A real-time guidance method for pelvic floor muscle exercise based on multi-parameter sensing of an abdominal belt

Abstract

This invention discloses a real-time guidance method for pelvic floor muscle exercises based on multi-parameter sensing using an abdominal belt, belonging to the field of physiological monitoring technology. The method includes: constructing a set of training rhythm windows based on respiratory phase sequences and training quality sequences; calculating the gating confidence of each training rhythm window in the set; generating a real-time guidance sequence; performing anomaly detection processing on the real-time guidance sequence to generate an anomaly prompt sequence; and simultaneously extracting training evaluation indicators for each training rhythm window to generate a training score record. This invention, by generating abdominal pressure change sequences and real-time guidance sequences, achieves collaborative analysis of the relationship between respiratory rhythm and abdominal pressure changes based on multi-source signals such as abdominal wall displacement and abdominal belt tension collected by an intelligent abdominal belt sensor, thereby improving the real-time guidance capability and training effect evaluation capability of the pelvic floor muscle training process.

Description

Technical Field

[0001] This invention relates to the field of physiological monitoring technology, and in particular to a method for real-time guidance of pelvic floor muscle movement based on multi-parameter sensing of an abdominal binder. Background Technology

[0002] With the development of rehabilitation medicine and smart wearable technology, pelvic floor muscle training is gradually evolving from traditional clinical guidance to digital and intelligent assisted training. Pelvic floor muscle training is commonly used in postpartum rehabilitation, urinary incontinence intervention, and pelvic floor dysfunction recovery. Its effectiveness largely depends on the trainee's correct control over respiratory rhythm, abdominal pressure changes, and abdominal muscle coordination. Current technologies commonly employ face-to-face movement guidance from rehabilitation therapists or monitoring pelvic floor muscle contraction signals with simple biofeedback devices and providing visual or auditory cues. In recent years, with the development of wearable sensing technology, some studies have begun to utilize smart sensors such as abdominal belts, pressure sensors, or inertial sensing components to monitor abdominal movements and provide training prompts to users via mobile terminals.

[0003] However, existing technologies still have certain shortcomings. For example, they often lack precise identification of respiratory phase structure during training guidance and lack a dynamic mechanism for constructing training windows. This results in training prompts typically being triggered at fixed rhythms or time intervals, making it difficult to adapt to changes in the respiratory rhythms of different users. Furthermore, in terms of training process monitoring, existing technologies often cannot distinguish abnormal movement states in detail. For example, abnormal increases in abdominal pressure, breath-holding states, or abdominal compensatory movements are difficult to accurately identify, leading to a lack of targeted training guidance and difficulty in quantifying the training process. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a real-time guidance method for pelvic floor muscle exercise based on abdominal belt multi-parameter sensing to solve the problem of difficulty in constructing training guidance timing and accurately identifying and quantifying abnormal training states.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] This invention provides a real-time guidance method for pelvic floor muscle exercise based on multi-parameter sensing of an abdominal belt. The method includes: collecting multi-parameter observation data of the abdominal belt during the user's wearing of the belt and performing unified temporal organization and observation consistency management to generate an abdominal belt observation sequence; identifying the respiratory cycle structure of the abdominal belt observation sequence and statistically analyzing respiratory cycle parameters to generate a respiratory phase sequence; using the respiratory phase sequence as a constraint, performing abdominal pressure response reconstruction processing on the abdominal belt observation sequence to generate an abdominal pressure change sequence; extracting abdominal pressure features from the abdominal pressure change sequence to form an abdominal pressure feature sequence; evaluating the rhythmic stability and abdominal wall stability of the abdominal belt observation sequence and the abdominal pressure feature sequence respectively based on the respiratory phase sequence to generate a training quality sequence; constructing a training rhythm window set based on the respiratory phase sequence and the training quality sequence; calculating the gating confidence of each training rhythm window in the training rhythm window set to generate a real-time guidance sequence; performing anomaly detection processing on the real-time guidance sequence to generate an anomaly prompt sequence; and simultaneously extracting training evaluation indicators for each training rhythm window to generate a training score record.

[0008] As a preferred embodiment of the real-time guidance method for pelvic floor muscle movement based on abdominal belt multi-parameter sensing described in this invention, the specific steps for generating the abdominal belt observation sequence are as follows:

[0009] Read abdominal wall displacement data, abdominal belt tension data and motion disturbance data from the multi-parameter observation data of the abdominal belt, and sort them according to the corresponding timestamps to generate an initial time series data sequence;

[0010] Bandpass filtering and abnormal mutation removal are performed on the initial time series data sequence to generate the abdominal zone observation sequence.

[0011] As a preferred embodiment of the real-time guidance method for pelvic floor muscle movement based on abdominal belt multi-parameter sensing described in this invention, the specific steps for generating the respiratory phase sequence are as follows:

[0012] Identify the positions of respiratory peaks and troughs in abdominal wall displacement data in abdominal band observation sequences, and generate respiratory peak-trough sequences;

[0013] The respiratory cycle boundary is obtained by the time interval between adjacent respiratory peaks or troughs in the respiratory peak-trough sequence. The respiratory cycle parameter sequence is generated by calculating the cycle length, inspiratory duration and expiratory duration of each respiratory cycle.

[0014] Based on the respiratory cycle parameter sequence, each time position in the abdominal band observation sequence is marked as the inspiratory phase, expiratory phase, and end-expiratory phase, generating a respiratory phase sequence.

[0015] As a preferred embodiment of the real-time guidance method for pelvic floor muscle movement based on multi-parameter sensing of an abdominal belt described in this invention, the specific steps for generating the abdominal pressure change sequence are as follows:

[0016] Based on the respiratory phase sequence, the time segments corresponding to the inspiratory phase, expiratory phase and end-expiratory phase in the abdominal belt observation sequence are identified, and the corresponding abdominal wall displacement data and abdominal belt tension data are extracted to generate an abdominal wall motion response sequence.

[0017] The rate of change of abdominal wall motion and the rate of change of abdominal belt tension in the abdominal wall motion response sequence are calculated, and continuity constraints and abnormal mutation correction are performed to generate an abdominal pressure change sequence.

[0018] As a preferred embodiment of the real-time guidance method for pelvic floor muscle movement based on abdominal belt multi-parameter sensing described in this invention, the specific steps for forming the abdominal pressure feature sequence are as follows:

[0019] The magnitude of abdominal pressure change between adjacent time positions in the abdominal pressure change sequence is calculated based on the time index, and the rate of change during the abdominal pressure increase and decrease phases is calculated to generate an abdominal pressure change rate sequence.

[0020] The maximum, minimum, and difference range of abdominal pressure change sequence within each respiratory cycle are statistically analyzed, and feature combination processing is performed with the abdominal pressure change rate sequence to generate an abdominal pressure feature sequence.

[0021] As a preferred embodiment of the real-time guidance method for pelvic floor muscle exercise based on abdominal belt multi-parameter sensing described in this invention, the specific steps for generating the training quality sequence are as follows:

[0022] Based on the respiratory phase sequence, the time segment corresponding to each respiratory cycle in the abdominal band observation sequence is identified, and the abdominal band observation sequence and abdominal pressure characteristic sequence are segmented into stages to generate a respiratory phase data set.

[0023] Based on the respiratory cycle parameter sequence, calculate the cycle deviation and phase duration deviation between adjacent respiratory cycles to generate a rhythm stability parameter sequence.

[0024] By reading abdominal wall displacement data from the respiratory phase dataset, the changes in abdominal wall displacement amplitude, displacement rate, and displacement fluctuation degree in each respiratory phase are calculated, and then combined with the rhythm stability parameter sequence according to the time index to generate a training quality sequence.

[0025] As a preferred embodiment of the real-time guidance method for pelvic floor muscle exercise based on abdominal belt multi-parameter sensing described in this invention, the specific steps for constructing the training rhythm window set are as follows:

[0026] Identify the time segment corresponding to the end of the exhalation phase in the respiratory phase sequence, determine the training effectiveness based on the training quality sequence, and generate a candidate training window sequence;

[0027] Based on the phase transition position of the breathing phase sequence and the stable interval of the training quality sequence in the candidate training window sequence, window expansion and time boundary adjustment are performed on each time segment to generate a training rhythm window set.

[0028] As a preferred embodiment of the real-time guidance method for pelvic floor muscle movement based on multi-parameter sensing of an abdominal belt according to the present invention, the specific steps for generating the real-time guidance sequence are as follows:

[0029] Calculate the signal consistency score between the abdominal band observation sequence and the abdominal pressure feature sequence in the training rhythm window set, and perform interference suppression correction processing in combination with motion interference data to generate a gated confidence sequence;

[0030] Based on the gating confidence sequence, gating output judgment is performed on the training rhythm window set to form a real-time guidance sequence.

[0031] As a preferred embodiment of the real-time guidance method for pelvic floor muscle movement based on abdominal belt multi-parameter sensing described in this invention, the specific steps for generating the abnormality prompt sequence are as follows:

[0032] Based on the real-time guidance sequence, the changing trend of the abdominal pressure characteristic sequence and the phase matching relationship of the respiratory phase sequence are analyzed, and abnormal rise in abdominal pressure, respiratory phase mismatch and respiratory interruption are identified, generating abnormal abdominal pressure marker sequence and abnormal respiratory marker sequence.

[0033] The changes in abdominal wall displacement amplitude, the degree of abdominal belt tension fluctuation, and the intensity of body motion corresponding to motion interference data in the abdominal belt observation sequence were analyzed. Abnormal increase in abdominal wall motion, enhanced abdominal compensation, and postural interference were identified, and an abnormal abdominal wall marker sequence was generated.

[0034] Based on the abnormal abdominal pressure marker sequence, abnormal breathing marker sequence, and abnormal abdominal wall marker sequence, the abnormality type of each training rhythm window is determined, and an abnormality prompt sequence is output.

[0035] As a preferred embodiment of the real-time guidance method for pelvic floor muscle exercise based on multi-parameter sensing of an abdominal belt described in this invention, the specific steps for generating training score records are as follows:

[0036] The contraction allowable window hit rate, hold allowable window hit rate, and relaxation allowable window hit rate corresponding to each training rhythm window are statistically analyzed to generate a motion execution consistency index.

[0037] Read the abnormal prompt information corresponding to each training rhythm window in the abnormal prompt sequence, and count the number of excessive abdominal pressure prompts, breath-holding prompts, abdominal compensation prompts and postural instability prompts, and generate an abnormal prompt statistical sequence;

[0038] Based on the consistency index of action execution and the statistical sequence of abnormal prompts, the training evaluation index corresponding to each training rhythm window is calculated, and the training score record is generated.

[0039] The beneficial effects of this invention are as follows: By generating an abdominal pressure change sequence, it achieves collaborative analysis of the relationship between respiratory rhythm and abdominal pressure changes based on multi-source signals such as abdominal wall displacement and abdominal belt tension collected by an intelligent abdominal belt sensor. This accurately characterizes the temporal coupling relationship between respiratory phase, abdominal wall movement, and abdominal pressure changes, improving the accuracy of movement recognition and the scientific nature of training guidance during pelvic floor muscle training. Simultaneously, by constructing a training rhythm window set based on the respiratory phase sequence and training quality sequence and calculating the gating confidence to generate a real-time guidance sequence, it achieves dynamic judgment and adaptive guidance of pelvic floor muscle training timing. This allows training prompts to be gated according to the user's real-time respiratory rhythm and abdominal pressure status, avoiding triggering incorrect training movements in cases of abnormal abdominal pressure or disordered respiratory rhythm, thus improving the real-time guidance capability and training effect evaluation capability of the pelvic floor muscle training process. Attached Figure Description

[0040] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0041] Figure 1 This is a flowchart of a method for real-time guidance of pelvic floor muscle movement based on multi-parameter sensing of an abdominal binder.

[0042] Figure 2 A flowchart for generating respiratory phase sequences and abdominal pressure change sequences.

[0043] Figure 3 A flowchart for generating training quality sequences and real-time guidance sequences.

[0044] Figure 4 A flowchart for generating anomaly alert sequences and training score records. Detailed Implementation

[0045] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0046] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0047] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0048] Reference Figures 1-4 This is one embodiment of the present invention, which provides a real-time guidance method for pelvic floor muscle movement based on abdominal binder multi-parameter sensing, including the following steps:

[0049] S1. Collect multi-parameter observation data of the abdominal belt during the period when the user wears the abdominal belt and perform unified time series organization and observation consistency management to generate the abdominal belt observation sequence.

[0050] S1.1 Read the abdominal wall displacement data, abdominal belt tension data and motion disturbance data from the abdominal belt multi-parameter observation data, and sort them according to the corresponding timestamps to generate an initial time series data sequence.

[0051] It should be noted that the multi-parameter observation data of the abdominal belt is a set of data continuously collected by the abdominal belt sensing components during the user's pelvic floor muscle training while wearing the abdominal belt device. It includes abdominal wall displacement data, abdominal belt tension data, motion disturbance data, and timestamp information. Among them, the abdominal wall displacement data is detected by the displacement sensing structure or strain sensitive structure set on the inside of the abdominal belt, which senses the slight deformation of the abdominal wall during breathing and changes in abdominal pressure and converts it into a displacement signal. The abdominal belt tension data is detected by the tension detection structure or pressure detection structure set in the abdominal belt structure, which senses the stretching changes of the abdominal belt caused by the expansion and contraction of the abdomen and converts it into a tension change signal. The motion disturbance data is collected by the inertial detection component set on the abdominal belt, which collects the acceleration change signal generated by the user's body movement to reflect the intensity of body movement or changes in posture. Based on the timestamp information, the abdominal wall displacement data, abdominal belt tension data, and motion disturbance data are organized according to a unified time index to generate an initial time-series data sequence arranged continuously in time.

[0052] S1.2 Perform bandpass filtering and abnormal mutation removal on the initial time series data sequence to generate the abdominal zone observation sequence.

[0053] It should be noted that, according to the time sequence, bandpass filtering is performed on the abdominal wall displacement data and abdominal belt tension data in the initial time series data sequence. Specifically, the abdominal wall displacement data and abdominal belt tension data are decomposed in the frequency domain and the frequency components corresponding to the respiratory rhythm are retained, while low-frequency drift components and high-frequency random disturbance components are filtered out, thereby obtaining the filtered displacement sequence and filtered tension sequence. The displacement change and tension change between adjacent time positions in the filtered displacement sequence and filtered tension sequence are calculated, and the change distribution of displacement change and tension change is statistically analyzed within a continuous time window. The time positions that deviate from the continuous change trend are identified based on the change distribution, and the corresponding data records are identified as abnormal change records. The displacement and tension values ​​corresponding to the abnormal change records are deleted, and linear interpolation compensation is performed based on the normal data values ​​corresponding to the adjacent time positions before and after the abnormal change records, so that the corrected data maintains continuous change in the time dimension, thereby generating the abdominal belt observation sequence.

[0054] It should also be noted that the displacement change is the difference between the abdominal wall displacement data corresponding to the current time position and the abdominal wall displacement data corresponding to the previous time position; the tension change is the difference between the abdominal belt tension data corresponding to the current time position and the abdominal belt tension data corresponding to the previous time position.

[0055] Low-frequency drift components refer to slowly changing signals with a period of change much larger than (e.g., more than about 10 seconds) than the respiratory cycle, while high-frequency random disturbance components refer to rapidly fluctuating signals with a period of change much smaller than (e.g., less than about 0.5 seconds) than the respiratory cycle.

[0056] S2. Identify the respiratory cycle structure of the abdominal band observation sequence and statistically analyze the respiratory cycle parameters to generate a respiratory phase sequence. Using the respiratory phase sequence as a constraint, perform abdominal pressure response reconstruction processing on the abdominal band observation sequence to generate an abdominal pressure change sequence.

[0057] S2.1 Identify the positions of respiratory peaks and troughs in the abdominal wall displacement data in the abdominal band observation sequence, and generate a respiratory peak-trough sequence.

[0058] It should be noted that abdominal wall displacement data from the abdominal band observation sequence are read and a continuous displacement change curve is formed in chronological order. The direction of change of displacement values ​​corresponding to adjacent positions before and after each time point is statistically analyzed in the continuous displacement change curve. When the displacement value corresponding to a certain time point is greater than the displacement value corresponding to the previous time point and greater than the displacement value corresponding to the next time point, the corresponding time point is identified as the respiratory peak position. When the displacement value corresponding to a certain time point is less than the displacement value corresponding to the previous time point and less than the displacement value corresponding to the next time point, the time point is identified as the respiratory trough position. The identified respiratory peak positions and respiratory trough positions are paired and organized in chronological order so that each respiratory peak position corresponds to the adjacent respiratory trough position, thereby generating a respiratory peak-trough sequence.

[0059] S2.2 Obtain the respiratory cycle boundary based on the time interval between adjacent respiratory peaks or troughs in the respiratory peak-trough sequence. Generate a respiratory cycle parameter sequence by calculating the cycle length, inspiratory duration, and expiratory duration of each respiratory cycle.

[0060] It should be noted that the time interval between adjacent respiratory peaks or troughs in the respiratory peak-trough sequence is read, and the time difference between adjacent peaks or troughs is calculated as the boundary of the respiratory cycle. Based on the start and end times of each respiratory cycle, the corresponding cycle length is calculated. Furthermore, based on the relative positions of the respiratory peaks and troughs, the inspiratory duration and expiratory duration of each respiratory cycle are calculated, where the inspiratory duration is the time interval from the position of the respiratory trough to the position of the next respiratory peak, and the expiratory duration is the time interval from the position of the respiratory peak to the position of the next respiratory trough. The cycle length, inspiratory duration, and expiratory duration of each respiratory cycle are organized into a respiratory cycle parameter sequence.

[0061] S2.3. Based on the respiratory cycle parameter sequence, mark each time position in the abdominal band observation sequence as the inspiratory phase, expiratory phase, and end-expiratory phase to generate a respiratory phase sequence.

[0062] It should be noted that the cycle length, inspiratory duration, and expiratory duration of each respiratory cycle in the respiratory cycle parameter sequence are read. If the current time position is between the respiratory trough and the next respiratory peak, the current time position is marked as the inspiratory phase; if the current time position is between the respiratory peak and the next respiratory trough, the current time position is marked as the expiratory phase; if the current time position is at the end of the expiratory phase, that is, in the later part of the time segment between the respiratory peak and the next respiratory trough in the current respiratory cycle, the current time position is marked as the end of the expiratory phase. All time positions in the abdominal band observation sequence are traversed, and a corresponding respiratory phase label is assigned to each time position to generate a respiratory phase sequence containing the inspiratory phase, expiratory phase, and end of the expiratory phase.

[0063] S2.4. Based on the respiratory phase sequence, identify the time segments corresponding to the inspiratory phase, expiratory phase, and end-expiratory phase in the abdominal belt observation sequence, and extract the corresponding abdominal wall displacement data and abdominal belt tension data to generate an abdominal wall motion response sequence.

[0064] It should be noted that the respiratory phase labels corresponding to each time position in the respiratory phase sequence are read, and the abdominal wall displacement data and abdominal belt tension data corresponding to each time position are retrieved in the abdominal belt observation sequence. The continuous time positions labeled as the inspiratory phase are divided into inspiratory time segments, the continuous time positions labeled as the expiratory phase are divided into expiratory time segments, and the continuous time positions labeled as the end of the expiratory phase are divided into end-expiratory time segments. The corresponding abdominal wall displacement data and abdominal belt tension data in each time segment are extracted respectively, and the abdominal wall displacement values ​​and abdominal belt tension values ​​corresponding to the same time position are paired and organized in chronological order to generate an abdominal wall motion response sequence that characterizes the abdominal wall motion state and abdominal belt stress state under different respiratory phases.

[0065] S2.5 Calculate the rate of change of abdominal wall motion and the rate of change of abdominal belt tension in the abdominal wall motion response sequence, and perform continuity constraint and abnormal mutation correction processing to generate the abdominal pressure change sequence.

[0066] It should be noted that abdominal wall displacement data and abdominal band tension data are read from the abdominal wall motion response sequence. The displacement change and tension change between adjacent time positions are calculated in chronological order, and the rate of change at each time position is calculated. The rate of change of abdominal wall motion is the ratio of the change of abdominal wall displacement at adjacent time positions to the time interval, and the rate of change of abdominal band tension is the ratio of the change of abdominal band tension at adjacent time positions to the time interval. A correspondence between the rate of change of abdominal wall motion and the rate of change of abdominal band tension is established at the same time position. Based on the synchronous changes of abdominal wall displacement and abdominal band tension in the same respiratory phase, the abdominal pressure response is generated. An increase in the rate of change of abdominal wall displacement and a continuous increase in abdominal band tension correspond to an increase in abdominal pressure, and a decrease in the rate of change of abdominal wall displacement and a continuous decrease in abdominal band tension correspond to a decrease in abdominal pressure, thus forming an abdominal pressure response sequence.

[0067] The continuous change relationship of the abdominal pressure response sequence is examined along the time direction. When there is a sudden change in the abdominal pressure response at a certain time position compared with the abdominal pressure response at the adjacent time positions, and the sudden change does not conform to the change direction formed by the consecutive time positions, the corresponding time position is marked as an abnormal change position. Abnormal change correction processing is performed on the abdominal pressure response corresponding to each abnormal change position, and interpolation correction is performed based on the abdominal pressure response corresponding to the adjacent normal time positions before and after the abnormal change position, so that the corrected abdominal pressure response remains continuous in the time dimension, thereby generating an abdominal pressure change sequence.

[0068] S3. Extract abdominal pressure features from the abdominal pressure change sequence to form an abdominal pressure feature sequence. Based on the respiratory phase sequence, evaluate the rhythmic stability and abdominal wall stability of the abdominal band observation sequence and the abdominal pressure feature sequence respectively, and generate a training quality sequence.

[0069] S3.1 Calculate the magnitude of abdominal pressure change between adjacent time positions in the abdominal pressure change sequence based on the time index, and calculate the rate of change of abdominal pressure during the rising and falling phases to generate a sequence of abdominal pressure change rates.

[0070] It should be noted that the abdominal pressure response values ​​corresponding to each time position in the abdominal pressure change sequence are read, and the abdominal pressure change amplitude between adjacent time positions is calculated according to the time index order. The abdominal pressure change amplitude is the difference between the abdominal pressure response value at the current time position and the abdominal pressure response value at the previous time position. The direction of abdominal pressure change is identified based on the sign change of the abdominal pressure change amplitude corresponding to consecutive time positions. The time segment where the abdominal pressure change amplitude corresponding to multiple consecutive time positions (such as three or more) is positive is defined as the abdominal pressure rising stage, and the time segment where the abdominal pressure change amplitude corresponding to multiple consecutive time positions is negative is defined as the abdominal pressure falling stage. The ratio of the abdominal pressure change amplitude to the corresponding time interval in each stage is used as the abdominal pressure change rate, and the abdominal pressure change rate sequence is organized according to the time order.

[0071] S3.2 Calculate the maximum, minimum and difference range of the abdominal pressure change sequence within each respiratory cycle, and perform feature combination processing with the abdominal pressure change rate sequence to generate an abdominal pressure feature sequence.

[0072] It should be noted that the abdominal pressure response quantity set corresponding to each respiratory cycle is extracted from the abdominal pressure change sequence; within the abdominal pressure response quantity set corresponding to each respiratory cycle, the maximum and minimum values ​​of the abdominal pressure response quantity are identified, and the difference between the maximum and minimum values ​​is taken as the abdominal pressure fluctuation range of the current respiratory cycle; the abdominal pressure change rate values ​​corresponding to the time segments of each respiratory cycle in the abdominal pressure change rate sequence are read, and a correspondence between the abdominal pressure fluctuation range and the abdominal pressure change rate is established according to the time position, so that each time position is simultaneously associated with the abdominal pressure fluctuation range of the corresponding respiratory cycle and the abdominal pressure change rate at that time position; the abdominal pressure fluctuation range and the abdominal pressure change rate are jointly organized according to a unified time index, so that each time position forms an abdominal pressure feature containing information on the amplitude and rate of abdominal pressure change, thereby generating an abdominal pressure feature sequence.

[0073] S3.3. Identify the time segments corresponding to each respiratory cycle in the abdominal belt observation sequence based on the respiratory phase sequence, and perform stage segmentation processing on the abdominal belt observation sequence and abdominal pressure characteristic sequence to generate a respiratory stage data set.

[0074] It should be noted that, based on the respiratory phase labels in the respiratory phase sequence, the start and end times of each respiratory cycle are identified according to the cyclical changes in time sequence. The time range between the start times of two adjacent respiratory cycles constitutes a respiratory cycle time segment. The abdominal wall displacement data and abdominal band tension data corresponding to each respiratory cycle time segment are extracted in the abdominal band observation sequence, and the abdominal pressure features within the corresponding time segment are extracted in the abdominal pressure feature sequence. The abdominal wall displacement data, abdominal band tension data, and abdominal pressure features within the same respiratory cycle are divided into inspiratory phase data, expiratory phase data, and end-expiratory phase data, respectively, and organized accordingly in time sequence, thereby generating a respiratory phase data set containing respiratory cycle identifiers, phase type identifiers, abdominal wall displacement data, abdominal band tension data, and abdominal pressure features.

[0075] S3.4. Based on the respiratory cycle parameter sequence, calculate the cycle deviation and phase duration deviation between adjacent respiratory cycles to generate a rhythm stability parameter sequence.

[0076] It should be noted that the cycle length, inspiratory duration, and expiratory duration corresponding to each respiratory cycle in the respiratory cycle parameter sequence are read, and the difference in cycle length between adjacent respiratory cycles is compared in chronological order to obtain the cycle deviation; at the same time, the difference in inspiratory duration and the difference in expiratory duration in the corresponding phases of adjacent respiratory cycles are compared to obtain the phase duration deviation; the cycle deviation and phase duration deviation are organized in the order of respiratory cycles so that each respiratory cycle corresponds to a set of rhythm stability parameter records, and all rhythm stability parameter records are integrated to generate a rhythm stability parameter sequence.

[0077] S3.5 By reading the abdominal wall displacement data in the respiratory phase dataset, calculate the changes in abdominal wall displacement amplitude, displacement change rate, and displacement fluctuation degree in each respiratory phase, and combine them with the rhythm stability parameter sequence according to the time index to generate a training quality sequence.

[0078] It should be noted that, based on the respiratory phase labels, abdominal wall displacement data records corresponding to the inspiratory phase, expiratory phase, and end-expiratory phase are extracted from the respiratory phase dataset. The maximum and minimum values ​​of abdominal wall displacement are identified in the abdominal wall displacement data records corresponding to each respiratory phase, and the difference between the maximum and minimum values ​​is calculated as the change in abdominal wall displacement amplitude. The ratio of the change in abdominal wall displacement between adjacent time positions to the corresponding time interval is calculated in chronological order to obtain the rate of change in abdominal wall displacement. The standard deviation of the change in abdominal wall displacement within each respiratory phase is used as the degree of fluctuation in abdominal wall displacement. The change in abdominal wall displacement amplitude, the rate of change in abdominal wall displacement, and the degree of fluctuation in abdominal wall displacement are used as abdominal wall motion features, and are aligned and combined with the rhythm stability parameters in the rhythm stability parameter sequence according to the time index to generate a training quality sequence.

[0079] S4. Based on the breathing phase sequence and the training quality sequence, construct a training rhythm window set, calculate the gating confidence of each training rhythm window in the training rhythm window set, and generate a real-time guidance sequence.

[0080] S4.1 Identify the time segment corresponding to the end of the exhalation phase of the respiratory phase sequence, and determine the training effectiveness based on the training quality sequence to generate a candidate training window sequence.

[0081] It should be noted that, based on the breathing phase labels corresponding to each time position in the breathing phase sequence, the training quality records at the corresponding time positions in the training quality sequence are read within each end-expiration time interval, and the changes in the training quality records in the time dimension within the current time interval are statistically analyzed. When the training quality records within a certain end-expiration time interval maintain a consistent change state within a continuous time range, the current end-expiration time interval is determined as the effective training time interval. All effective training time intervals are organized in chronological order to generate a candidate training window sequence.

[0082] It should also be noted that a continuous time range refers to a time segment consisting of multiple consecutive sampling time positions within the end-expiratory phase, such as the time range corresponding to three or more consecutive sampling time positions.

[0083] Within a continuous time range, if the direction of change of training quality parameters corresponding to adjacent time positions in the training quality sequence remains consistent, then the current training quality record is determined to satisfy a consistent change state.

[0084] The parameters of abdominal wall displacement amplitude change, abdominal wall displacement change rate, abdominal wall displacement fluctuation degree and rhythm stability are organized according to a unified time index, so that each time position forms a corresponding training quality value. The training quality value corresponding to each time position is defined as the training quality parameter. The training quality parameter is the value of a single time position, the training quality record is a set of data records corresponding to a certain time position, and the training quality sequence is all training quality records organized in chronological order.

[0085] S4.2. Based on the phase transition position of the breathing phase sequence and the stable interval of the training quality sequence in the candidate training window sequence, perform window expansion and time boundary adjustment for each time segment to generate a training rhythm window set.

[0086] It should be noted that the process involves reading the time segments corresponding to each candidate training window in the candidate training window sequence, and retrieving the positions of breathing phase label changes adjacent to the current time segment in the breathing phase sequence. The time position where the breathing phase label changes from one phase to another is determined as the breathing phase transition position. The time range of the corresponding breathing phase is determined between the breathing phase transition positions, and the training quality records corresponding to this time range in the training quality sequence are read. By statistically analyzing the changes in training quality records at continuous time positions, the time interval where the training quality remains stable is identified and used as the training stability interval. Centered on the time segment corresponding to the candidate training window, and combined with the breathing phase boundary and the time range of the training stability interval, the start and end time positions of the candidate training window are expanded and adjusted in terms of time boundary, so that the adjusted time segments are simultaneously within the training stability interval and do not cross adjacent breathing phase transition positions. The time segments adjusted by the time boundary are organized in chronological order to generate a training rhythm window set.

[0087] It should also be noted that the training quality record maintaining stable change means that the direction of change of the training quality parameter is consistent or the change amplitude is within a similar range at consecutive time positions. For example, the direction of change of the training quality parameter is the same or the difference in the change amplitude does not exceed a certain proportion of the average change amplitude of that period (such as 10% to 30%). This is set based on the statistical characteristics that the change amplitude of respiratory and abdominal pressure signals at adjacent sampling points is usually small under normal physiological conditions, while body movement interference or abnormal fluctuations will significantly increase the change amplitude. It is used to distinguish between normal continuous change and abnormal fluctuation.

[0088] S4.3 Calculate the signal consistency score between the abdominal band observation sequence and the abdominal pressure feature sequence in the training rhythm window set, and perform interference suppression correction processing in combination with motion interference data to generate a gated confidence sequence.

[0089] The expression for calculating the signal consistency score is:

[0090] ;

[0091] in, This represents the signal consistency score corresponding to the training rhythm window. This indicates the number of time positions contained within the current training rhythm window. Indicates the first Rate of change of abdominal wall movement at each time point Indicates the first The rate of change of abdominal pressure at each time point This represents a very small positive number, used to avoid division by zero when the denominator is close to zero at the same time as the rate of change of abdominal wall movement and the rate of change of abdominal pressure. For example, it can be 0.0000001.

[0092] It should be noted that motion interference data in the abdominal band observation sequence is read, and the change amplitude of motion interference data at each time position is calculated as the motion intensity. The corresponding motion intensity index is obtained by statistically analyzing the average value of motion intensity within the training rhythm window. Based on the motion intensity index, interference suppression correction processing is performed on the signal consistency score so that the signal consistency score corresponding to the training rhythm window with higher motion intensity is lower. The gating confidence corresponding to each training rhythm window is obtained, and the gating confidence sequence is organized in chronological order.

[0093] S4.4. Based on the gating confidence sequence, perform gating output judgment on the training rhythm window set to form a real-time guidance sequence.

[0094] It should be noted that the time segments corresponding to each training rhythm window in the training rhythm window set are read, and the gated confidence records corresponding to each time segment are obtained from the gated confidence sequence; the average value of the gated confidence within each time segment is calculated as the window confidence corresponding to the current training rhythm window, and the difference between the maximum and minimum values ​​of the gated confidence within each training rhythm window is calculated as the window's built-in confidence fluctuation value; the average gated confidence corresponding to all training rhythm windows is sorted according to the numerical order, and the median value of the sorting results is determined as the window confidence criterion. Simultaneously, all windows... The built-in confidence fluctuation values ​​are sorted according to their numerical magnitude, and the median value of the sorted results is determined as the window stability criterion. When the average gating confidence of a training rhythm window is not lower than the window confidence criterion, and the corresponding built-in confidence fluctuation value of the window is not higher than the window stability criterion, the training rhythm window is determined as a valid training window; otherwise, it is determined as an invalid training window. The time position of the valid training window in the respiratory cycle and the corresponding window type identifier are integrated to generate training action prompt records, and the training action prompt records are organized in chronological order to generate a real-time guidance sequence.

[0095] S5. Perform anomaly detection processing on the real-time guidance sequence, generate anomaly prompt sequence, and extract training evaluation indicators for each training rhythm window to generate training score records.

[0096] S5.1 Based on the real-time guidance sequence, analyze the changing trend of the abdominal pressure characteristic sequence and the stage matching relationship of the respiratory phase sequence, and identify abnormal rise in abdominal pressure, respiratory stage mismatch and respiratory interruption state, and generate abdominal pressure abnormality marker sequence and respiratory abnormality marker sequence.

[0097] It should be noted that the direction of change of the rate of change of abdominal pressure is statistically analyzed in chronological order at consecutive time points. When the rate of change of abdominal pressure is positive at multiple consecutive time points (e.g., three or more time points) and the corresponding range of abdominal pressure fluctuation continues to expand, the corresponding time segment is identified as the abnormal rise segment of abdominal pressure, and an abnormal abdominal pressure label is assigned to the corresponding time position in the abnormal abdominal pressure labeling sequence. The corresponding action type of each training action prompt in the real-time guidance sequence is read and matched with the breathing phase label at the corresponding time position. When the action type requires performing a contraction or holding action and the breathing phase label corresponds to the inspiratory phase, or when the action type requires performing a relaxation action and the breathing phase label corresponds to the end of expiration, the action type is identified. During the terminal phase, the corresponding time position is determined as the respiratory phase mismatch position, and a respiratory phase mismatch mark is assigned to the respiratory abnormality mark sequence. At the same time, the conversion relationship between adjacent respiratory phase labels is counted in the respiratory phase sequence. When the respiratory phase label remains unchanged for a long time within a certain time segment and the duration exceeds the average duration of multiple adjacent respiratory cycles, the time segment is determined as the respiratory interruption segment, and a respiratory interruption mark is assigned to the time position corresponding to the time segment in the respiratory abnormality mark sequence. All abdominal pressure abnormality marks are organized in chronological order to form an abdominal pressure abnormality mark sequence, and all respiratory phase mismatch marks and respiratory interruption marks are organized in chronological order to form a respiratory abnormality mark sequence.

[0098] S5.2 Analyze the changes in abdominal wall displacement amplitude, abdominal belt tension fluctuation degree and body motion intensity corresponding to motion interference data in the abdominal belt observation sequence, and identify abnormal increase in abdominal wall motion, enhanced abdominal compensation and posture interference state, and generate an abnormal abdominal wall marker sequence.

[0099] It should be noted that the abdominal wall displacement amplitude change record corresponding to the real-time guidance sequence is read, and the abdominal belt tension data and motion disturbance data in the abdominal belt observation sequence are read simultaneously. The standard deviation of the difference between the abdominal belt tension value and the average abdominal belt tension value in the corresponding time segment is calculated as the degree of abdominal belt tension fluctuation. The amplitude of change between adjacent time positions is calculated for the motion disturbance data in chronological order, and the average amplitude is statistically analyzed within a continuous time window to characterize the body motion intensity of the corresponding time segment. After obtaining abnormal judgment characteristics such as abdominal wall displacement amplitude change, abdominal belt tension fluctuation degree, and body motion intensity, the average value of the corresponding characteristic values ​​for multiple adjacent respiratory cycles (e.g., three or more respiratory cycles) is statistically analyzed. As a reference feature value, the abnormal judgment feature value of the current time segment is compared with the reference feature value. Among them, the change in abdominal wall displacement amplitude is used to identify abnormal increase in abdominal wall movement, the degree of abdominal belt tension fluctuation is used to identify enhanced abdominal compensation, and the intensity of body movement is used to identify the state of posture disturbance. If the abnormal judgment feature value of the current time segment is higher than the corresponding reference feature value at multiple consecutive time positions (e.g., three or more time positions), the current time segment is determined as an abnormal change segment. Each identified abnormal segment is assigned a corresponding abnormal label according to the abnormality type, including abdominal wall movement abnormality label, abdominal compensation abnormality label, and posture disturbance abnormality label, and organized in chronological order to generate an abdominal wall abnormality label sequence.

[0100] S5.3. Based on the abnormal abdominal pressure marker sequence, abnormal breathing marker sequence, and abnormal abdominal wall marker sequence, determine the abnormality type of each training rhythm window and output the abnormality prompt sequence.

[0101] It should be noted that the time segments corresponding to each training rhythm window in the training rhythm window set are read, and the corresponding abnormal marker records are extracted from the abdominal pressure abnormal marker sequence, respiratory abnormal marker sequence, and abdominal wall abnormal marker sequence; the occurrence frequency of each abnormal marker is counted within the time segment corresponding to each training rhythm window; the occurrence frequency of abdominal pressure abnormal markers, respiratory abnormal markers, and abdominal wall abnormal markers are compared with the total number of time positions within the training rhythm window, respectively. When a certain type of abnormal marker appears in multiple consecutive time positions (e.g., three or more time positions) within the training rhythm window, the current training rhythm window is determined as an abnormal training window of the corresponding abnormal type; corresponding abnormal prompt records are generated for each identified abnormal training window according to the abnormal type, and organized in chronological order to output the abnormal prompt sequence.

[0102] S5.4 Statistically analyze the hit rate of the contraction allowed window, the hold allowed window, and the relaxation allowed window corresponding to each training rhythm window, and generate an action execution consistency index.

[0103] It should be noted that the time segments corresponding to each training rhythm window in the training rhythm window set are read, and the corresponding training action prompt records are extracted from the real-time guidance sequence. Simultaneously, abdominal wall displacement and abdominal pressure change records at the corresponding time positions are read from the abdominal belt observation sequence. Based on the action type corresponding to the training action prompt records, the training rhythm windows are divided into contraction-allowed windows, hold-allowed windows, and relaxation-allowed windows. Within each type of window, the direction of abdominal wall displacement change and the direction of abdominal pressure change are statistically analyzed to ensure they meet the corresponding action requirements. Specifically, when the training action prompt record is a contraction action, the direction of abdominal wall displacement change is analyzed to be in the contraction direction, and the abdominal pressure change shows an upward trend. The number of time positions is calculated; when the training action prompt is recorded as a hold action, the number of time positions where the abdominal wall displacement remains stable in a continuous time position and the abdominal pressure changes within a small fluctuation range is counted; when the training action prompt is recorded as a relaxation action, the number of time positions where the abdominal wall displacement changes in the direction of relaxation and the abdominal pressure changes show a decreasing trend is counted; the ratio of the number of time positions that meet the action requirements to the total number of time positions in the corresponding training rhythm window is calculated to obtain the contraction allowable window hit rate, hold allowable window hit rate, and relaxation allowable window hit rate, and these are organized according to the order of the training rhythm windows to generate an action execution consistency index.

[0104] S5.5 Read the abnormal prompt information corresponding to each training rhythm window in the abnormal prompt sequence, and count the number of excessive abdominal pressure prompts, breath-holding prompts, abdominal compensation prompts, and postural instability prompts, and generate an abnormal prompt statistical sequence.

[0105] It should be noted that the time position and abnormality type identifier corresponding to each abnormality prompt record in the abnormality prompt sequence are read, and the training rhythm window corresponding to each time position is retrieved in the training rhythm window set; the abnormality prompt records are classified according to the abnormality type identifier. When the abnormality type identifier corresponds to an abnormal increase in abdominal pressure, the current abnormality prompt record is defined as an excessive abdominal pressure prompt record; when the abnormality type identifier corresponds to a respiratory interruption or respiratory phase mismatch, the current abnormality prompt record is defined as a breath-holding prompt record; when the abnormality type identifier corresponds to an enhanced abdominal compensation state, the current abnormality prompt record is defined as an abdominal compensation prompt record; when the abnormality type identifier corresponds to a postural disturbance state, the current abnormality prompt record is identified as a postural instability prompt record; the occurrence frequency of the four types of abnormality prompt records is counted in the time segment corresponding to each training rhythm window, and the excessive abdominal pressure prompt, breath-holding prompt, abdominal compensation prompt, and postural instability prompt frequency corresponding to each training rhythm window are obtained; the frequency of the four types of abnormality prompts corresponding to each training rhythm window is organized according to the order of the training rhythm window to generate an abnormality prompt statistical sequence.

[0106] S5.6 Based on the consistency index of action execution and the statistical sequence of abnormal prompts, calculate the training evaluation index corresponding to each training rhythm window and generate training score records.

[0107] It should be noted that the corresponding window identifier, action execution consistency index, number of four types of abnormal prompts and training evaluation index should be recorded in the time sequence of the training rhythm window, and a training score record should be formed.

[0108] The expression for calculating the training evaluation index is:

[0109] ;

[0110] ;

[0111] ;

[0112] ;

[0113] ;

[0114] in, Indicates the first Training evaluation metrics corresponding to each training rhythm window. Indicates the first The consistency index of action execution corresponding to each training rhythm window. Indicates the first The number of excessive abdominal pressure alerts corresponding to each training rhythm window. Indicates the first The number of breath-holding prompts corresponding to each training rhythm window. Indicates the first The number of abdominal compensation prompts corresponding to each training rhythm window. Indicates the first The number of posture instability alerts corresponding to each training rhythm window. Indicates the first The total number of time positions within each training rhythm window. Indicates the first Density of abnormal alerts corresponding to each training rhythm window Indicates the first The average density of the four types of abnormality alerts in each training rhythm window. Indicates the first The dispersion of the density of four types of abnormality cues within a training rhythm window. Indicates the first Multiple abnormal concurrent coupling strengths within a training rhythm window This represents the natural index.

[0115] In summary, this invention achieves the collaborative analysis of the relationship between respiratory rhythm and abdominal pressure changes by generating an abdominal pressure change sequence, based on multi-source signals such as abdominal wall displacement and abdominal belt tension collected by an intelligent abdominal belt sensor. This accurately characterizes the temporal coupling relationship between respiratory phase, abdominal wall movement, and abdominal pressure changes, improving the accuracy of movement recognition and the scientific nature of training guidance during pelvic floor muscle training. Simultaneously, by constructing a training rhythm window set based on the respiratory phase sequence and training quality sequence and calculating gating confidence to generate a real-time guidance sequence, it achieves dynamic determination and adaptive guidance of pelvic floor muscle training timing. This allows training prompts to be gated according to the user's real-time respiratory rhythm and abdominal pressure status, avoiding triggering incorrect training movements under abnormal abdominal pressure or disordered respiratory rhythm, thus enhancing the real-time guidance capability and training effect evaluation capability of the pelvic floor muscle training process.

[0116] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for real-time guidance of pelvic floor muscle exercise based on multi-parameter sensing of an abdominal belt, characterized in that, include: Collect multi-parameter observation data of the abdominal belt during the period when the user wears the abdominal belt and perform unified time series organization and observation consistency management to generate abdominal belt observation sequence; The respiratory cycle structure of the abdominal binder observation sequence is identified and respiratory cycle parameters are statistically analyzed to generate a respiratory phase sequence. Using the respiratory phase sequence as a constraint, abdominal pressure response reconstruction processing is performed on the abdominal binder observation sequence to generate an abdominal pressure change sequence. The abdominal pressure features of the abdominal pressure change sequence are extracted to form an abdominal pressure feature sequence. Based on the respiratory phase sequence, the rhythmic stability and abdominal wall stability of the abdominal band observation sequence and the abdominal pressure feature sequence are evaluated respectively to generate a training quality sequence. Based on the respiratory phase sequence and the training quality sequence, a training rhythm window set is constructed, the gating confidence of each training rhythm window in the training rhythm window set is calculated, and a real-time guidance sequence is generated. Anomaly detection processing is performed on the real-time guidance sequence to generate anomaly prompt sequences. At the same time, training evaluation indicators are extracted from each training rhythm window to generate training score records.

2. The real-time guidance method of pelvic floor muscle exercise based on abdominal belt multi-parameter sensing according to claim 1, wherein, The specific steps for generating the abdominal band observation sequence are as follows: Read abdominal wall displacement data, abdominal belt tension data and motion disturbance data from the multi-parameter observation data of the abdominal belt, and sort them according to the corresponding timestamps to generate an initial time series data sequence; Bandpass filtering and abnormal mutation removal are performed on the initial time series data sequence to generate the abdominal zone observation sequence.

3. The real-time guidance method of pelvic floor muscle exercise based on abdominal belt multi-parameter sensing according to claim 2, wherein, The specific steps for generating the respiratory phase sequence are as follows: Identify the positions of respiratory peaks and troughs in abdominal wall displacement data in abdominal band observation sequences, and generate respiratory peak-trough sequences. The respiratory cycle boundary is obtained by the time interval between adjacent respiratory peaks or troughs in the respiratory peak-trough sequence. The respiratory cycle parameter sequence is generated by calculating the cycle length, inspiratory duration and expiratory duration of each respiratory cycle. Based on the respiratory cycle parameter sequence, each time position in the abdominal band observation sequence is marked as the inspiratory phase, expiratory phase, and end-expiratory phase, generating a respiratory phase sequence.

4. The real-time guidance method of pelvic floor muscle exercise based on abdominal belt multi-parameter sensing according to claim 2 or 3, characterized in that, The specific steps for generating the abdominal pressure change sequence are as follows: Based on the respiratory phase sequence, the time segments corresponding to the inspiratory phase, expiratory phase and end-expiratory phase in the abdominal belt observation sequence are identified, and the corresponding abdominal wall displacement data and abdominal belt tension data are extracted to generate an abdominal wall motion response sequence. The rate of change of abdominal wall motion and the rate of change of abdominal belt tension in the abdominal wall motion response sequence are calculated, and continuity constraints and abnormal mutation correction are performed to generate an abdominal pressure change sequence.

5. The real-time guidance method of pelvic floor muscle exercise based on abdominal belt multi-parameter sensing of claim 4, wherein, The specific steps for forming the abdominal pressure feature sequence are as follows: The magnitude of abdominal pressure change between adjacent time positions in the abdominal pressure change sequence is calculated based on the time index, and the rate of change during the abdominal pressure increase and decrease phases is calculated to generate an abdominal pressure change rate sequence. The maximum, minimum, and difference range of abdominal pressure change sequence within each respiratory cycle are statistically analyzed, and feature combination processing is performed with the abdominal pressure change rate sequence to generate an abdominal pressure feature sequence.

6. The real-time guidance method of pelvic floor muscle exercise based on abdominal belt multi-parameter sensing of claim 5, wherein, The specific steps for generating the training quality sequence are as follows: Based on the respiratory phase sequence, the time segment corresponding to each respiratory cycle in the abdominal band observation sequence is identified, and the abdominal band observation sequence and abdominal pressure characteristic sequence are segmented into stages to generate a respiratory phase data set. Based on the respiratory cycle parameter sequence, calculate the cycle deviation and phase duration deviation between adjacent respiratory cycles to generate a rhythm stability parameter sequence. By reading abdominal wall displacement data from the respiratory phase dataset, the changes in abdominal wall displacement amplitude, displacement rate, and displacement fluctuation degree in each respiratory phase are calculated, and then combined with the rhythm stability parameter sequence according to the time index to generate a training quality sequence.

7. The real-time guidance method of pelvic floor muscle exercise based on abdominal belt multi-parameter sensing of claim 1, wherein, The specific steps for constructing the training rhythm window set are as follows: Identify the time segment corresponding to the end of the exhalation phase in the respiratory phase sequence, determine the training effectiveness based on the training quality sequence, and generate a candidate training window sequence; Based on the phase transition position of the breathing phase sequence and the stable interval of the training quality sequence in the candidate training window sequence, window expansion and time boundary adjustment are performed on each time segment to generate a training rhythm window set.

8. The real-time guidance method of pelvic floor muscle exercise based on abdominal belt multi-parameter sensing of claim 7, wherein, The specific steps for generating the real-time guidance sequence are as follows: Calculate the signal consistency score between the abdominal band observation sequence and the abdominal pressure feature sequence in the training rhythm window set, and perform interference suppression correction processing in combination with motion interference data to generate a gated confidence sequence; Based on the gating confidence sequence, gating output judgment is performed on the training rhythm window set to form a real-time guidance sequence.

9. The real-time guidance method of pelvic floor muscle exercise based on abdominal belt multi-parameter sensing of claim 8, wherein, The specific steps for generating the anomaly message sequence are as follows: Based on the real-time guidance sequence, the changing trend of the abdominal pressure characteristic sequence and the phase matching relationship of the respiratory phase sequence are analyzed, and abnormal rise in abdominal pressure, respiratory phase mismatch and respiratory interruption are identified, generating abnormal abdominal pressure marker sequence and abnormal respiratory marker sequence. The changes in abdominal wall displacement amplitude, the degree of abdominal belt tension fluctuation, and the body motion intensity corresponding to motion interference data in the abdominal belt observation sequence were analyzed. Abnormal increase in abdominal wall motion, enhanced abdominal compensation, and postural interference were identified, and an abnormal abdominal wall marker sequence was generated. Based on the abnormal abdominal pressure marker sequence, abnormal breathing marker sequence, and abnormal abdominal wall marker sequence, the abnormality type of each training rhythm window is determined, and an abnormality prompt sequence is output.

10. The real-time guidance method for pelvic floor muscle movement based on abdominal belt multi-parameter sensing as described in claim 7 or 9, characterized in that, The specific steps for generating the training score record are as follows: The contraction allowable window hit rate, hold allowable window hit rate, and relaxation allowable window hit rate corresponding to each training rhythm window are statistically analyzed to generate a motion execution consistency index. Read the abnormal prompt information corresponding to each training rhythm window in the abnormal prompt sequence, and count the number of excessive abdominal pressure prompts, breath-holding prompts, abdominal compensation prompts and postural instability prompts, and generate an abnormal prompt statistical sequence; Based on the consistency index of action execution and the statistical sequence of abnormal prompts, the training evaluation index corresponding to each training rhythm window is calculated, and the training score record is generated.

Images