Intelligent traditional Chinese medicine rehabilitation monitoring method and system based on multi-modal data fusion

Abstract

The invention provides an intelligent traditional Chinese medicine rehabilitation monitoring method and system based on multi-modal data fusion, and relates to the technical field of intelligent medical treatment, and the method comprises the steps: obtaining user detection data at a plurality of moments of a monitoring period; rehabilitation scheme information and historical user detection data are acquired; according to the historical user detection data, the user detection data and the rehabilitation scheme information, determining a traditional Chinese medicine rehabilitation state coefficient; determining a comprehensive prediction risk coefficient according to the traditional Chinese medicine rehabilitation state coefficient; and monitoring according to the traditional Chinese medicine rehabilitation state coefficient and the comprehensive prediction risk coefficient. According to the invention, effectiveness, comprehensiveness and logicality of traditional Chinese medicine rehabilitation monitoring can be improved.

Description

Technical Field

[0001] This invention relates to the field of smart healthcare technology, and in particular to a smart TCM rehabilitation monitoring method and system based on multimodal data fusion. Background Technology

[0002] In related technologies, current TCM rehabilitation monitoring relies heavily on physicians' subjective assessments (e.g., observation, auscultation, inquiry, and palpation) or simple measurements of single physiological parameters (e.g., heart rate, blood oxygen). It lacks an objective, continuous, and multi-dimensional quantitative system, making it difficult to provide rehabilitation treatment with adjustment suggestions that are consistent with TCM logic and are interpretable. In other words, related technologies are unable to improve the effectiveness, comprehensiveness, and logic of TCM rehabilitation monitoring.

[0003] The information disclosed in the background section of this application is intended only to enhance the understanding of the general background of this application and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0004] This invention provides a smart TCM rehabilitation monitoring method and system based on multimodal data fusion, which can solve the technical problem that related technologies are unable to improve the effectiveness, comprehensiveness and logic of TCM rehabilitation monitoring.

[0005] According to a first aspect of the present invention, a smart TCM rehabilitation monitoring method based on multimodal data fusion is provided, comprising: Acquire user detection data at multiple points during the monitoring period; Obtain rehabilitation program information and historical user testing data; The TCM rehabilitation status coefficient is determined based on the historical user detection data, the user detection data, and the rehabilitation plan information. Based on the aforementioned TCM rehabilitation status coefficient, a comprehensive predictive risk coefficient is determined; Monitoring is conducted based on the TCM rehabilitation status coefficient and the comprehensive prediction risk coefficient.

[0006] According to the present invention, the TCM rehabilitation status coefficient is determined based on the historical user detection data, the user detection data, and the rehabilitation program information, including: Based on the user detection data, physiological regulation data, local circulation data, and recovery execution data are determined. The physiological regulation data includes: low-frequency power of heart rate variability, low-frequency power of ideal heart rate variability, and short-term coefficient of variation of RR interval. The local circulation data includes: average skin temperature at the rehabilitation site, average skin temperature at the corresponding point on the healthy side, change in average skin temperature at the rehabilitation site, and change in average skin temperature at the corresponding point on the healthy side. The recovery execution data includes: actual completed activity amount, visual analog value of pain, and joint range of motion achievement rate. Based on the historical user detection data, determine the historical average heart rate variability low-frequency power; Based on the rehabilitation plan information, determine the target activity level and rehabilitation stage; Based on the target activity level, the historical average heart rate variability low-frequency power, the rehabilitation stage, the physiological regulation data, the local circulation data, and the recovery execution data, a first dynamic weight, a second dynamic weight, and a third dynamic weight are determined. The TCM rehabilitation status coefficient is determined based on the first dynamic weight, the second dynamic weight, the third dynamic weight, the target activity level, the historical average heart rate variability low-frequency power, the physiological regulation data, the local circulation data, and the recovery execution data.

[0007] According to the present invention, determining a first dynamic weight, a second dynamic weight, and a third dynamic weight based on the target activity level, the historical average heart rate variability low-frequency power, the rehabilitation stage, the physiological regulation data, the local circulation data, and the recovery execution data includes: Based on the described rehabilitation stages, determine the function values ​​for the first rehabilitation stage, the second rehabilitation stage, and the third rehabilitation stage. Based on the target activity level, the historical average heart rate variability low-frequency power, the physiological regulation data, the local circulation data, and the recovery execution data, the first dimension confidence level, the second dimension confidence level, and the third dimension confidence level are determined. Based on the function values ​​of the first rehabilitation stage, the second rehabilitation stage, the third rehabilitation stage, the confidence level of the first dimension, the confidence level of the second dimension, and the confidence level of the third dimension, the first dynamic weight, the second dynamic weight, and the third dynamic weight are determined.

[0008] According to the present invention, determining the first-dimensional confidence level, the second-dimensional confidence level, and the third-dimensional confidence level based on the target activity level, the historical average heart rate variability low-frequency power, the physiological regulation data, the local circulation data, and the recovery execution data includes: according to the formula:

[0009] Determine the confidence level of the first dimension at time i in the monitoring period. Second dimension confidence and the third dimension confidence Where min is the function for finding the minimum value. and For preset coefficients, Let be the short-term variation coefficient of the RR interval at time i in the monitoring period. The low-frequency power of heart rate variability at time i in the monitoring cycle. For a preset time period, The low-frequency power of heart rate variability at a preset time point before the i-th moment of the monitoring cycle. For ideal heart rate variability, low-frequency power, The historical average heart rate variability low-frequency power, This represents the average skin temperature change at the rehabilitation site at time i during the monitoring period. This represents the average skin temperature change at the corresponding point on the healthy side at time i of the monitoring cycle. The average skin temperature at the rehabilitation site at time i of the monitoring period. The average skin temperature at the corresponding point on the healthy side at the i-th moment of the monitoring cycle. For the target activity level, This represents the actual amount of activity completed at time i during the monitoring period. The visual simulation value of pain at the i-th moment of the monitoring period.

[0010] According to the present invention, a TCM rehabilitation state coefficient is determined based on the first dynamic weight, the second dynamic weight, the third dynamic weight, the target activity level, the historical average heart rate variability low-frequency power, the physiological regulation data, the local circulation data, and the recovery execution data, including: according to the formula:

[0011] Determine the TCM rehabilitation status coefficient at time i of the monitoring period. ,in, , and Here, is the preset coefficient, and max is the function to find the maximum value. The first dynamic weight at the i-th moment of the monitoring period, The second dynamic weight is the weight at the i-th time point of the monitoring period. The third dynamic weight is the value at the i-th time point of the monitoring period. The low-frequency power of heart rate variability at time i in the monitoring cycle. For ideal heart rate variability, low-frequency power, Low-frequency power of historical average heart rate variability The short-term variation coefficient of the RR interval at time i in the monitoring period. The average skin temperature at the rehabilitation site at time i of the monitoring period. The average skin temperature at the corresponding point on the healthy side at the i-th moment of the monitoring cycle. This represents the average skin temperature change at the rehabilitation site at time i during the monitoring period. This represents the average skin temperature change at the corresponding point on the healthy side. For the target activity level, This represents the actual amount of activity completed at time i during the monitoring period. The visual analog value of pain at time i in the monitoring cycle. This represents the joint range of motion compliance rate at the i-th moment of the monitoring cycle.

[0012] According to the present invention, a comprehensive predictive risk coefficient is determined based on the TCM rehabilitation status coefficient, including: Based on the TCM rehabilitation state coefficient, determine the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, and the recovery execution sub-dimension coefficient; Based on the physiological regulation subdimension coefficient, the local circulation subdimension coefficient, and the recovery execution subdimension coefficient, determine the standard deviation of the physiological regulation subdimension coefficient, the average value of the physiological regulation subdimension coefficient, the standard deviation of the local circulation subdimension coefficient, the average value of the local circulation subdimension coefficient, the standard deviation of the recovery execution subdimension coefficient, and the average value of the recovery execution subdimension coefficient within a historical preset time period. Determine the total number of samples, the number of first-dimension rises, the number of second-dimension rises, and the number of third-dimension rises for the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, and the recovery execution sub-dimension coefficient within the historical preset time period; The comprehensive prediction risk coefficient is determined based on the first dynamic weight, the second dynamic weight, the third dynamic weight, the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, the recovery execution sub-dimension coefficient, the standard deviation of the physiological regulation sub-dimension coefficient, the average value of the physiological regulation sub-dimension coefficient, the standard deviation of the local circulation sub-dimension coefficient, the average value of the local circulation sub-dimension coefficient, the standard deviation of the recovery execution sub-dimension coefficient, the average value of the recovery execution sub-dimension coefficient, the total number of samplings, the number of times the first dimension rebounds, the number of times the second dimension rebounds, and the number of times the third dimension rebounds.

[0013] According to the present invention, a comprehensive prediction risk coefficient is determined based on the first dynamic weight, the second dynamic weight, the third dynamic weight, the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, the recovery execution sub-dimension coefficient, the standard deviation of the physiological regulation sub-dimension coefficient, the average value of the physiological regulation sub-dimension coefficient, the standard deviation of the local circulation sub-dimension coefficient, the average value of the local circulation sub-dimension coefficient, the standard deviation of the recovery execution sub-dimension coefficient, the average value of the recovery execution sub-dimension coefficient, the total number of samplings, the number of times the first dimension rebounds, the number of times the second dimension rebounds, and the number of times the third dimension rebounds, including: according to the formula:

[0014] Determine the comprehensive prediction risk coefficient at time i of the monitoring period. ,in, As the first preset weight, The second preset weight is K, which is a preset constant. The first dynamic weight at the i-th moment of the monitoring period, The second dynamic weight is the weight at the i-th time point of the monitoring period. The third dynamic weight is the value at the i-th time point of the monitoring period. To find the maximum value function, The physiological regulation subdimension coefficients at the i-th time point of the monitoring cycle. The length of the historical time period is preset. The physiological regulation subdimension coefficients are the values ​​corresponding to the historical preset time period before the i-th moment of the monitoring cycle. This refers to the number of times the first dimension rebounds within the historical preset time period corresponding to the i-th moment of the monitoring period. This represents the total number of samples taken within the historical preset time period corresponding to the i-th moment of the monitoring period. The standard deviation of the physiological regulation subdimension coefficient within the historical preset time period corresponding to the i-th moment of the monitoring cycle. This represents the average value of the physiological regulation subdimension coefficients within the historical preset time period corresponding to the i-th moment of the monitoring cycle. For the local cyclic subdimension coefficients at the i-th time point of the monitoring period, The local cyclic subdimension coefficients are the values ​​corresponding to the historical preset time periods prior to the i-th moment of the monitoring period. This refers to the number of times the second dimension rebounds within the historical preset time period corresponding to the i-th moment of the monitoring period. The standard deviation of the local cyclic subdimension coefficient within the historical preset time period corresponding to the i-th moment of the monitoring cycle. This represents the average value of the local cyclic multidimensional coefficients within the historical preset time period corresponding to the i-th moment of the monitoring cycle. The recovery execution dimension coefficient is the value at time i of the monitoring period. The recovery execution multidimensional coefficients are the historical preset time periods corresponding to the i-th time point in the monitoring period. This refers to the number of times the third dimension rebounds within the historical preset time period corresponding to the i-th moment of the monitoring period. The standard deviation of the recovery execution multidimensional coefficient within the historical preset time period corresponding to the i-th moment of the monitoring period. It represents the average value of the recovery execution multidimensional coefficients within the historical preset time period corresponding to the i-th moment of the monitoring cycle.

[0015] According to a second aspect of the present invention, a smart TCM rehabilitation monitoring system based on multimodal data fusion is provided, comprising: The real-time data module is used to acquire user detection data at multiple points in the monitoring cycle; The historical data module is used to obtain rehabilitation program information and historical user test data; The rehabilitation coefficient module is used to determine the TCM rehabilitation status coefficient based on the historical user detection data, the user detection data, and the rehabilitation plan information. The risk prediction module is used to determine the comprehensive risk prediction coefficient based on the TCM rehabilitation status coefficient. The rehabilitation monitoring module is used to monitor the rehabilitation status based on the TCM rehabilitation status coefficient and the comprehensive prediction risk coefficient.

[0016] Technical Effects: According to the present invention, user testing data can be acquired in real time through various testing devices. Based on historical user testing data, user testing data, and rehabilitation plan information, the patient's recovery status can be assessed to determine the TCM rehabilitation status coefficient. Furthermore, risk prediction can be performed based on the TCM rehabilitation status coefficient to determine a comprehensive predicted risk coefficient. Rehabilitation monitoring can then be conducted based on the TCM rehabilitation status coefficient and the comprehensive predicted risk coefficient, improving the effectiveness, comprehensiveness, and logical consistency of TCM rehabilitation monitoring. When determining the confidence levels of the first, second, and third dimensions, these levels can be determined based on target activity levels, historical average heart rate variability low-frequency power, physiological regulation data, local circulation data, and recovery execution data. During the calculation process, the confidence levels of the three dimensions of data—physiological regulation data, local circulation data, and recovery execution data—can be accurately analyzed, improving the accuracy and comprehensiveness of the first, second, and third dimension confidence levels. When determining the TCM rehabilitation status coefficient, it can be based on the first dynamic weight, second dynamic weight, third dynamic weight, target activity level, historical average heart rate variability low-frequency power, physiological regulation data, local circulation data, and recovery execution data. During the calculation process, accurate analysis of physiological regulation, blood circulation, and recovery execution status can be performed, improving the comprehensiveness and accuracy of the TCM rehabilitation status coefficient. When determining the comprehensive predictive risk coefficient, it can be based on the first dynamic weight, second dynamic weight, third dynamic weight, physiological regulation sub-dimension coefficient, local circulation sub-dimension coefficient, recovery execution sub-dimension coefficient, standard deviation of physiological regulation sub-dimension coefficient, mean of physiological regulation sub-dimension coefficient, standard deviation of local circulation sub-dimension coefficient, mean of local circulation sub-dimension coefficient, standard deviation of recovery execution sub-dimension coefficient, mean of recovery execution sub-dimension coefficient, total number of samples, number of first dimension rises, number of second dimension rises, and number of third dimension rises. During the calculation process, future risks can be predicted based on the rise and stability of each sub-dimension coefficient, improving the accuracy of the comprehensive predictive risk coefficient.

[0017] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Other features and aspects of the invention will become clearer from the following detailed description of exemplary embodiments with reference to the accompanying drawings. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other embodiments can be obtained based on these drawings without creative effort. Figure 1 An exemplary flowchart of a smart TCM rehabilitation monitoring method based on multimodal data fusion according to an embodiment of the present invention is shown. Figure 2 An exemplary schematic diagram illustrating the determination of the TCM rehabilitation status coefficient according to an embodiment of the present invention is shown; Figure 3 An exemplary schematic diagram illustrating the determination of the comprehensive prediction risk coefficient according to an embodiment of the present invention is shown; Figure 4 A block diagram of a smart TCM rehabilitation monitoring system based on multimodal data fusion according to an embodiment of the present invention is shown as an example. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments.

[0021] Figure 1 An exemplary flowchart illustrates a smart TCM rehabilitation monitoring method based on multimodal data fusion according to an embodiment of the present invention, the method comprising: Step S1: Acquire user detection data at multiple points in the monitoring cycle; Step S2: Obtain rehabilitation plan information and historical user test data; Step S3: Determine the TCM rehabilitation status coefficient based on the historical user detection data, the user detection data, and the rehabilitation plan information; Step S4: Determine the comprehensive prediction risk coefficient based on the TCM rehabilitation status coefficient; Step S5: Monitor the situation based on the TCM rehabilitation status coefficient and the comprehensive prediction risk coefficient.

[0022] The intelligent TCM rehabilitation monitoring method based on multimodal data fusion according to embodiments of the present invention can acquire user detection data in real time through various detection devices, and evaluate the patient's recovery status based on historical user detection data, user detection data and rehabilitation plan information to determine the TCM rehabilitation status coefficient. Furthermore, risk prediction can be performed based on the TCM rehabilitation status coefficient to determine the comprehensive prediction risk coefficient, and rehabilitation monitoring can be performed based on the TCM rehabilitation status coefficient and the comprehensive prediction risk coefficient, thereby improving the effectiveness, comprehensiveness and logic of TCM rehabilitation monitoring.

[0023] According to one embodiment of the present invention, in step S1, user detection data is acquired at multiple moments during the monitoring period.

[0024] For example, the monitoring period is 30 minutes apart. During the monitoring period, user test data is acquired through devices such as PPG / ECG wristbands, surface electromyography patches, infrared thermal imagers, inertial measurement units, and human-computer interaction interfaces (for entering questionnaires and subjective feelings), and the real-time collected user test data is stored in the database.

[0025] According to one embodiment of the present invention, in step S2, rehabilitation program information and historical user detection data are obtained.

[0026] For example, information about rehabilitation plans (such as target activity levels and rehabilitation stages) can be obtained through pre-defined rehabilitation programs, and historical user test data can be obtained through databases.

[0027] According to an embodiment of the present invention, in step S3, a TCM rehabilitation status coefficient is determined based on the historical user detection data, the user detection data, and the rehabilitation program information.

[0028] Figure 2 An exemplary schematic diagram illustrating the determination of the TCM rehabilitation status coefficient according to an embodiment of the present invention is shown.

[0029] According to an embodiment of the present invention, step S3 includes: Step S31: Based on the user detection data, determine physiological regulation data, local circulation data, and recovery execution data. The physiological regulation data includes: low-frequency power of heart rate variability, low-frequency power of ideal heart rate variability, and short-term coefficient of variation of RR interval. The local circulation data includes: average skin temperature at the rehabilitation position, average skin temperature at the corresponding point on the healthy side, change in average skin temperature at the rehabilitation position, and change in average skin temperature at the corresponding point on the healthy side. The recovery execution data includes: actual completed activity amount, visual analog value of pain, and joint range of motion achievement rate. Step S32: Determine the historical average heart rate variability low-frequency power based on the historical user detection data; Step S33: Determine the target activity level and rehabilitation stage based on the rehabilitation plan information; Step S34: Determine the first dynamic weight, the second dynamic weight, and the third dynamic weight based on the target activity level, the historical average heart rate variability low-frequency power, the rehabilitation stage, the physiological regulation data, the local circulation data, and the recovery execution data. Step S35: Determine the TCM rehabilitation state coefficient based on the first dynamic weight, the second dynamic weight, the third dynamic weight, the target activity level, the historical average heart rate variability low-frequency power, the physiological regulation data, the local circulation data, and the recovery execution data.

[0030] For example, based on user test data, physiological regulation data (representing the patient's life energy and physiological regulatory functions), local circulation data (representing the patient's material nutrition supply and local circulation status), and recovery execution data (representing the patient's mental will, active participation, and functional execution ability) are determined. Among them, physiological regulation data includes: heart rate variability low-frequency power (reflecting the synergistic effect of sympathetic nerves and humoral regulation (e.g., the renin-angiotensin system). In traditional Chinese medicine theory, this is highly related to the functions of "qi" in promoting, warming, defending, and consolidating. It can be obtained by collecting electrocardiogram signals through PPG / ECG sensors, calculating the RR interval sequence, and then performing spectral analysis (e.g., Lomb-Scargle) to obtain 0.04-0.The system determines the 15Hz frequency band power, ideal heart rate variability low-frequency power (the system queries a preset mapping table based on the results of a standard TCM constitution questionnaire filled out by the user during initialization, such as Qi deficiency constitution or blood stasis constitution, to obtain a personalized ideal heart rate variability low-frequency power), and RR interval short-term variability coefficient (based on the RR interval sequence obtained from the same PPG / ECG sensor, the ratio of its standard deviation to the mean within a short time window (such as 30 seconds). Local circulation data includes: average skin temperature at the rehabilitation site (measured and calculated using an infrared thermal imager or a high-precision thermal sensor, aimed at the rehabilitation site (specific area), and the corresponding temperature on the healthy side). The data included: average skin temperature at specific points (measured and calculated using the same equipment at symmetrical anatomical locations on the healthy side of the body); average skin temperature change at the rehabilitation location (the change in average skin temperature at the current rehabilitation location compared to 30 minutes prior; if a temperature decrease occurred, the change was negative); and average skin temperature change at corresponding points on the healthy side (the change in average skin temperature at the current corresponding points on the healthy side compared to 30 minutes prior). Recovery execution data included: actual activity level (monitored in real-time by motion sensors such as inertial measurement units, and the equivalent activity level calculated in kilocalories); and pain visual model. The system includes: a simulated pain level (reported by the patient on a scale of 0-10 via an app or human-computer interaction device) and a joint range of motion target achievement rate (calculated via IMU, with values ​​between 0 and 1); based on historical user data, the historical average heart rate variability low-frequency power (HRVLP) is determined. HRVLP represents an individual baseline; it is measured at rest during the initial rehabilitation phase or a stable period, and the average of multiple measurements is taken as the historical average HRVLP; based on rehabilitation program information, the target activity level is determined (the target activity level set by the rehabilitation physician or system algorithm, e.g., ...). The rehabilitation process is divided into three phases: energy expenditure and recovery phase (e.g., acute phase, recovery phase, and consolidation phase). Based on target activity levels, historical average heart rate variability low-frequency power, rehabilitation phase, physiological regulation data, local circulation data, and recovery execution data, a first dynamic weight is determined for physiological regulation data, a second dynamic weight for local circulation data, and a third dynamic weight for recovery execution data. Based on these dynamic weights, the patient's recovery status is assessed to determine the TCM rehabilitation status coefficient.

[0031] According to an embodiment of the present invention, step S34 includes: Step S341: Determine the function values ​​for the first rehabilitation stage, the second rehabilitation stage, and the third rehabilitation stage based on the rehabilitation stages. Step S342: Determine the first dimension confidence level, the second dimension confidence level, and the third dimension confidence level based on the target activity level, the historical average heart rate variability low-frequency power, the physiological regulation data, the local circulation data, and the recovery execution data. Step S343: Determine the first dynamic weight, the second dynamic weight, and the third dynamic weight based on the first rehabilitation stage function value, the second rehabilitation stage function value, the third rehabilitation stage function value, the first dimension confidence level, the second dimension confidence level, and the third dimension confidence level.

[0032] For example, the function values ​​for the first rehabilitation stage are designed based on the "physiological regulation data" aspect within the three rehabilitation stages; the function values ​​for the second rehabilitation stage are designed based on the "local circulation data" aspect; and the function values ​​for the third rehabilitation stage are designed based on the "recovery execution data" aspect. When the rehabilitation stages are the acute phase, recovery phase, and consolidation phase, the function values ​​for the first rehabilitation stage are 0.3, 0, and -0.1, respectively; the function values ​​for the second rehabilitation stage are 0.2, 0.1, and 0, respectively; and the function values ​​for the third rehabilitation stage are -0.2, 0.1, and 0.3. The function values ​​for the first, second, and third rehabilitation stages are... The specific logic for setting the function value of the rehabilitation stage is as follows: (1) In the acute phase, the focus is on "qi" and "blood". During this phase, the vital energy is damaged, and the stability of vital signs (physiological regulation data) is crucial. It is necessary to increase its assessment weight to "strengthen the body and consolidate the foundation". Therefore, the function value of the first rehabilitation stage of this phase is set to 0.3. During this phase, blood stasis is stagnant, and local circulation (local circulation data) is the direct window for observing the condition. It is necessary to pay appropriate attention to "activate blood and remove blood stasis". Therefore, the function value of the second rehabilitation stage of this phase is set to 0.2. During this phase, patients will inevitably have low active activity and compliance (recovery execution data) due to pain and immobilization. Over-reliance on this dimension for assessment will lead to distortion. Therefore, its weight is reduced. Therefore, the function value of the third rehabilitation stage of this stage is set to -0.2; (2) Blood regulation and mental nourishment during the recovery period. The physiological functions tend to be stable during this stage. The weight of the "physiological regulation data" regresses the confidence of the data itself. Therefore, the function value of the first rehabilitation stage of this stage is set to 0. The focus of this stage is to promote blood circulation and eliminate swelling. The emphasis on the "local circulation data" dimension is maintained. Therefore, the function value of the second rehabilitation stage of this stage is set to 0.1. Active rehabilitation is encouraged at this stage. The patient's participation and willpower (recovery execution data) become the key driving force for rehabilitation progress. Their weight needs to be increased. Therefore, the function value of the third rehabilitation stage of this stage is set to 0.1. The function value of the recovery stage is set to 0.1; (3) Strengthen the "spirit" and consolidate the "form" in the consolidation period. The physiological indicators (physiological regulation data) in this stage may have returned to normal or even become better due to training. Their discrimination is reduced, so the weight is slightly reduced. Therefore, the function value of the first rehabilitation stage in this stage is set to -0.1. The local circulation in this stage has usually returned to normal. The weight is adjusted to the confidence of the regression data. Therefore, the function value of the second rehabilitation stage in this stage is set to 0. The rehabilitation effect in this stage is highly dependent on the patient's persistence and the quality of the movement (recovery execution data). Its weight should be greatly increased to "nourish both body and spirit". Therefore, the function value of the third rehabilitation stage in this stage is set to 0.3.Based on the target activity level, historical average heart rate variability low-frequency power, physiological regulation data, local circulation data, and recovery execution data, assess the confidence level of the data in each dimension to determine the confidence levels of the first, second, and third dimensions. Based on the function values ​​of the first, second, and third rehabilitation stages, the confidence levels of the first, second, and third dimensions, determine the first, second, and third dynamic weights, as shown below. The first dynamic weight is determined in this way, and the second and third dynamic weights can be determined in a similar manner.

[0033] According to an embodiment of the present invention, step S342 includes: determining the confidence level of the first dimension at the i-th moment of the monitoring period according to formula (1). Second dimension confidence and the third dimension confidence , (1) Where min is the function for finding the minimum value. and For preset coefficients, Let be the short-term variation coefficient of the RR interval at time i in the monitoring period. The low-frequency power of heart rate variability at time i in the monitoring cycle. For a preset time period, The low-frequency power of heart rate variability at a preset time point before the i-th moment of the monitoring cycle. For ideal heart rate variability, low-frequency power, The historical average heart rate variability low-frequency power, This represents the average skin temperature change at the rehabilitation site at time i during the monitoring period. This represents the average skin temperature change at the corresponding point on the healthy side at time i of the monitoring cycle. The average skin temperature at the rehabilitation site at time i of the monitoring period. The average skin temperature at the corresponding point on the healthy side at the i-th moment of the monitoring cycle. For the target activity level, This represents the actual amount of activity completed at time i during the monitoring period. The visual simulation value of pain at the i-th moment of the monitoring period.

[0034] According to one embodiment of the present invention, The short-term variation coefficient of the RR interval at time i in the monitoring period. The larger the value, the more volatile the heart rate variability (large standard deviation), indicating an unstable signal that may be affected by motion artifacts, short-term emotional fluctuations, or measurement interference. An unstable raw signal naturally leads to a lower reliability of the derived indicators. The absolute value of the difference between the low-frequency heart rate variability power at time i of the monitoring period and the low-frequency heart rate variability power at a preset time point before time i of the monitoring period represents the instantaneous fluctuation amplitude of the low-frequency heart rate variability power. The larger the instantaneous fluctuation amplitude, the greater the noise and instability of the low-frequency heart rate variability power signal itself. The lower the confidence level of the first dimension, the smaller the error in judging the long-term state due to transient interference. It can be set to 10 seconds. This indicates that the larger the short-term coefficient of variation of the RR interval, The smaller the value, the lower the confidence level of the first dimension of the physiological regulation data when the signal is noisy. This is the ratio of the difference between the low-frequency heart rate variability power and the ideal low-frequency heart rate variability power at time i of the monitoring period to the historical average low-frequency heart rate variability power. It represents the relative degree to which the current low-frequency heart rate variability power deviates from the ideal low-frequency heart rate variability power. The lower the relative deviation between the low-frequency heart rate variability power and the ideal low-frequency heart rate variability power, the better. The closer the value is to 1, the greater the relative deviation between the low-frequency power of heart rate variability and the ideal low-frequency power of heart rate variability, the more likely the reading is abnormal (e.g., due to interference from stress, caffeine, etc.). The smaller the value, the lower the confidence level of the first dimension. Used to adjust the impact of signal stability on confidence level. It can be set to 2.

[0035] According to one embodiment of the present invention, This is the ratio of the average skin temperature change at the rehabilitation site to the average skin temperature change at the corresponding point on the healthy side at the i-th moment of the monitoring period. In the actual system process, a very small constant is added to the denominator to prevent the denominator from being zero. This represents the synchronicity of the temperature change rate between the rehabilitation site and the healthy side. Changes in ambient temperature (e.g., air conditioning on, blankets covering) will affect the skin temperature of both the affected and healthy sides simultaneously. If the temperature change trends and magnitudes of the two are highly synchronized ( If the reading is close to 1, then it can be assumed that the current temperature reading mainly reflects environmental changes, and the "blood supply" information of the affected area itself may be masked. However, at least the measurement is consistent and reliable. If the two changes are not synchronized ( The larger the difference between 1 and 1, the more likely it is that the measurement point is unstable, the sensor is not making good contact, or there are unique pathological changes in the affected area. The smaller the value, the lower the confidence level of the second dimension. This refers to the absolute temperature difference between the average skin temperature at the rehabilitation site and the corresponding average skin temperature on the healthy side at time i of the monitoring cycle. In a stable environment, a certain temperature difference is a normal reflection of pathology (e.g., the temperature at the affected area is higher during the inflammatory phase). However, an excessively large temperature difference may indicate that the local condition is extremely unstable (e.g., severe ischemia or severe infection). In this case, skin temperature as a single indicator may no longer be able to stably reflect subtle changes in "blood supply," or the sensor may be operating outside its normal range. The lower the confidence level of the second dimension, the more likely it is to be affected. The confidence level for the second dimension is determined based on two factors: the synchronicity of the rate of temperature change between the rehabilitation site and the healthy side, and the absolute temperature difference between the average skin temperature at the rehabilitation site and the corresponding average skin temperature at the healthy side. Used to adjust the effect of temperature difference on confidence level It can be set to 0.2 (°C) - ¹).

[0036] According to one embodiment of the present invention, This represents the ratio of the actual amount of activity completed to the target amount of activity at time i in the monitoring period. Indicates taking The minimum value of 1 and the above minimum value processing can be used to find the minimum value. The upper limit is 1. If the patient completes the plan completely or even exceeds it, it indicates a high level of subjective effort and cooperation, and the data is reliable. If the completion rate is very low, it may mean that the patient is wearing the device incorrectly. In this case, the actual amount of activity completed cannot effectively reflect the patient's accurate condition. The lower the confidence level of the third dimension, the more reliable the data becomes. The ratio of the pain visual analog value at time i of the monitoring period to 10 is used to normalize the pain visual analog value. The larger the pain visual analog value, the lower the confidence of the third dimension, indicating that severe pain will seriously affect the patient's attention. When the patient is in high pain, he / she cannot concentrate and cannot focus on the key points of the movement. This means that under the condition of high pain, his / her behavioral data can no longer objectively and stably reflect his / her original movement status and rehabilitation ability, resulting in a decrease in the reliability of objective behavioral data.

[0037] In this way, the confidence levels of the first, second, and third dimensions can be determined based on the target activity level, historical average heart rate variability low-frequency power, physiological regulation data, local circulation data, and recovery execution data. During the calculation process, the confidence levels of the three dimensions of physiological regulation data, local circulation data, and recovery execution data can be accurately analyzed, improving the accuracy and comprehensiveness of the confidence levels of the first, second, and third dimensions.

[0038] According to an embodiment of the present invention, step S35 includes: determining the TCM rehabilitation status coefficient at the i-th moment of the monitoring period according to formula (2). , (2) in, , and Here, is the preset coefficient, and max is the function to find the maximum value. The first dynamic weight at the i-th moment of the monitoring period, The second dynamic weight is the weight at the i-th time point of the monitoring period. The third dynamic weight is the value at the i-th time point of the monitoring period. The low-frequency power of heart rate variability at time i in the monitoring cycle. For ideal heart rate variability, low-frequency power, The historical average heart rate variability low-frequency power, The short-term variation coefficient of the RR interval at time i in the monitoring period. The average skin temperature at the rehabilitation site at time i of the monitoring period. The average skin temperature at the corresponding point on the healthy side at the i-th moment of the monitoring cycle. This represents the average skin temperature change at the rehabilitation site at time i during the monitoring period. This represents the average skin temperature change at the corresponding point on the healthy side. For the target activity level, This represents the actual amount of activity completed at time i during the monitoring period. The visual analog value of pain at time i in the monitoring cycle. This represents the joint range of motion compliance rate at the i-th moment of the monitoring cycle.

[0039] According to one embodiment of the present invention, This indicator, representing the low-frequency power of heart rate variability at time i of the monitoring cycle, comprehensively reflects sympathetic tone and humoral regulation (such as the renin-angiotensin system). In the context of Traditional Chinese Medicine, it is associated with the driving and regulating function of "qi." This represents the absolute value of the difference between the low-frequency power of heart rate variability and the ideal low-frequency power of heart rate variability at time i of the monitoring cycle. The larger this absolute value, the further the regulatory function of "qi" deviates from its optimal equilibrium state. The historical average heart rate variability low-frequency power is used as an individualized calibration benchmark. The short-term variation coefficient of the RR interval at time i in the monitoring period. The larger the number, the greater the degree of heart rhythm disorder. In traditional Chinese medicine theory, this is a sign of "disordered Qi". The larger the value, the better. The smaller the value, This indicates a disordered state in the flow of "qi". The smaller the size, the more severe the disorder in the flow of "qi". This indicates that the physiological regulatory status is determined based on the regulatory function and disordered operation of Qi. It can be set to 2.

[0040] According to one embodiment of the present invention, This is the ratio of the average skin temperature at the rehabilitation site and the average skin temperature at the corresponding point on the healthy side at time i of the monitoring cycle. This ratio reflects the relative blood supply level between the affected area and healthy tissue. The more adequate the blood supply, the more metabolic heat is generated, and the higher the skin temperature is usually. The larger the value, the better. middle, It can be set to 0.7. Judging blood circulation solely based on temperature at a single moment may be inaccurate, therefore, by... Setting it to 0.5 weakens the influence of the average skin temperature at the rehabilitation site on the TCM rehabilitation status coefficient. The difference between the average skin temperature change at the rehabilitation site and the average skin temperature change at the corresponding point on the healthy side at the i-th moment of the monitoring period represents the cooling rate of the affected area minus the cooling rate of the healthy area. This represents the absolute value of the average skin temperature change at the corresponding point on the healthy side, and is the absolute value of the temperature change at the healthy site. It serves as the normalization benchmark. During the calculation process, the system usually adds a very small constant to the denominator to prevent the denominator from being zero. The larger the ratio, the more pronounced the abnormal cooling trend of the affected area relative to healthy tissue. To take 0 and The maximum value indicates that only the case where the affected area cools down faster than the healthy area is considered. When an abnormally accelerated cooling of the affected area is detected, A value less than 1 lowers the TCM rehabilitation status coefficient, issuing a warning that "blood circulation may be deteriorating." It can be set to 1.5. It indicates the patient's blood supply status.

[0041] According to one embodiment of the present invention, This is the ratio of the actual amount of activity completed to the target amount of activity at time i of the monitoring period. The larger this ratio, the higher the patient's adherence to the pre-set rehabilitation plan, the better the patient's willingness and ability to perform rehabilitation, and the better the recovery progress. This indicates the patient's subjective pain level. The smaller the age, the higher the patient's stability and concentration during the recovery process, and the better their executive function. To monitor the joint range of motion target achievement rate at time i of the monitoring cycle, a higher joint range of motion target achievement rate indicates that the patient's rehabilitation movements are more standardized and the recovery execution is better. This indicates the patient's recovery progress.

[0042] According to one embodiment of the present invention, The TCM rehabilitation status coefficient is determined based on three aspects: physiological regulation, blood circulation, and recovery performance.

[0043] In this way, the TCM rehabilitation status coefficient can be determined based on the first dynamic weight, the second dynamic weight, the third dynamic weight, the target activity level, the historical average heart rate variability low-frequency power, physiological regulation data, local circulation data, and recovery execution data. During the calculation process, the physiological regulation status, blood circulation status, and recovery execution status can be accurately analyzed, thus improving the comprehensiveness and accuracy of the TCM rehabilitation status coefficient.

[0044] According to one embodiment of the present invention, in step S4, a comprehensive prediction risk coefficient is determined based on the TCM rehabilitation status coefficient.

[0045] Figure 3 An exemplary schematic diagram illustrating the determination of the comprehensive prediction risk coefficient according to an embodiment of the present invention is shown.

[0046] According to an embodiment of the present invention, step S4 includes: Step S41: Based on the TCM rehabilitation state coefficient, determine the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, and the recovery execution sub-dimension coefficient; Step S42: Based on the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, and the recovery execution sub-dimension coefficient, determine the standard deviation of the physiological regulation sub-dimension coefficient, the average value of the physiological regulation sub-dimension coefficient, the standard deviation of the local circulation sub-dimension coefficient, the average value of the local circulation sub-dimension coefficient, the standard deviation of the recovery execution sub-dimension coefficient, and the average value of the recovery execution sub-dimension coefficient within a historical preset time period. Step S43: Determine the total number of samplings, the number of first dimension rises, the number of second dimension rises, and the number of third dimension rises for the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, and the recovery execution sub-dimension coefficient within the historical preset time period. Step S44: Determine the comprehensive prediction risk coefficient based on the first dynamic weight, the second dynamic weight, the third dynamic weight, the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, the recovery execution sub-dimension coefficient, the standard deviation of the physiological regulation sub-dimension coefficient, the average value of the physiological regulation sub-dimension coefficient, the standard deviation of the local circulation sub-dimension coefficient, the average value of the local circulation sub-dimension coefficient, the standard deviation of the recovery execution sub-dimension coefficient, the average value of the recovery execution sub-dimension coefficient, the total number of samplings, the number of times the first dimension rebounds, the number of times the second dimension rebounds, and the number of times the third dimension rebounds.

[0047] For example, based on the TCM rehabilitation state coefficient, the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, and the recovery execution sub-dimension coefficient are determined. In formula (2), For physiological regulation dimension coefficients, For local cyclic dimension coefficients, To recover the performance of the sub-dimension coefficients; when the recovery phase is acute, the historical preset time period can be set to 6-12 hours (for high-frequency monitoring); when the recovery phase is recovery period, the historical preset time period can be set to 24 hours (for daily trend monitoring); when the recovery phase is consolidation period, the historical preset time period can be set to 3 days (for long-term adverse monitoring). Obtain the standard deviation, mean, average, and average values ​​of the physiological regulation sub-dimension coefficients, local circulation sub-dimension coefficients, local circulation sub-dimension coefficients, recovery performance sub-dimension coefficients, and recovery performance sub-dimension coefficients within the historical preset time period prior to the current moment; determine the total number of samplings, the number of first-dimension rebounds, and the number of second-dimension rebounds within the historical preset time period. The number of increases and the number of rebounds in the third dimension are calculated. For example, the number of rebounds in the first dimension is the number of times the physiological regulation sub-dimension coefficient has rebounded compared to the previous moment. Similarly, the number of rebounds in the second dimension and the number of rebounds in the third dimension are determined. The total number of samples is the number of moments within the historical preset time period. Based on the first dynamic weight, the second dynamic weight, the third dynamic weight, the physiological regulation sub-dimension coefficient, the local loop sub-dimension coefficient, the recovery execution sub-dimension coefficient, the standard deviation of the physiological regulation sub-dimension coefficient, the average of the physiological regulation sub-dimension coefficient, the standard deviation of the local loop sub-dimension coefficient, the average of the local loop sub-dimension coefficient, the standard deviation of the recovery execution sub-dimension coefficient, the average of the recovery execution sub-dimension coefficient, the total number of samples, the number of rebounds in the first dimension, the number of rebounds in the second dimension, and the number of rebounds in the third dimension, the future risk situation is predicted, and the comprehensive predicted risk coefficient is determined.

[0048] According to an embodiment of the present invention, step S44 includes: determining the comprehensive prediction risk coefficient at the i-th moment of the monitoring period according to formula (3). , (3) in, As the first preset weight, The second preset weight is K, which is a preset constant. The first dynamic weight at the i-th moment of the monitoring period, The second dynamic weight is the weight at the i-th time point of the monitoring period. The third dynamic weight is the value at the i-th time point of the monitoring period. To find the maximum value function, The physiological regulation subdimension coefficients at the i-th time point of the monitoring cycle. The length of the historical time period is preset. The physiological regulation subdimension coefficients are the values ​​corresponding to the historical preset time period before the i-th moment of the monitoring cycle. This refers to the number of times the first dimension rebounds within the historical preset time period corresponding to the i-th moment of the monitoring period. This represents the total number of samples taken within the historical preset time period corresponding to the i-th moment of the monitoring period. The standard deviation of the physiological regulation subdimension coefficient within the historical preset time period corresponding to the i-th moment of the monitoring cycle. This represents the average value of the physiological regulation subdimension coefficients within the historical preset time period corresponding to the i-th moment of the monitoring cycle. For the local cyclic subdimension coefficients at the i-th time point of the monitoring period, The local cyclic subdimension coefficients are the values ​​corresponding to the historical preset time periods prior to the i-th moment of the monitoring period. This refers to the number of times the second dimension rebounds within the historical preset time period corresponding to the i-th moment of the monitoring period. The standard deviation of the local cyclic subdimension coefficient within the historical preset time period corresponding to the i-th moment of the monitoring cycle. This represents the average value of the local cyclic multidimensional coefficients within the historical preset time period corresponding to the i-th moment of the monitoring cycle. The recovery execution dimension coefficient is the value at time i of the monitoring period. The recovery execution multidimensional coefficients are the historical preset time periods corresponding to the i-th time point in the monitoring period. This refers to the number of times the third dimension rebounds within the historical preset time period corresponding to the i-th moment of the monitoring period. The standard deviation of the recovery execution multidimensional coefficient within the historical preset time period corresponding to the i-th moment of the monitoring period. It represents the average value of the recovery execution multidimensional coefficients within the historical preset time period corresponding to the i-th moment of the monitoring cycle.

[0049] According to one embodiment of the present invention, The relative difference between the physiological regulation subdimension coefficient at a time corresponding to a preset historical time period prior to the i-th time of the monitoring period and the physiological regulation subdimension coefficient at the i-th time of the monitoring period indicates the degree of change of the physiological regulation subdimension coefficient within the preset historical time period. Indicates taking 0 and The maximum value, as described above, can be used to determine the degree of decrease in the physiological regulation subdimension coefficient over a preset historical time period. When the physiological regulation subdimension coefficient does not decrease... The value is 0 when the physiological regulation subdimension coefficient decreases. The value is , The ratio of the number of times the first dimension rebounds within the historical preset time period corresponding to the i-th moment of the monitoring period to the total number of samples within the historical preset time period corresponding to the i-th moment of the monitoring period is larger. The larger the ratio, the more times the physiological regulation subdimension coefficient rebounds compared to the previous time point within the historical preset time period. Even if the overall value decreases, frequent rebounds in the middle (fluctuation recovery) mean that the risk of continued decline is relatively small. The smaller the ratio, the more drastic the decline trend of the physiological regulation subdimension coefficient within the historical preset time period (no obvious rebound) is, and the greater the risk of continued decline. The predictive risk of a sustained decline in the physiological regulation sub-dimension coefficient is represented by the same factor. and These represent the predicted risk of continued decline in the partial loop subdimension coefficient and the recovery execution subdimension coefficient, respectively.

[0050] According to one embodiment of the present invention, The ratio of the standard deviation to the mean of the physiological regulation subdimension coefficients within the historical preset time period corresponding to the i-th moment of the monitoring cycle represents the discrimination coefficient of the physiological regulation subdimension coefficients within the historical preset time period. The larger the value, the greater the fluctuation of the physiological regulation sub-dimension coefficient within the preset time period, the higher the degree of instability, and the higher the risk of instability in the predicted state of the physiological regulation sub-dimension. Similarly, and These represent the predicted state instability risk of the local loop subdimension coefficient and the recovery execution subdimension coefficient, respectively.

[0051] According to one embodiment of the present invention, This indicates that the comprehensive predicted risk of the physiological regulation dimension is determined based on the predicted risk of continued decline and the predicted risk of state instability, whereby... and These can be set to 0.7 and 0.3 respectively, and similarly, and These represent the combined predicted risks of the local loop dimension and the recovery execution dimension, respectively.

[0052] According to one embodiment of the present invention, , and These are the corresponding weights for the physiological regulation dimension, the local circulation dimension, and the recovery execution dimension. These corresponding weights are inversely related to the first dynamic weight, the second dynamic weight, and the third dynamic weight, respectively. The specific design logic is as follows: The health status of the human body is the result of the dynamic balance between various systems. When one aspect (such as the physiological regulation dimension) is too prominent or becomes the main contradiction, it often means that other aspects (such as the local circulation dimension or the recovery execution dimension) may have relative deficiencies or hidden dangers. These are the weak links where pathogens may take advantage of the weakness. When setting the comprehensive prediction risk coefficient, it is necessary to set higher corresponding weights for the dimensions corresponding to the weak links.

[0053] In this way, the comprehensive prediction risk coefficient can be determined based on the first dynamic weight, the second dynamic weight, the third dynamic weight, the physiological regulation sub-dimension coefficient, the local loop sub-dimension coefficient, the recovery execution sub-dimension coefficient, the standard deviation of the physiological regulation sub-dimension coefficient, the average of the physiological regulation sub-dimension coefficient, the standard deviation of the local loop sub-dimension coefficient, the average of the local loop sub-dimension coefficient, the standard deviation of the recovery execution sub-dimension coefficient, the average of the recovery execution sub-dimension coefficient, the total number of samples, the number of times the first dimension rebounds, the number of times the second dimension rebounds, and the number of times the third dimension rebounds. During the calculation process, future risks can be predicted based on the rebound and stability of each sub-dimension coefficient, thus improving the accuracy of the comprehensive prediction risk coefficient.

[0054] According to one embodiment of the present invention, in step S5, monitoring is performed based on the TCM rehabilitation status coefficient and the comprehensive prediction risk coefficient.

[0055] For example, by determining the predicted average and standard deviation of the comprehensive predicted risk coefficient at multiple points in the monitoring period, when the TCM rehabilitation status coefficient is less than 0.65 and the comprehensive predicted risk coefficient is greater than the predicted average plus twice the predicted standard deviation, it indicates that the rehabilitation status is already very poor and the risk is rising rapidly. This is the most critical situation, requiring the activation of the highest level of response. When the TCM rehabilitation status coefficient is less than 0.8 and the slope of the straight line determined by fitting the comprehensive predicted risk coefficient and the points in the monitoring period remains positive, it indicates that the status is declining and risk factors are accumulating. It is necessary to find the root cause and adjust the rehabilitation plan. When the TCM rehabilitation status coefficient is greater than or equal to 0.8, but the comprehensive predicted risk coefficient is greater than the predicted average plus twice the predicted standard deviation, it indicates that the current status looks good, but there are already signs of risk. This is the golden window for early intervention.

[0056] The intelligent TCM rehabilitation monitoring method based on multimodal data fusion according to embodiments of the present invention can acquire user detection data in real time through various detection devices, and assess the patient's recovery status based on historical user detection data, user detection data, and rehabilitation plan information to determine the TCM rehabilitation status coefficient. Furthermore, risk prediction can be performed based on the TCM rehabilitation status coefficient to determine a comprehensive predicted risk coefficient, and rehabilitation monitoring can be conducted based on the TCM rehabilitation status coefficient and the comprehensive predicted risk coefficient, thus improving the effectiveness, comprehensiveness, and logicality of TCM rehabilitation monitoring. When determining the confidence levels of the first, second, and third dimensions, these levels can be determined based on target activity levels, historical average heart rate variability low-frequency power, physiological regulation data, local circulation data, and recovery execution data. During the calculation process, the confidence levels of the three dimensions of data—physiological regulation data, local circulation data, and recovery execution data—can be accurately analyzed, improving the accuracy and comprehensiveness of the first, second, and third dimension confidence levels. When determining the TCM rehabilitation status coefficient, it can be based on the first dynamic weight, second dynamic weight, third dynamic weight, target activity level, historical average heart rate variability low-frequency power, physiological regulation data, local circulation data, and recovery execution data. During the calculation process, accurate analysis of physiological regulation, blood circulation, and recovery execution status can be performed, improving the comprehensiveness and accuracy of the TCM rehabilitation status coefficient. When determining the comprehensive predictive risk coefficient, it can be based on the first dynamic weight, second dynamic weight, third dynamic weight, physiological regulation sub-dimension coefficient, local circulation sub-dimension coefficient, recovery execution sub-dimension coefficient, standard deviation of physiological regulation sub-dimension coefficient, mean of physiological regulation sub-dimension coefficient, standard deviation of local circulation sub-dimension coefficient, mean of local circulation sub-dimension coefficient, standard deviation of recovery execution sub-dimension coefficient, mean of recovery execution sub-dimension coefficient, total number of samples, number of first dimension rises, number of second dimension rises, and number of third dimension rises. During the calculation process, future risks can be predicted based on the rise and stability of each sub-dimension coefficient, improving the accuracy of the comprehensive predictive risk coefficient.

[0057] Figure 4 An exemplary block diagram of a smart TCM rehabilitation monitoring system based on multimodal data fusion according to an embodiment of the present invention is shown, the system comprising: The real-time data module is used to acquire user detection data at multiple points in the monitoring cycle; The historical data module is used to obtain rehabilitation program information and historical user test data; The rehabilitation coefficient module is used to determine the TCM rehabilitation status coefficient based on the historical user detection data, the user detection data, and the rehabilitation plan information. The risk prediction module is used to determine the comprehensive risk prediction coefficient based on the TCM rehabilitation status coefficient. The rehabilitation monitoring module is used to monitor the rehabilitation status based on the TCM rehabilitation status coefficient and the comprehensive prediction risk coefficient.

[0058] This invention can be a method, apparatus, system, and / or computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions loaded thereon for performing various aspects of the invention.

[0059] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are merely examples and do not limit the present invention. The objectives of the present invention have been fully and effectively achieved. The functions and structural principles of the present invention have been demonstrated and explained in the embodiments, and any variations or modifications may be made to the implementation of the present invention without departing from the stated principles.

Claims

1. A smart traditional Chinese medicine rehabilitation monitoring method based on multi-modal data fusion, characterized in that, The method comprises the following steps: acquiring user detection data at multiple time points in a monitoring period; acquiring rehabilitation program information and historical user detection data; determining a traditional Chinese medicine rehabilitation state coefficient according to the historical user detection data, the user detection data, and the rehabilitation program information; determining a comprehensive prediction risk coefficient according to the traditional Chinese medicine rehabilitation state coefficient; monitoring according to the traditional Chinese medicine rehabilitation state coefficient and the comprehensive prediction risk coefficient; determining a traditional Chinese medicine rehabilitation state coefficient according to the historical user detection data, the user detection data, and the rehabilitation program information, comprising: determining physiological adjustment data, local circulation data, and recovery execution data according to the user detection data, wherein the physiological adjustment data comprises heart rate variability low-frequency power, ideal heart rate variability low-frequency power, and RR interval short-term variation coefficient, the local circulation data comprises rehabilitation position average skin temperature, healthy side corresponding point average skin temperature, rehabilitation position average skin temperature change amount, and healthy side corresponding point average skin temperature change amount, and the recovery execution data comprises actual completed activity amount, pain visual analog value, and joint range of motion compliance rate; determining historical average heart rate variability low-frequency power according to the historical user detection data; determining target activity amount and rehabilitation stage according to the rehabilitation program information; determining first dynamic weight, second dynamic weight, and third dynamic weight according to the target activity amount, the historical average heart rate variability low-frequency power, the rehabilitation stage, the physiological adjustment data, the local circulation data, and the recovery execution data; determining a traditional Chinese medicine rehabilitation state coefficient according to the first dynamic weight, the second dynamic weight, the third dynamic weight, the target activity amount, the historical average heart rate variability low-frequency power, the physiological adjustment data, the local circulation data, and the recovery execution data. 2.The intelligent Chinese medicine rehabilitation monitoring method based on multi-modal data fusion according to claim 1, characterized in that, determining first dynamic weight, second dynamic weight, and third dynamic weight according to the target activity amount, the historical average heart rate variability low-frequency power, the rehabilitation stage, the physiological adjustment data, the local circulation data, and the recovery execution data, comprising: determining first rehabilitation stage function value, second rehabilitation stage function value, and third rehabilitation stage function value according to the rehabilitation stage; determining first dimension confidence, second dimension confidence, and third dimension confidence according to the target activity amount, the historical average heart rate variability low-frequency power, the physiological adjustment data, the local circulation data, and the recovery execution data; determining first dynamic weight, second dynamic weight, and third dynamic weight according to the first rehabilitation stage function value, the second rehabilitation stage function value, the third rehabilitation stage function value, the first dimension confidence, the second dimension confidence, and the third dimension confidence. 3.The intelligent Chinese medicine rehabilitation monitoring method based on multi-modal data fusion according to claim 2, characterized in that, determining first dimension confidence, second dimension confidence, and third dimension confidence according to the target activity amount, the historical average heart rate variability low-frequency power, the physiological adjustment data, the local circulation data, and the recovery execution data, comprising: according to the formula: a first-dimension confidence degree of the i th moment of the monitoring period a second-dimension confidence degree and a third-dimension confidence degree wherein min is a minimum function, and is a preset coefficient, is a short-term coefficient of variation of RR intervals at the i th moment of the monitoring period, is a low-frequency power of heart rate variability at the i th moment of the monitoring period, is a preset time period, is a low-frequency power of heart rate variability at a preset time point before the i th moment of the monitoring period, is an ideal low-frequency power of heart rate variability, is a historical average low-frequency power of heart rate variability, is an average skin temperature change amount of the rehabilitation position at the i th moment of the monitoring period, is an average skin temperature change amount of the corresponding point of the healthy side at the i th moment of the monitoring period, is an average skin temperature of the rehabilitation position at the i th moment of the monitoring period, is an average skin temperature of the corresponding point of the healthy side at the i th moment of the monitoring period, is a target activity amount, is an actual completed activity amount at the i th moment of the monitoring period, is a visual analogue scale of pain at the i th moment of the monitoring period. 4.The intelligent Chinese medicine rehabilitation monitoring method based on multi-modal data fusion according to claim 1, characterized in that, According to the first dynamic weight, the second dynamic weight, the third dynamic weight, the target activity amount, the historical average heart rate variability low frequency power, the physiological regulation data, the local circulation data and the recovery execution data, a traditional Chinese medicine rehabilitation state coefficient is determined, including: according to the formula: Determine the TCM rehabilitation state coefficient of the i th moment of the monitoring period , wherein , and is a preset coefficient, max is a maximum value function, is the first dynamic weight of the i th moment of the monitoring period, is the second dynamic weight of the i th moment of the monitoring period, is the third dynamic weight of the i th moment of the monitoring period, is the heart rate variability low frequency power of the i th moment of the monitoring period, is the ideal heart rate variability low frequency power, is the historical average heart rate variability low frequency power is the RR interval short-term variation coefficient of the i th moment of the monitoring period, is the rehabilitation position average skin temperature of the i th moment of the monitoring period, is the healthy side corresponding point average skin temperature of the i th moment of the monitoring period, is the rehabilitation position average skin temperature change amount of the i th moment of the monitoring period, is the healthy side corresponding point average skin temperature change amount, is the target activity amount, is the actual completed activity amount of the i th moment of the monitoring period, is the visual analogue scale of pain of the i th moment of the monitoring period, is the joint range of motion compliance rate of the i th moment of the monitoring period. 5.The intelligent Chinese medicine rehabilitation monitoring method based on multi-modal data fusion according to claim 1, characterized in that, According to the traditional Chinese medicine rehabilitation state coefficient, a comprehensive prediction risk coefficient is determined, including: According to the traditional Chinese medicine rehabilitation state coefficient, a physiological regulation sub-dimension coefficient, a local circulation sub-dimension coefficient and a recovery execution sub-dimension coefficient are determined; According to the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient and the recovery execution sub-dimension coefficient, a physiological regulation sub-dimension coefficient standard deviation, a physiological regulation sub-dimension coefficient average value, a local circulation sub-dimension coefficient standard deviation, a local circulation sub-dimension coefficient average value, a recovery execution sub-dimension coefficient standard deviation and a recovery execution sub-dimension coefficient average value in a historical preset time period are determined; The total sampling number, the first dimension rebound number, the second dimension rebound number and the third dimension rebound number of the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient and the recovery execution sub-dimension coefficient in the historical preset time period are determined; According to the first dynamic weight, the second dynamic weight, the third dynamic weight, the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, the recovery execution sub-dimension coefficient, the physiological regulation sub-dimension coefficient standard deviation, the physiological regulation sub-dimension coefficient average value, the local circulation sub-dimension coefficient standard deviation, the local circulation sub-dimension coefficient average value, the recovery execution sub-dimension coefficient standard deviation, the recovery execution sub-dimension coefficient average value, the total sampling number, the first dimension rebound number, the second dimension rebound number and the third dimension rebound number, a comprehensive prediction risk coefficient is determined.

6. The intelligent Chinese medicine rehabilitation monitoring method based on multi-modal data fusion according to claim 5, characterized in that, According to the first dynamic weight, the second dynamic weight, the third dynamic weight, the physiological regulation sub-dimension coefficient, the local circulation sub-dimension coefficient, the recovery execution sub-dimension coefficient, the physiological regulation sub-dimension coefficient standard deviation, the physiological regulation sub-dimension coefficient average value, the local circulation sub-dimension coefficient standard deviation, the local circulation sub-dimension coefficient average value, the recovery execution sub-dimension coefficient standard deviation, the recovery execution sub-dimension coefficient average value, the total sampling number, the first dimension rebound number, the second dimension rebound number and the third dimension rebound number, a comprehensive prediction risk coefficient is determined, including: according to the formula: determining a comprehensive prediction risk coefficient of an i-th moment of a monitoring period , wherein is a first preset weight value, is a second preset weight value, K is a preset constant, is a first dynamic weight of an i-th moment of a monitoring period, is a second dynamic weight of an i-th moment of a monitoring period, is a third dynamic weight of an i-th moment of a monitoring period, is a maximum value taking function, is a physiological adjustment sub-dimension coefficient of an i-th moment of a monitoring period, is a length of a historical preset time period, is a physiological adjustment sub-dimension coefficient of a moment corresponding to a historical preset time period before an i-th moment of a monitoring period, is a first dimension rebound number in a historical preset time period corresponding to an i-th moment of a monitoring period, is a total sampling number in a historical preset time period corresponding to an i-th moment of a monitoring period, is a physiological adjustment sub-dimension coefficient standard deviation in a historical preset time period corresponding to an i-th moment of a monitoring period, is a physiological adjustment sub-dimension coefficient average value in a historical preset time period corresponding to an i-th moment of a monitoring period, is a local cycle sub-dimension coefficient of an i-th moment of a monitoring period, is a local cycle sub-dimension coefficient of a moment corresponding to a historical preset time period before an i-th moment of a monitoring period, is a second dimension rebound number in a historical preset time period corresponding to an i-th moment of a monitoring period, is a local cycle sub-dimension coefficient standard deviation in a historical preset time period corresponding to an i-th moment of a monitoring period, is a local cycle sub-dimension coefficient average value in a historical preset time period corresponding to an i-th moment of a monitoring period, is a recovery execution sub-dimension coefficient of an i-th moment of a monitoring period, is a recovery execution sub-dimension coefficient of a moment corresponding to a historical preset time period before an i-th moment of a monitoring period, is a third dimension rebound number in a historical preset time period corresponding to an i-th moment of a monitoring period, is a recovery execution sub-dimension coefficient standard deviation in a historical preset time period corresponding to an i-th moment of a monitoring period, is a recovery execution sub-dimension coefficient average value in a historical preset time period corresponding to an i-th moment of a monitoring period.

7. A smart traditional Chinese medicine rehabilitation monitoring system based on multi-modal data fusion, characterized in that, For performing the method of any one of claims 1-6, comprising: A real-time data module is configured to acquire user detection data at multiple time points of a monitoring period; A historical data module is configured to acquire rehabilitation scheme information and historical user detection data; A rehabilitation coefficient module is configured to determine a traditional Chinese medicine rehabilitation state coefficient according to the historical user detection data, the user detection data and the rehabilitation scheme information; A prediction risk module is configured to determine a comprehensive prediction risk coefficient according to the traditional Chinese medicine rehabilitation state coefficient; A rehabilitation monitoring module is configured to monitor according to the traditional Chinese medicine rehabilitation state coefficient and the comprehensive prediction risk coefficient.

Images