights after considering respiratory synchronization quality and phase deviation, Let the weighting coefficients satisfy: In a preferred embodiment, it is preferable to: 3. Target sliding analysis window determination: The target sliding analysis window is the sliding analysis window with the largest overall evaluation value: in, Number the target sliding analysis window.

[0076] This invention uses a comprehensive evaluation value instead of a single peak-to-trough difference in heart rate as the basis for determining the resonant frequency. The comprehensive evaluation value simultaneously considers the intensity of respiratory-induced heart rate oscillations, the quality of actual respiratory follow-up, and the stability of changes in heart rate intervals, and reduces the impact of low-quality windows on the results through window validity weighting.

[0077] Specifically, the formula for single-resonance frequency output is: In determining the target sliding analysis window Then, based on the center time after phase compensation Tracing back to the corresponding instantaneous guided breathing rate: in, The user's resonant frequency respiratory rate is the output of this test.

[0078] If the target sliding analysis window does not perform phase compensation, then we can set: thereby: To improve the stability of the results, a local weighted average can also be taken near the target sliding analysis window: in, The target sliding analysis window and its adjacent sliding analysis windows form a local neighborhood.

[0079] After determining the target sliding analysis window, the system does not directly use the uncorrected animation time for frequency backtracking. Instead, it corrects the center time of the sliding analysis window based on the phase offset between the real respiratory signal and the guided respiratory signal, and determines the corresponding instantaneous respiratory rate based on the corrected time, thereby improving the accuracy of the resonant frequency respiratory rate determination.

[0080] Specifically, long-term treatment management: 1. Initial candidate interval: Upon first use, the system determines the initial candidate interval based on a preset range, historical detection results, or stepwise coarse detection results: For example: If the user already has a historical resonant frequency, it can be set as follows: in, This is the initial expanded width.

[0081] 2. Multiple tests during the training period Let a training cycle be For example, a week or a month. Executed within this period. Secondary sliding resonant frequency detection. The candidate resonant frequencies are output from the second detection: It also outputs the corresponding detection quality parameters, including the average respiratory synchronization quality score. Peak resolution Heartbeat-related physiological signal quality score .

[0082] The average respiratory synchronization quality score can be defined as: in, For the first The number of sliding analysis windows included in each detection.

[0083] 3. Peak resolution: Peak sharpness is used to determine whether the overall evaluation curve of this test has a clear peak. Let the first peak be... The maximum comprehensive evaluation value in this test was The second largest comprehensive evaluation value is ,but: Alternatively, the difference between the target window and adjacent windows can be used as the definition: in, For the first The comprehensive evaluation value of the target sliding analysis window in the second detection.

[0084] 4. Determination of the validity of a single test: No. Whether a test is valid can be determined based on the following conditions: in, , , These are the minimum thresholds for average respiratory synchronicity quality, peak clarity, and heart rate-related physiological signal quality, respectively.

[0085] 5. Inspection quality weight: For valid detection results, the system calculates the quality weight based on the detection quality parameters: in: In a preferred embodiment, it is preferable to: 6. Stage-specific resonant frequencies: The phase resonant frequencies of the current training cycle can be obtained through weighted aggregation: in, This is the phased resonant frequency output for the current training cycle.

[0086] In another embodiment, the stage resonant frequency can also be determined using a weighted median to reduce the impact of extreme detection results on the periodic output value.

[0087] 7. Candidate intervals will be dynamically updated in the next period: After the current training period ends, the candidate interval for the next period can be dynamically updated based on the phased resonant frequency: in, To expand the detection range for the next cycle.

[0088] If the current cycle detection quality is high, a narrower detection range will be used: If the current cycle's detection quality is low, a wider detection range will be used: in: If the number of valid detections in the current cycle is insufficient: Then the initial candidate interval will be restored in the next cycle: The long-term treatment management process is not simply about recording multiple resonant frequency respiratory rates. Instead, it involves filtering and quality-weighted aggregation of test results based on the respiratory synchronization quality, peak clarity of the comprehensive assessment value, and the quality of cardiac-related physiological signals for each test, to obtain the staged resonant frequency at the training cycle scale. The system further dynamically adjusts the candidate detection range for the next training cycle based on the staged resonant frequency and detection quality, ensuring accuracy while reducing unnecessary wide-range scans in subsequent tests.

[0089] In this embodiment, in step S3, the real breathing signal is acquired through a breathing belt, a nasal / oral airflow sensor, chest and abdominal impedance, or a breathing waveform extracted from an electrocardiogram signal; in step S5, before defining the sliding analysis window, the method further includes: calculating the breathing synchronization quality score corresponding to each sliding analysis window based on the frequency deviation, phase deviation, period integrity, and / or amplitude stability between the real breathing signal and the progressive breathing guidance animation; when the breathing synchronization quality score is lower than a preset threshold, the sliding analysis window is downweighted or removed.

[0090] In this embodiment, a phase compensation step is also included: the correspondence between the sliding analysis window and the instantaneous respiratory rate is corrected in time based on the phase offset between the real respiratory signal and the progressive respiratory guidance animation; when the phase deviation is less than a first threshold, time correction is performed; when the phase deviation is greater than a second threshold, the corresponding sliding window is removed or its weight is reduced.

[0091] In this embodiment, the comprehensive evaluation value is obtained by weighting the heart rate oscillation amplitude, respiratory synchronization quality score, and heart rate variability stability index; wherein, the heart rate variability stability index is calculated by the root mean square standard deviation of the intra-beat interval sequence within the sliding analysis window.

[0092] In this embodiment, the weights of heart rate oscillation amplitude, respiratory synchrony quality score, and heart rate variability stability index are 0.6, 0.3, and 0.1, respectively.

[0093] In this embodiment, after determining the resonant frequency respiratory rate in step S6, a long-term treatment management step, namely step S7, is included. Specifically, this involves: determining candidate intervals based on the user's initial test, historical test results, or a step-by-step coarse test; within a training cycle (e.g., one week or one month), performing multiple sliding fine tests within the candidate intervals (i.e., executing steps S1 to S6 multiple times) to obtain multiple valid test results; eliminating invalid results with respiratory synchronization quality below a threshold; and weighted aggregation of the valid results to output a weekly or monthly staged resonant frequency for updating subsequent respiratory training parameters. This step enables long-term treatment management of the user's resonant frequency, adapting to changes in different dates, physical conditions, and training stages.

[0094] like Figure 9 As shown, in some embodiments, the long-term treatment management steps are further refined into the following long-term RF determination and periodic fine-tuning process. When a user uses the system for the first time or when historical RF data is unreliable, the system first performs a wide range of resonance frequency detection to obtain the initial resonance frequency, which is then used as a reference benchmark for subsequent periodic fine-tuning. After entering long-term training, the system uses a preset time length as the training cycle. At the beginning of each training cycle, the system establishes the candidate detection interval for the current cycle based on the stage RF of the previous cycle. When the detection quality of the previous cycle is high and the results are stable, the current cycle uses a narrower candidate interval for fine detection. When the effective detection of the previous cycle is insufficient, the results fluctuate greatly, or the user interrupts training and restarts, the current cycle expands the candidate interval or restores the initial candidate interval.

[0095] Within each training cycle, the system performs sliding radiofrequency (RF) detection multiple times within the candidate interval. After each detection, the system determines the validity of the detection result based on the quality of respiratory tracking, the quality of cardiac-related physiological signals, and the peak clarity of the comprehensive evaluation value. Detection results with ineffective respiratory tracking, insufficient cardiac signal quality, or unclear peaks are marked as invalid and excluded from the determination of the current cycle's phased RF.

[0096] Once a sufficient number of valid detection results are obtained within the current training cycle, the system weights and determines multiple candidate RFs based on the quality of each detection, resulting in the current cycle's interim RF. If the number of valid detection results is insufficient or the dispersion of multiple candidate RFs is too high, the system does not perform a significant update, but instead uses the RF from the previous cycle or makes conservative fine-tuning. The system compares the current cycle's interim RF with the previous cycle's RF; when the difference is within an acceptable range, it directly updates the training frequency for the next cycle; when the difference exceeds the preset maximum adjustment range, it performs limited fine-tuning according to the maximum adjustment range. After the current cycle ends, the system dynamically determines the detection range for the next cycle based on the number of valid detections, detection quality, and RF stability in this cycle, thus forming a long-term RF management process of "initial benchmark establishment—cycle review—quality screening—interim determination—limited fine-tuning—next cycle range update".

[0097] The following explains the specific content of the respiratory synchronization quality assessment and comprehensive assessment values: To further improve the accuracy of resonant frequency detection, this embodiment synchronously acquires the user's actual breathing signal in step S3. The actual breathing signal can be obtained through a breathing belt, a nasal / mouth airflow sensor, chest and abdominal impedance, or a breathing waveform extracted from an electrocardiogram signal.

[0098] This invention compares real breathing signals with progressive breathing guidance animations to identify the inspiratory start point, expiratory start point, respiratory cycle length, and respiratory amplitude of real breathing. It calculates frequency deviation, phase deviation, cycle integrity, and amplitude stability, and obtains a respiratory synchronization quality score after normalization and weighting. When the score is below a preset threshold, the corresponding sliding analysis window is downweighted or removed. Unlike existing technologies that assume the user fully follows the guidance, this invention actively detects the user's actual following status through the respiratory synchronization quality score, avoiding the use of window data from when the user is not following or only partially following for resonant frequency determination, thereby improving detection accuracy.

[0099] Based on this, this embodiment uses a comprehensive evaluation value instead of a single heart rate peak-to-trough difference. The comprehensive evaluation value is obtained by weighting the heart rate oscillation amplitude (the difference between the peak and trough of the heart rate wave within the sliding window), the respiratory synchronization quality score, and the heart rate variability stability index (the root mean square standard deviation of the heartbeat interval sequence). In a preferred embodiment, the weights of the three are 0.6, 0.3, and 0.1, respectively. This comprehensive evaluation value can reduce the impact of motion artifacts, abnormal pulsations, or respiratory asynchrony on the detection results.

[0100] In addition, when the user's actual breathing lags behind the guidance signal but the cycle length is basically the same, the system performs time correction based on the phase offset: if the phase deviation is less than 0.5 seconds, the correspondence between the sliding analysis window and the breathing frequency is corrected; if it is greater than 1.5 seconds, the window is removed; windows in between are retained but have their weight reduced.

[0101] To adapt to the characteristic of the resonant frequency changing with the training phase, this embodiment also provides a long-term dynamic update mechanism: the initial detection can use a wide-range scan or a step-by-step coarse detection to obtain the initial candidate interval; in subsequent training cycles (such as one week or one month), a sliding fine detection is performed near the candidate interval, and multiple effective detection results are weighted and aggregated according to quality scores to output the stage resonant frequency for the next stage of training parameter settings.

[0102] In some embodiments, to avoid misjudgment of resonant frequency due to discrepancies between the user's actual breathing and the progressive breathing guidance animation, this invention introduces a breathing synchronization quality score during the sliding analysis process. Specifically, while the user follows the progressive breathing guidance animation for breathing training, the system synchronously acquires real breathing signals and compares them with the guided breathing signals. The system calculates the breathing synchronization quality score corresponding to each sliding analysis window based on the frequency deviation between the real and guided breathing frequencies, the phase deviation between the real and guided breathing signals, the number of complete breathing cycles within the sliding window, and the stability of the real breathing amplitude.

[0103] In some embodiments, the system further performs time correction based on the phase offset between the actual respiratory signal and the guided respiratory signal. If the actual respiratory signal only lags or leads the guided respiratory signal overall, but its periodic structure is still basically consistent with the guided signal, the system corrects the center time of the sliding analysis window and backtracks the corresponding respiratory frequency from the time-instantaneous guided respiratory frequency mapping function based on the corrected time. If the phase offset exceeds a preset threshold, the system reduces the weight of the corresponding sliding window or removes it.

[0104] In some embodiments, the system uses a comprehensive evaluation value instead of a single heart rate peak-to-trough difference as the basis for determining the resonant frequency. The comprehensive evaluation value is obtained by weighting the normalized heart rate oscillation amplitude, respiratory synchronization quality score, and heart rate variability stability index, and further multiplying it by a sliding window effectiveness weight. In this way, the system can simultaneously consider the intensity of respiratory-induced heart rate oscillations, the user's actual respiratory following quality, and the stability of heartbeat-related physiological signals, thereby reducing the impact of abnormal pulsations, motion artifacts, and respiratory asynchrony on the determination of the resonant frequency.

[0105] In some embodiments, the long-term treatment management steps are used to dynamically update the user's staged resonant frequency on a weekly or monthly scale. Specifically, the system performs sliding resonant frequency detection multiple times within a training cycle to obtain multiple candidate resonant frequency respiratory rates, and determines the validity of the detection results based on the average respiratory synchronization quality score, peak clarity of the comprehensive assessment value, and quality score of heartbeat-related physiological signals corresponding to each detection. The system discards invalid detection results and weights and aggregates the valid detection results according to the detection quality weight to obtain the staged resonant frequency of the current training cycle. After the current training cycle ends, the system dynamically adjusts the candidate detection range for the next training cycle based on the staged resonant frequency and detection quality. When the detection quality is high, the next cycle uses a narrower range centered on the staged resonant frequency for fine detection; when the detection quality is low or the number of valid detections is insufficient, the detection range is expanded or the initial detection range is restored.

[0106] The present invention relates to a personalized resonant frequency breathing training intelligent guidance system for optimizing resonant frequency breathing training methods. Trainees follow the animated breathing, and the software dynamically adjusts the breathing frequency sequentially. It has the advantages of real-time feedback of physiological indicators, no need for professional guidance, and intelligent personalized matching, which increases the fun and learning of the breathing training process and has a certain positive effect on the promotion of resonant frequency breathing training.

[0107] The sliding analysis window method used in the detection of resonant frequency respiratory rate in the embodiment of the present invention for optimizing the resonant frequency breathing training method eliminates the need for multiple mechanical repetitions of the traditional discrete ladder method, thereby improving the efficiency and resolution accuracy of RF detection. It greatly optimizes the detection process of resonant frequency respiratory rate from a completely new perspective and improves the way resonant frequency breathing training strategies are formulated.

[0108] Experimental example: Ten subjects were selected to determine the resonant frequency of breathing using both the traditional step method and the sliding method of this invention, and the magnitudes of the resonant breathing frequencies determined by the two methods were compared.

[0109] The specific procedure for the stepwise breathing method is as follows: Subjects determine their breathing pattern based on existing resonant frequencies and follow the breathing instructions sequentially: 8 BPM, 7.5 BPM, 7 BPM, 6.5 BPM, 6 BPM, 5.5 BPM, 5 BPM, and 4.5 BPM. Physiological signals are recorded during the process, and relevant indicators are calculated.

[0110] The specific procedure for the sliding method is as follows: In the sliding method experiment, the subject first undergoes two minutes of deep breathing to adapt and adjust. Then, based on the subject's characteristics, the detection range of the resonant frequency respiratory rate (the upper and lower limits of the possible values ​​of the resonant respiratory frequency) and the total sliding detection time are set. During the detection process, the respiratory frequency guided by the animation will gradually decrease from the maximum respiratory rate to the minimum respiratory rate. The subject only needs to follow the animation guidance to complete the test in one go, without the need for multiple interruptions in groups. After the measurement, the subject's resonant frequency will be automatically identified based on the algorithm of the resonant frequency respiratory rate.

[0111] like Figure 8 The results show a comparative analysis of the resonant frequencies calculated using the step method and the sliding method. From an average perspective, the RF results measured by the step method are on average 0.08 bpm higher than those measured by the sliding method, which is closely related to the different resolutions of the two methods. Of the 10 pairs of RFs measured using both the step method and the sliding method, 8 pairs (80%) of the sliding method results fall within the 0.5 bpm resolution range of the step method results. In the other two pairs of RFs with an absolute error exceeding 0.5 bpm, one difference was 0.614 (subject 3), and the other was 0.629 (subject 9).

[0112] The breathing training method for optimizing resonant frequency according to embodiments of the present invention has the following advantages: 1. The method for optimizing resonant frequency breathing training according to the embodiments of the present invention can be easily used by ordinary users. The built-in algorithm can customize a personalized breathing feedback training program for users, including resonant frequency breathing rate and inspiratory-to-expiratory ratio. The human-computer interaction guided by animation enables users who have not received professional medical training to complete breathing feedback training independently.

[0113] 2. Compared with traditional resonant frequency respiratory rate detection methods, the sliding method used in this invention eliminates the need for mechanically time-consuming repetitive testing at each step respiratory rate (e.g., 6.5, 6.0, 5.5, 5.0, 4.5 breaths / minute). The new method of detecting resonant frequency by successively decreasing the respiratory rate and sliding analysis provides high accuracy and greatly reduces the required time, simplifying the resonant respiratory rate measurement process and improving the experience and efficiency of resonant frequency breathing training.

[0114] 4. Compared with other breathing training methods, the breathing training method for optimizing resonant frequency in this embodiment of the invention incorporates the inspiratory-to-expiratory ratio into the variable parameters of the breathing feedback training program. Through the algorithm, a suitable inspiratory-to-expiratory ratio is matched for different individuals, and a personalized breathing training plan is formulated to achieve the best effect of breathing training.

[0115] Example 2 like Figure 10 As shown, this embodiment of the invention also provides a breathing training system for optimizing resonant frequency, comprising: The parameter preset module is used to obtain the user's resonant frequency respiratory rate detection range and total sliding detection time as preset parameters; An animation generation module is used to generate a progressive breathing guidance animation based on the preset parameters, in which the breathing rate decreases from the maximum value to the minimum value within the detection range in one breath, and to record the instantaneous breathing rate corresponding to each breathing cycle. The physiological acquisition module is used to synchronously acquire the user's physiological signals to obtain raw physiological signals and acquire the user's real breathing signals during the breathing training process while the user follows the progressive breathing guidance animation. The signal processing module is used to preprocess the raw physiological signal and identify heartbeat feature points, extract the heartbeat interval sequence, and generate a curve of the heartbeat interval changing over time based on the heartbeat interval sequence to obtain the heartbeat interval time curve. The sliding analysis module is used to sequentially delineate multiple sliding analysis windows on the heartbeat interval time curve according to preset sliding analysis window parameters. For each sliding analysis window, a respiratory synchronization quality score is calculated based on at least one of the following: frequency deviation, phase deviation, period integrity, and amplitude stability between the actual respiratory signal and the progressive respiratory guidance animation. If the respiratory synchronization quality score is lower than a preset threshold, the sliding analysis window is downweighted or removed. The heart rate oscillation amplitude within the remaining sliding analysis windows is combined with the respiratory synchronization quality score to calculate the comprehensive evaluation value of each sliding analysis window, resulting in a comprehensive evaluation value sequence. The resonance frequency determination module is used to determine the target sliding analysis window corresponding to the maximum value in the comprehensive evaluation value sequence, and to backtrack to obtain the breathing frequency corresponding to the target sliding analysis window based on the progressive breathing guidance animation, as the user's resonance frequency breathing rate; The long-term treatment management module is used to call the parameter preset module, animation generation module, physiological acquisition module, signal processing module, sliding analysis module, and resonance frequency determination module multiple times within a training cycle to obtain multiple effective resonance frequency respiratory rates. Invalid results with respiratory synchronization quality scores below the threshold are eliminated, and the effective resonance frequency respiratory rates are weighted and aggregated to output the stage resonance frequencies on a weekly or monthly scale for updating subsequent respiratory training parameters.

[0116] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for optimizing resonant frequency breathing training, characterized in that, Includes the following steps: Step S1: Obtain the user's resonant frequency respiratory rate detection range and total sliding detection time as preset parameters; Step S2: Based on the preset parameters, generate a progressive breathing guidance animation in which the respiratory rate decreases from the maximum value to the minimum value within the detection range in one breath, and record the instantaneous respiratory rate corresponding to each breathing cycle; Step S3: During the user's breathing training by following the progressive breathing guidance animation, the user's physiological signals are collected synchronously to obtain the original physiological signals, and the user's real breathing signals are also collected. Step S4: Preprocess the raw physiological signal and identify heartbeat feature points, extract the heartbeat interval sequence, and generate a curve of heartbeat interval changing with time based on the heartbeat interval sequence to obtain the heartbeat interval time curve; Step S5: According to the preset sliding analysis window parameters, multiple sliding analysis windows are sequentially defined on the heartbeat interval time curve; for each sliding analysis window, a respiratory synchronization quality score is calculated based on at least one of the frequency deviation, phase deviation, period integrity, and amplitude stability between the real respiratory signal and the progressive respiratory guidance animation; if the respiratory synchronization quality score is lower than a preset threshold, the sliding analysis window is downweighted or removed; the heart rate oscillation amplitude in the remaining sliding analysis windows is combined with the respiratory synchronization quality score to calculate the comprehensive evaluation value of each sliding analysis window, thus obtaining a comprehensive evaluation value sequence; Step S6: Determine the target sliding analysis window corresponding to the maximum value in the comprehensive evaluation value sequence, and backtrack to obtain the breathing frequency corresponding to the target sliding analysis window based on the progressive breathing guidance animation, as the user's resonant frequency breathing rate; Step S7: Execute steps S1 to S6 multiple times within a training cycle to obtain multiple effective resonant frequency respiratory rates. Eliminate invalid results with respiratory synchronization quality scores below the threshold, and perform weighted aggregation on the effective resonant frequency respiratory rates to output the stage resonant frequencies on a weekly or monthly scale for updating subsequent respiratory training parameters.

2. The method for optimizing resonant frequency breathing training according to claim 1, characterized in that, In step S4, the raw physiological signal is preprocessed, specifically including: filtering the raw physiological signal using a Butterworth bandpass filter, wherein the Butterworth bandpass filter consists of a 1Hz high-pass filter and a 30Hz low-pass filter.

3. The method for optimizing resonant frequency breathing training according to claim 1, characterized in that, In step S4, the inter-heartbeat sequence is extracted, specifically including: A heartbeat feature point detection algorithm is used to identify heartbeat feature points in preprocessed physiological signals; Calculate the time interval between adjacent heartbeat feature points to obtain the initial heartbeat interval sequence; The initial heartbeat interval sequence is subjected to outlier removal and stabilization processing. Outliers that do not conform to physiological characteristics are replaced with the harmonic mean of their adjacent elements to obtain the heartbeat interval sequence.

4. The method for optimizing resonant frequency breathing training according to claim 1, characterized in that, In step S4, generating the heartbeat interval time curve specifically includes: interpolating the heartbeat interval sequence into a continuous time curve using the spline cubic spline interpolation method to obtain the heartbeat interval time curve.

5. The method for optimizing resonant frequency breathing training according to claim 1, characterized in that, In step S5, the window length of the sliding analysis window is 1 minute, and the step size for each slide is 15 seconds. The step size corresponds to the movement distance of the center point of the sliding analysis window.

6. The method for optimizing resonant frequency breathing training according to claim 1, characterized in that, In step S5, the comprehensive evaluation value within each sliding analysis window is calculated, specifically including: extracting the heart rate peak and heart rate trough within the current sliding analysis window based on the heart rate interval curve, calculating the stable difference between the heart rate peak and the heart rate trough, and using the stable difference as a component of the comprehensive evaluation value within the current sliding analysis window.

7. The method for optimizing resonant frequency breathing training according to claim 1, characterized in that, In step S6, after determining the resonant frequency respiratory rate, the step of matching the inspiratory-to-expiratory ratio is also included: The user's breathing rate is fixed at the resonant frequency breathing rate, and multiple preset inspiratory-to-expiratory ratio modes are switched sequentially, generating corresponding inspiratory-to-expiratory ratio guidance animations. While the user follows the breathing training animation for each inspiratory-to-expiratory ratio, the user's physiological signals are collected synchronously, and the heart rate variability index is calculated based on the collected physiological signals for each inspiratory-to-expiratory ratio mode. The maximum value among the various heart rate variability indicators is selected, and the inspiratory-to-expiratory ratio corresponding to the maximum value is determined as the optimal inspiratory-to-expiratory ratio for the user.

8. The method for optimizing resonant frequency breathing training according to claim 1, characterized in that, Step S3 also includes synchronously collecting the user's breathing signals; the breathing signals are used for real-time interactive feedback during breathing training to help the user synchronize with the progressive breathing guidance animation.

9. The method for optimizing resonant frequency breathing training according to claim 8, characterized in that, In step S3, the user's respiratory signal acquired synchronously is preprocessed, specifically including filtering the respiratory signal using a Butterworth filter with a passband range of 0.03Hz to 0.7Hz.

10. A breathing training system for optimizing resonant frequency, characterized in that, include: The parameter preset module is used to obtain the user's resonant frequency respiratory rate detection range and total sliding detection time as preset parameters; An animation generation module is used to generate a progressive breathing guidance animation based on the preset parameters, in which the breathing rate decreases from the maximum value to the minimum value within the detection range in one breath, and to record the instantaneous breathing rate corresponding to each breathing cycle. The physiological acquisition module is used to synchronously acquire the user's physiological signals to obtain raw physiological signals and acquire the user's real breathing signals during the breathing training process while the user follows the progressive breathing guidance animation. The signal processing module is used to preprocess the raw physiological signal and identify heartbeat feature points, extract the heartbeat interval sequence, and generate a curve of the heartbeat interval changing over time based on the heartbeat interval sequence to obtain the heartbeat interval time curve. The sliding analysis module is used to sequentially delineate multiple sliding analysis windows on the heartbeat interval time curve according to preset sliding analysis window parameters. For each sliding analysis window, a respiratory synchronization quality score is calculated based on at least one of the following: frequency deviation, phase deviation, period integrity, and amplitude stability between the actual respiratory signal and the progressive respiratory guidance animation. If the respiratory synchronization quality score is lower than a preset threshold, the sliding analysis window is downweighted or removed. The heart rate oscillation amplitude within the remaining sliding analysis windows is combined with the respiratory synchronization quality score to calculate the comprehensive evaluation value of each sliding analysis window, resulting in a comprehensive evaluation value sequence. The resonance frequency determination module is used to determine the target sliding analysis window corresponding to the maximum value in the comprehensive evaluation value sequence, and to backtrack to obtain the breathing frequency corresponding to the target sliding analysis window based on the progressive breathing guidance animation, as the user's resonance frequency breathing rate; The long-term treatment management module is used to call the parameter preset module, animation generation module, physiological acquisition module, signal processing module, sliding analysis module, and resonance frequency determination module multiple times within a training cycle to obtain multiple effective resonance frequency respiratory rates. Invalid results with respiratory synchronization quality scores below the threshold are eliminated, and the effective resonance frequency respiratory rates are weighted and aggregated to output the stage resonance frequencies on a weekly or monthly scale for updating subsequent respiratory training parameters.

Images