BCW theory-based personalized rehabilitation exercise method for patients with rheumatoid arthritis

Abstract

The invention relates to the technical field of digital health, in particular to a personalized rehabilitation exercise method for rheumatoid arthritis patients based on a BCW theory, which comprises the following steps: acquiring joint angle change and calculating a pain delay response coefficient, detecting fluctuation of supporting force and stabilizing force to obtain a joint load tolerance value, analyzing respiratory coordination to generate a respiratory coordination index, and performing rehabilitation training on the respiratory coordination index. The method comprises the steps that a joint angle sensor and pain intensity data are analyzed, a BCW capacity factor weight value is generated in a weighted mode, a fuzzy control algorithm is input to conduct hierarchical mapping adjustment on action amplitude frequency and rest intervals, and exercise adaptability feedback data are generated through monitoring completion degree and pain feedback. The exercise intensity is dynamically adjusted in real time through a data model, a personalized rehabilitation scheme is automatically optimized, the problems that subjective feedback is relied on, and physiological changes cannot be responded in time are solved, more accurate exercise amount adjustment is provided, the situation that the joint burden is increased due to improper exercise intensity is avoided, and the rehabilitation process is effectively promoted.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of digital health, in particular to a personalized rehabilitation exercise method for rheumatoid arthritis patients based on the BCW theory. BACKGROUND

[0002] The technical field of digital health involves monitoring, managing and intervening in human health status using information technology, networking and intelligent means. Its core tasks include health data collection, analysis and application, covering wearable device monitoring, physiological signal collection, health information management systems, disease management platforms and remote medical services. This technical field integrates medicine, information technology and sports science to provide digital support for individualized health management and disease rehabilitation, and is widely used in chronic disease monitoring, rehabilitation guidance and clinical decision support.

[0003] Among them, the traditional personalized rehabilitation exercise method for rheumatoid arthritis patients refers to the exercise prescription and training plan developed by doctors' experience or general rehabilitation guide manual for rheumatoid arthritis patients at different stages of the disease to improve joint function and motor ability. The traditional method uses questionnaires to assess the patient's condition, physical examination to determine joint mobility, and then selects appropriate exercise types and intensity based on the experience of rehabilitation physicians, such as joint mobility training, muscle strength training, hydrotherapy training or light aerobic exercise, etc. This method mainly uses regular manual assessment and face-to-face guidance as the main means, relying on patient subjective feedback and medical staff observation to adjust the rehabilitation program, lacking continuous objective data support and dynamic individualized adjustment basis.

[0004] The existing technology mainly relies on subjective assessment and experience judgment, lacking real-time and objective data support. In the process of individualized rehabilitation, the development and adjustment of exercise prescription are based on patient feedback and regular examination results, which makes the response to changes in joint load and physiological response not timely. The traditional method fails to effectively capture the patient's immediate state, causing the execution of the rehabilitation plan to deviate from the patient's actual needs, leading to inappropriate exercise amount, affecting treatment effectiveness, and even increasing joint pressure or delaying rehabilitation progress. SUMMARY

[0005] In order to solve the technical problems that the existing technology mainly relies on subjective assessment and experience judgment, lacking real-time and objective data support. In the process of individualized rehabilitation, the development and adjustment of exercise prescription are based on patient feedback and regular examination results, which makes the response to changes in joint load and physiological response not timely. The traditional method fails to effectively capture the patient's immediate state, causing the execution of the rehabilitation plan to deviate from the patient's actual needs, leading to inappropriate exercise amount, affecting treatment effectiveness, and even increasing joint pressure or delaying rehabilitation progress, the embodiments of the present application provide a personalized rehabilitation exercise method for rheumatoid arthritis patients based on the BCW theory, comprising the following steps:

[0006] To achieve the above object, the application adopts the following technical scheme: a personalized rehabilitation exercise method for rheumatoid arthritis patients based on the BCW theory, comprising the following steps: S1: Obtain the joint angle change sequence through the joint angle sensor and perform time series difference processing, monitor the pain intensity time series data, perform time difference calculation on the joint angle change sequence angle peak time point and the pain intensity time series data pain peak time point, and generate a pain delay response coefficient; S2: Call the pain delay response coefficient, detect the joint support force value and the stable force fluctuation amplitude in the action bearing stage, perform weighted ratio operation to obtain the joint load tolerance value, collect the inspiration time and expiration time of five consecutive breathing cycles, and input them into the variance analysis algorithm to calculate and generate a breathing coordination index; S3: Based on the breathing coordination index, perform threshold segmentation and trend consistency judgment on the joint load tolerance value, perform positive weighted processing when the numerical change direction is the same, and perform negative weighted processing when the direction is opposite, and generate a BCW ability factor weight value; S4: Call the BCW ability factor weight value to input the fuzzy control algorithm hierarchical mapping, collect the action amplitude parameter, action frequency parameter and rest interval parameter of the current training period, and adjust and calculate the three types of parameters according to the hierarchical results, and generate a personalized exercise intensity configuration scheme.

[0007] As a further scheme of the application, the pain delay response coefficient includes the time difference between the pain peak and the joint angle peak, the joint angle change rate and the pain intensity change rate, the joint load tolerance value includes the joint support force fluctuation amplitude, the stable force amplitude and the weighted ratio, the breathing coordination index includes the inspiration and expiration time ratio, the breathing frequency stability and the cycle variability, and the BCW ability factor weight value includes the trend consistency weight, the weighted direction value and the threshold adjustment coefficient.

[0008] As a further scheme of the application, the specific steps of S1 are: S101: Obtain the joint angle change sequence through the joint angle sensor, extract the angle values of adjacent sampling points and calculate the angle difference, compare the angle difference with the time interval to obtain the angular velocity distribution, record the extreme point time index according to the amplitude change, and generate a set of angle peak time points; S102: Detect the pain intensity time series data, extract the pain values of adjacent sampling points and calculate the change rate, compare the change rate with the fluctuation amplitude threshold to determine the local peak candidate point, filter the time index point according to the time interval and the intensity stability range, and generate a set of pain peak time points; S103: Call the pain peak time point set and the angle peak time point set, match the corresponding time points and calculate the time difference, form a delay sequence, take the average, quantify the average value as a factor to represent the time response relationship, and generate a pain delay response coefficient.

[0009] As a further scheme of the present application, the specific steps of S2 are: S201: Call the pain delay response coefficient, detect the joint support force value in the action load bearing stage, calculate the support force instantaneous value sequence, compare it with the pain delay response coefficient matrix item by item, calculate the time node response adjustment factor, and perform weighted superposition to generate the joint support force adjustment value; S202: Based on the joint support force adjustment value, detect the stable force fluctuation amplitude in the load bearing stage, perform the ratio operation of the fluctuation amplitude and the support force adjustment value, weight and smooth the comparison value sequence in the time window, extract the aggregated interval data, and generate the joint load tolerance value; S203: According to the joint load tolerance value, collect the inspiration and expiration time length data of five breathing cycles, obtain the inspiration-expiration ratio time length sequence, input the sequence into the variance analysis algorithm for calculation, difference and normalize the variance result and the breathing coordination reference value, and generate the breathing coordination index.

[0010] As a further scheme of the present application, the joint support force adjustment value refers to the corrected support force value calculated based on the joint support force instantaneous value sequence and the pain delay response coefficient matrix; The joint load tolerance value refers to the result of the ratio of the stable force fluctuation amplitude in the load bearing stage and the joint support force adjustment value after time window weighting and smoothing; The difference normalization refers to the operation mode of standardizing the difference between the variance analysis result and the breathing coordination reference value.

[0011] As a further scheme of the present application, the specific steps of S3 are: S301: Based on the breathing coordination index, calculate the signal fluctuation rate in the continuous time window and record the direction identifier, perform segmented mapping according to the joint load tolerance threshold interval, extract the average amplitude and direction identifier combination, and obtain the joint load segmented vector set; S302: According to the joint load segmented vector set, compare the adjacent time period direction identifier data, call the average amplitude result, calculate the direction matching proportion according to the time sequence order and convert it into a matrix form, and obtain the direction consistency judgment matrix; S303: For the direction relationship data in the direction consistency judgment matrix, call the breathing coordination index and the joint load tolerance value for normalization processing and perform positive and negative weighting operation to generate the BCW ability factor weight value.

[0012] As a further scheme of the present application, the joint load tolerance threshold interval is a numerical interval of a corresponding bearable load range of the joint under a specific motion condition. The direction consistency determination matrix is a matrix data structure corresponding to the direction relationship of the joint load change of adjacent time periods.

[0013] As a further scheme of the present application, the specific steps of S4 are: S401: calling the BCW capability factor weight value input fuzzy control algorithm hierarchical mapping, performing fuzzy processing on the input weight sequence according to the grade interval boundary, calculating the weight component membership degree and aggregation, performing classification calculation according to the difference of the grade interval boundary, and generating a fuzzy grade mapping matrix; S402: based on the fuzzy grade mapping matrix, collecting the action amplitude parameter, the action frequency parameter and the rest interval parameter, calculating the mean value and the volatility rate of the time sequence sample, comparing the mean value and the grade center difference value and performing proportional adjustment to obtain a parameter adjustment coefficient set; S403: according to the parameter adjustment coefficient set, performing weighted operation on the action amplitude parameter, the action frequency parameter and the rest interval parameter, aggregating to form a configuration parameter vector, performing proportional distribution calculation according to the weight distribution relationship, and generating a personalized exercise intensity configuration scheme.

[0014] As a further scheme of the present application, the method further comprises a S5 step: S5: performing parameter configuration based on the adjusted action amplitude parameter, action frequency parameter and rest interval parameter in the personalized exercise intensity configuration scheme, monitoring the actual action completion degree and pain feedback intensity, performing deviation calculation, and generating exercise adaptability feedback data; The exercise adaptability feedback data includes actual action completion degree deviation, pain feedback difference and intensity adaptation adjustment amount.

[0015] As a further scheme of the present application, the specific steps of S5 are: S501: based on the adjusted action amplitude parameter, action frequency parameter and rest interval parameter in the personalized exercise intensity configuration scheme, establishing a parameter configuration set, performing instruction sequence index mapping, collecting displacement and angle signals, calculating the change rate and performing frequency decomposition, and generating an action execution feature matrix; S502: calling the action execution feature matrix, performing matching degree calculation on the displacement and angle change amount, calculating the action difference value and the action completion degree threshold value comparison, recording the frame index below the threshold value, performing time sequence aggregation and normalization processing, and generating an action completion degree evaluation data set; S503: According to the action completion degree evaluation data set and the pain feedback intensity signal, the completion rate and the pain average in the time window are extracted, the correlation coefficient and the deviation baseline value difference are calculated, the deviation value sequence is constructed and smoothed and weighted integrated, and the exercise adaptability feedback data is generated.

[0016] Compared with the prior art, the advantages and positive effects of the present application are that: In the present application, through the difference analysis of the joint angle sensor and the pain intensity data, the joint load tolerance value and the respiratory coordination index are accurately quantified. Using this data model to dynamically adjust the exercise intensity in real time, the personalized rehabilitation plan can be automatically optimized based on the physiological response of the patient. This scheme solves the problem that the traditional method relies too much on subjective feedback and cannot respond to physiological changes in patients in a timely manner, and can provide more accurate exercise adjustment, reduce the over or insufficient exercise intensity, and avoid the joint burden caused by inappropriate exercise plans, thereby more effectively promoting the rehabilitation process. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0018] Figure 1 The step flowchart of the present application is shown in the figure. Figure 2 The S1 refinement schematic diagram of the present application is shown in the figure. Figure 3 The S2 refinement schematic diagram of the present application is shown in the figure. Figure 4 The S3 refinement schematic diagram of the present application is shown in the figure. Figure 5 The S4 refinement schematic diagram of the present application is shown in the figure. Figure 6 The S5 refinement schematic diagram of the present application is shown in the figure. DETAILED DESCRIPTION

[0019] The technical solutions in the present application will be described below in combination with the drawings.

[0020] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0021] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, they convey the same meaning.

[0022] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0023] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0024] Please see Figure 1 This invention provides a personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory, including the following steps: S1: Acquire joint angle change sequence through joint angle sensor and perform time-series difference processing, monitor pain intensity time-series data, calculate the time difference between the angle peak time point of joint angle change sequence and the pain peak time point of pain intensity time-series data, and generate pain delay response coefficient; S2: Call the pain delay response coefficient, detect the joint support force value and stability force fluctuation amplitude during the weight-bearing phase of the movement, perform weighted ratio calculation to obtain the joint load tolerance value, collect the inspiratory and expiratory durations of five consecutive respiratory cycles, and input them into the variance analysis algorithm to calculate and generate respiratory coordination index. S3: Based on the respiratory coordination index, threshold segmentation and trend consistency judgment are performed on the joint load tolerance value. When the numerical change direction is the same, positive weighting is performed, and when the direction is opposite, negative weighting is performed to generate BCW ability factor weight value. S4: Call the BCW ability factor weight value input fuzzy control algorithm for hierarchical mapping, collect the motion amplitude parameters, motion frequency parameters and rest interval parameters of the current training cycle, adjust and calculate the three types of parameters according to the hierarchical results, and generate a personalized training intensity configuration scheme. S5: Based on the adjusted movement amplitude parameters, movement frequency parameters, and rest interval parameters in the personalized exercise intensity configuration scheme, execute parameter configuration, monitor the actual movement completion degree and pain feedback intensity, perform deviation calculation, and generate exercise adaptation feedback data; The delayed response coefficient for pain includes the time difference between the peak pain level and the peak joint angle, the rate of change of joint angle, and the rate of change of pain intensity. The joint load tolerance value includes the fluctuation amplitude of joint support force, the amplitude of stability force, and the weighted ratio. The respiratory coordination index includes the ratio of inspiratory to expiratory time, respiratory rate stability, and periodic variability. The BCW ability factor weight value includes the trend consistency weight, the weighted direction value, and the threshold adjustment coefficient. The exercise adaptation feedback data includes the deviation of actual movement completion, the difference in pain feedback, and the intensity adaptation adjustment amount.

[0025] Please see Figure 2 The specific steps of S1 are as follows: S101: Obtain the joint angle change sequence through the joint angle sensor, extract the angle values ​​of adjacent sampling points and calculate the angle difference, compare the angle difference with the time interval to obtain the angular velocity distribution, record the extreme point time index according to the amplitude change, and generate the angle peak time point set; By fixing a sampling frequency at the patient's knee joint The electronic angle measuring instrument records the patient's position. The sensor collects knee flexion and extension angle data at each sampling point during a complete "seated knee extension" rehabilitation exercise within seconds. (in Each outputs an angle value. The unit is degrees (°), thus forming a system that includes Angle change sequence of data points Next, the angle values ​​of adjacent sampling points are extracted from the sequence. For example, the angle values ​​of the first sampling point are extracted. The sampling point and the first The angle value of each sampling point, i.e. and And perform a subtraction operation on these two values ​​to calculate the angle difference. For example, in Angle of time collection for ,exist Angle of time collection for The angle difference is Then, the calculated angle difference time interval with preset (That is, the reciprocal of the sampling period, here it is) The comparison actually involves performing a division operation to obtain the instantaneous angular velocity at that point in time. In the aforementioned example, the angular velocity is Repeat this process for adjacent data points in the entire joint angle change sequence to generate a set of angular velocity distribution data. Then, based on the original angle sequence The amplitude change is used to record the time index of the extreme point. Specifically, this process involves traversing the entire angle sequence. For each angle value (in from arrive ), and simultaneously call the angle value of its previous point. and the angle value of the next point Determine whether it satisfies and If the condition is met, then the current point's time index will be... Record the data. For example, during a period of motion, some angle data were collected, as shown in Table 1. By analyzing the data in Table 1 point by point, it can be found that at certain time points... At that time, its angle value Greater than the previous time point Angle value And greater than the next time point Angle value ,therefore, The time index of a peak angle point is determined and added to a set. The same judgment process is performed on the data points throughout the entire motion process to collect the time indices of peak points that meet the conditions, and finally generate the set of peak angle time points.

[0026] Table 1: Knee Joint Angle and Angular Velocity Monitoring Data Time index (second) Joint angle (°) Angular velocity (° / s) 2.5 50.5 12.0 2.6 51.8 13.0 2.7 54.1 23.0 2.8 55.2 11.0 2.9 54.8 -4.0 3.0 54.2 -6.0 As shown in Table 1, this table lists the patients' rehabilitation exercises during which... Instant The continuous knee joint angle data collected over a period of seconds, and the instantaneous angular velocity calculated from the data, demonstrate the peak process of the angle changing from rising to falling.

[0027] S102: Detect time-series data of pain intensity, extract pain values ​​from adjacent sampling points and calculate the rate of change, compare the rate of change with the fluctuation amplitude threshold to determine local peak candidate points, filter time index points according to time interval and intensity stability range, and generate a set of pain peak time points; When simultaneously detecting the time-series data of pain intensity reported by the patient, specifically, the patient holds a device that can output... to The pressure sensor uses continuous numerical values; higher values ​​indicate stronger pain. This device operates similarly to a joint angle sensor. Data is collected at a frequency to generate a pain intensity sequence. Then, the pain values ​​of two adjacent sampling points were extracted from the sequence. and The difference in pain values ​​is obtained by performing a subtraction operation. And divide the difference by the time interval. This allows for the calculation of the rate of change in pain. For example, in Pain values ​​recorded at the time for ,exist Pain values ​​recorded at the time for The rate of change in pain is... The unit is per second. Next, the calculated rate of change in pain will be... With a preset fluctuation range threshold For comparison, this threshold is set with reference to the standard deviation of pain signals at rest, specifically by collecting data from patients when they are not engaged in any activity. Pain data for minutes, calculate its standard deviation ,like for Then the threshold Set as Double the standard deviation, i.e. Units per second, when the calculated rate of change Greater than For example Then the corresponding time point (This is) A local peak candidate point is marked as a local peak. After determining the candidate points, the time index of the candidate points is filtered according to the time interval and intensity stability range. First, a minimum time interval is set. Its value is set based on clinical experience. Iterate through the candidate points; if the time difference between any two candidate points is less than... Then, only the time index of the point with the larger pain intensity value is retained. Secondly, a stable intensity range is set. Its setting is based on the minimum perceptible difference in pain values, and is generally set to... For each candidate point that passed the time interval screening, for example, in [number of units], [number of units] units. Pain value at the location Then check its after Within seconds (i.e.) The pain value of (a number of data points), if this Pain value at each point and In comparison, the increases did not exceed ,Right now to The values ​​are all less than Then it is officially confirmed. For a time index of a pain peak point, the time index points that have passed the above filtering process are collected to generate a set of pain peak time points.

[0028] S103: Call the pain peak time point set and the angle peak time point set, match the corresponding time points and calculate the time difference, form a delay sequence, take the average value, quantify the average value into a factor to characterize the time response relationship, and generate the pain delay response coefficient. First, the generated pain peak time point set and angle peak time point set are retrieved. Assuming the obtained angle peak time point set is... The obtained set of pain peak time points is Next, the two time point sets are matched, with the matching rule being that the peak time point for each angle is [matching rule]. Find the time closest to and later than the peak pain time point from the set of peak pain time points. Time point At the same time, set a maximum matching window. for ,Right now Must be smaller than If the value exceeds the threshold, it is considered that there is no corresponding pain peak at that angle peak point. According to this rule, the first angle peak point... With pain peak Matching, second angular peak point and Matching, third angle peak point and Matching is performed, and all matched pairs meet the time window requirements. After matching is completed, the time points of each matched pair are analyzed. Perform the subtraction operation and calculate the time difference between the two. For the above example, the calculated time differences are as follows: , , This forms a delayed sequence that includes the time difference. Then, the values ​​in the delayed sequence are summed and divided by the number of elements in the sequence to calculate the average time difference. ,Right now Finally, this average delay time The data is quantified and mapped to a pre-defined factor table, which is set based on the clinical interpretation of pain response time. The specific intervals are: Level 1: Level 2: Greater than 0.5s and less than or equal to 1s; Level 3: Greater than 1s and less than or equal to 2s; Level 4: The calculated average Compared with the interval, we found Therefore, the average value falls into the Level 2 range. This quantified level "Level 2" is used as the final factor to characterize the time response relationship between movement and pain, and finally generates the pain delayed response coefficient.

[0029] Please see Figure 3 The specific steps of S2 are as follows: S201: Call the pain delay response coefficient, detect the joint support force value during the weight-bearing phase of the movement, calculate the instantaneous support force value sequence, compare it item by item with the pain delay response coefficient matrix, calculate the time node response adjustment factor, and perform weighted superposition to generate the joint support force adjustment value. The generated pain delay response coefficient, quantified as "Level 2," is invoked within a pre-defined response coefficient lookup table (a table that maps pain response levels to specific adjustment values). In the matrix, this level corresponds to a set of adjustment factors. Subsequently, as the patient performed the rehabilitation exercise of "from sitting to standing," a pressure sensor array placed on the soles of the patient's feet detected the joint support force during the weight-bearing phase of the movement. The frequency of continuous acquisition of ground reaction force data, the acquisition process is continuous seconds, total obtained These data points directly constitute the instantaneous value sequence of the support force. For example, in the core load-bearing phase of the movement (from Instant arrival Continuous data collected in seconds The force values ​​are respectively Next, this set of instantaneous support force values ​​will be compared with the corresponding "Level 2" adjustment factor sequence found in the pain delay response coefficient matrix. A step-by-step comparison is performed. This comparison essentially prepares the adjustment factors for each time point in the subsequent weighted calculation. When, the response adjustment factor is ,exist When, the response adjustment factor is And so on, and then, for this Set a set of weighting coefficients for each core time node. This coefficient is set according to the importance of each stage of the movement; the closer to the moment of maximum load, the higher the weight. Specifically, the setting is based on a quantitative assessment of the impact of multiple stages on joint stability derived from biomechanical analysis of the movement by clinical rehabilitation experts. Next, the instantaneous value of the support force at each time point. Time-based response adjustment factor and weighting coefficients The three terms are multiplied, and then the products are summed. The calculation process is as follows: First term product The product of the second term The product of the third term The product of the fourth term The product of the fifth term Finally, add these five results together. The final joint support force adjustment value is generated. .

[0030] S202: Based on the joint support force adjustment value, detect the stability force fluctuation amplitude during the load-bearing stage, perform the ratio calculation of the fluctuation amplitude and the support force adjustment value, perform time window weighted smoothing on the comparison value sequence, extract the aggregated interval data, and generate the joint load tolerance value; Based on the generated joint support force adjustment value During the same "from sitting to standing" weight-bearing phase, the fluctuation amplitude of the stability force during the weight-bearing phase is detected by analyzing the ground reaction force data sequence collected by pressure sensors. Specifically, this detection process involves calculating the absolute difference in support force values ​​between two consecutive sampling points to quantify the force fluctuation. For example, the aforementioned support force sequence can be used. (Unit: N), calculated fluctuation amplitude sequence for , , , The fluctuation amplitude sequence is obtained. Subsequently, the fluctuation range was executed. With support force adjustment value The ratio operation divides each element in the fluctuation amplitude sequence by 1 / 2. Generate a dimensionless ratio sequence. The calculation process is as follows , , , The ratio sequence is obtained. Next, this ratio sequence is subjected to time-window weighted smoothing, with a time window size of [value missing]. Time window and corresponding weighting coefficient The weighting coefficient is set to highlight the influence of the central data point while also taking into account the trend of the preceding and following data, thus affecting the sequence. Perform smoothing calculation on the second point in the calculation, and call... Three values, calculate Similarly, for sequences Perform smoothing calculations on the third point in the calculation, and call... Three values, calculate The smoothed sequence becomes Finally, aggregated interval data is extracted from this smoothed sequence. This aggregated interval is defined as the interval with the highest force value during the entire load-bearing stage. The time period, in this example, corresponds to the smoothed time. and Given two data points, perform an average operation on these two data points. The final joint load tolerance value is generated. .

[0031] S203: Based on the joint load tolerance value, collect the inspiratory and expiratory duration data of five respiratory cycles, obtain the inspiratory-expiratory ratio duration sequence, input the sequence into the variance analysis algorithm for calculation, and perform differential normalization between the variance result and the respiratory coordination benchmark value to generate a respiratory coordination index. Based on the obtained joint load tolerance values While the patient maintained a semi-squatting position close to this load level, the inspiratory and expiratory duration data for five complete respiratory cycles were continuously collected using a respiratory strap sensor worn on the patient's chest and abdomen. The specific data collected are shown in Table 2. Next, the inspiratory-to-expiratory duration sequence was obtained based on the data in Table 2. This process involved dividing the inspiratory duration of each respiratory cycle by the corresponding expiratory duration. The calculation process was as follows: the ratio for cycle one was... The ratio of period two is The ratio of period three is The ratio of period four is The ratio of period five is Thus, the inspiratory-to-expiratory ratio time sequence was obtained. Then, input this sequence into the variance calculation process, first calculating the variance of the sequence. average Then calculate the square of the difference between each data point and the mean. , , , , Sum the five squared differences and divide by the number of data points. The variance results were obtained. Then, this variance result is compared with a pre-defined respiratory coordination baseline. Differential normalization was performed, and the baseline value was derived from the statistical mean of the variance of the respiratory ratio in healthy adults at rest, and set as [value missing]. At the same time, a maximum expected variance of imbalance is set. This value is derived from statistics on patients with respiratory dysfunction. The normalization calculation first obtains the difference between the calculated variance and the benchmark value. Then divide this difference by the difference between the maximum variance and the benchmark value, i.e. The final respiratory coordination index is generated as follows .

[0032] Table 2: Duration data of five respiratory cycles Breathing cycle Inhalation duration (second) Exhalation duration (second) One 2.2 2.6 Two 2.4 2.7 Three 2.1 2.5 Four 2.5 2.9 Five 2.3 2.6 As shown in Table 2, this table records the time taken for each inhalation and exhalation in five consecutive respiratory cycles under specific load conditions.

[0033] Please see Figure 4 The specific steps of S3 are as follows: S301: Based on the respiratory coordination index, calculate the signal fluctuation rate within a continuous time window and record the direction identifier. Perform segmented mapping according to the joint load tolerance threshold range, extract the combination of average amplitude and direction identifier, and obtain the joint load segmented vector set. Based on the generated respiratory coordination index, its initial value is During ongoing rehabilitation monitoring, this indicator was continuously sampled, with a time window set as [missing information]. Seconds, continuous monitoring A time series signal is formed by dividing the time window. First, calculate the rate of fluctuation of the signal within a continuous time window. Specifically, calculate the difference between the index values ​​of two adjacent time windows. For example, the fluctuation of the second window compared to the first window is... The fluctuation of the third window compared to the second window is The fluctuation of the fourth window compared to the third window is At the same time, the direction of the fluctuation value is recorded according to its sign; a positive value is recorded as... Negative values ​​are denoted as Thus, the direction identifier sequence is obtained as follows: Next, segmented mapping is performed based on a preset joint load tolerance threshold range. This threshold range is set with reference to joint load tolerance values ​​from a large number of clinical cases (a... arrive The distribution of dimensionless values ​​(between 0 and 0.1) is divided into three levels: low tolerance zone (≥0 and <0.1), medium tolerance zone (≥0.1 and <0.2), and high tolerance zone (≥0.2). In the calculated joint load tolerance value This value falls into the medium tolerance zone. The mapping rule is to adjust the calculated volatility rate according to its tolerance zone. Here, since it is in the medium tolerance zone, no adjustment is made to the volatility rate. If it were in the low tolerance zone, the volatility rate would be multiplied by a coefficient. If it is in a high tolerance zone, multiply by a coefficient. Then, the average amplitude within each time window is extracted and combined with the direction indicator generated at the end of the window. The average amplitude is the average value of the sampling points within the window, which is simplified here to be the average of the starting and ending values ​​of the window. For example, the average amplitude of the first window is... The direction identifier generated at its endpoint is The average amplitude of the second window is Its direction is indicated as The average amplitude of the third window is Its direction is indicated as The specific data is shown in Table 3. The average amplitude and the corresponding direction indicator are combined into vectors. For example, the first vector is... The second vector is The third vector is This ultimately generates a set of joint load segment vectors.

[0034] Table 3: Time Series Data of Joint Load-Related Indicators Time window Breathing coordination index Joint load tolerance value Fluctuation direction identifier 1 0.0485 0.123 (starting) 2 0.0510 0.118 +1 3 0.0490 0.125 -1 4 0.0525 0.130 +1 As shown in Table 3, this table lists the continuous The changes in respiratory coordination indicators and joint load tolerance values ​​within a monitoring time window, and the directional indicators calculated based on the changes in respiratory coordination indicators.

[0035] S302: Based on the joint load segmented vector set, compare the direction identification data of adjacent time periods, call the average amplitude result, calculate the direction matching ratio according to the time series order and convert it into matrix form to obtain the direction consistency judgment matrix. Based on the generated joint load segmentation vector set, which contains the direction identifier sequence generated at the end of each time window. First, the directional identifier data of adjacent time periods in this sequence are compared. Specifically, the first identifier is extracted. and the second identifier The two are compared to determine if they are equal. If they are not equal, the second identifier is extracted. And the third identifier The comparison also yielded unequal results. After the comparison was completed, the calculated average amplitude result for each window was retrieved. And calculate the direction matching ratio according to the time series order. The calculation process is as follows: in the total number of... In each of the adjacent comparisons, the number of times the direction indicators are equal is: Therefore, the proportion of direction matching is... This single matching ratio result is then expanded into a matrix to describe the specific patterns of directional changes in more detail, constructing a... A matrix, where rows represent the starting direction, columns represent the ending direction, and matrix elements... Recorded from direction Change direction Number of times, direction Defined as state ,direction Defined as state In the sequence There exists a time from arrive The transformation, and a transition from arrive The transformation, therefore, in the matrix (from arrive The value of ) , (from arrive The value of ) ,and (from arrive )and (from arrive The conversion of ) did not occur, and its value was ) The count values ​​are then filled into the matrix, resulting in the final direction consistency determination matrix: .

[0036] S303: For the directional relationship data in the directional consistency judgment matrix, the respiratory coordination index and joint load tolerance value are called for normalization and positive / negative weighted calculation, using the following formula: ; Generate BCW capability factor weight values; in, Represents the BCW capability factor weight value. This represents the normalized respiratory coordination index for the i-th time window. This represents the normalized joint load tolerance value for the i-th time window. For direction identifiers, Number of time windows; Based on the described directional relationship data, and by referencing the time series data of respiratory coordination indices and joint load tolerance values ​​(see Table 3), these two sets of data were normalized. First, a reference range for normalization was set. Based on long-term monitoring of rheumatoid arthritis patients, the reasonable fluctuation range of respiratory coordination indices was set as follows: The reasonable fluctuation range of joint load tolerance value is set as follows: Normalization calculation uses a linear mapping method to map the original data to... Interval, for respiratory coordination index sequence Its normalized sequence Calculation process ,get For joint load tolerance value sequence Normalized Calculation process ,get Next, a positive-to-negative weighted operation is performed, which is accomplished using the formula: ; In this formula, This is the final generated BCW capability factor weight value. It is the first Normalized respiratory coordination index for a time window It is the first The joint load tolerance value after normalization within a time window This represents the total number of time windows, here it is... , As a direction identifier, it represents the first... Within a window, the two indicators relative to the previous window ( Whether the directions of change are consistent; if they increase or decrease simultaneously, then... If one increases and the other decreases, then it is... For the first window ( Since there is no preceding window, its direction identifier is conventionally set to... For subsequent windows, based on the trend of the original data in Table 3, in When R increases and L decreases, therefore ,exist When R decreases and L increases, therefore ,exist When R increases and L increases, therefore Substituting the numerical values ​​into the formula, the calculation for the numerator is as follows: ; ; ; ; The summation of the numerators is ; The calculation of the denominator is as follows: ; ; ; ; The summation of the squared terms in the denominator is: ; Take its square root to get Final calculation The weight value of the BCW capability factor is... The advantage of the formula lies in its dynamic combination of two dimensions: respiratory coordination (a reflection of the body's internal homeostasis) and joint load tolerance (a manifestation of external physical capabilities), and the introduction of a weighting factor for directional consistency. The addition of a term (e.g., the denominator term) allows the weight calculation to consider not only the instantaneous values ​​of the two indicators but, more importantly, capture their coordinated changes over time. Simultaneously, the denominator term, acting as a normalized distance metric, adjusts the overall scale of the weight values. This result indicates that the coordination between the two indicators is relatively weak because the final weight values... It is a relatively low positive value. A positive value indicates that the product of the two indicators contributes positively overall, but the value is small, indicating inconsistency in their changes (due to...). The negative values ​​introduced by the item offset most of the positive correlation, and this weight value will serve as an important basis for adjusting the subsequent rehabilitation exercise program.

[0037] Please see Figure 5 The specific steps of S4 are as follows: S401: Call the BCW capability factor weight value input to the fuzzy control algorithm for hierarchical mapping, perform fuzzification processing on the input weight sequence according to the level interval boundary, calculate and aggregate the membership degree of the weight components, perform classification calculation according to the difference of the level interval boundary, and generate a fuzzy level mapping matrix. The generated BCW capability factor weight value is called, and its specific value is: This value is input into a preset grading mapping rule, which, based on clinical experience, sets the range of BCW capability factor weight values. The ability is divided into three fuzzy levels: insufficient, moderate, and good. Each level is defined by a triangular membership function, and the boundary parameters of the level intervals are shown in Table 4. The weight values ​​of the input are based on the level interval boundaries. The fuzzification process is performed by calculating the membership degree of the value for each fuzzy level. For the "insufficient ability" level, the core point is... , range ,because If it is not within this interval, its membership degree is For the "moderate ability" level, the core point is: , range Its membership degree is calculated as follows: For the "Good Ability" level, the core point is: , range ,because If it is outside the left rising edge of this level, its membership degree is calculated as follows: The result is negative, therefore the membership degree is taken as... However, to reflect the transition region, a linear calculation method is used. If the value falls in the overlapping part of the two intervals, the membership degree is calculated separately. It does not fall within the overlap area of ​​"moderate ability" and "good ability". Therefore, its membership degree to "good ability" is... The calculated membership degrees of the weight components are aggregated to obtain a membership vector: (Insufficient ability:) Moderate ability: Good ability: Then, based on the difference in the central values ​​of the rank intervals, a classification calculation is performed, and the membership vector is compared with the central value of the rank. A weighted average is used to calculate a clear rating value. This value will be used for subsequent adjustments, and finally the level definition parameters will be combined with the calculated membership vector to generate a fuzzy level mapping matrix.

[0038] Table 4: Definition Table of Fuzzy Levels for BCW Capability Factors Fuzzy grade Lower limit of interval Grade core Upper limit of interval Inadequate ability 0 0.1 0.25 Moderate ability 0.15 0.3 0.45 Good ability 0.35 0.5 1.0 As shown in Table 4, this table defines three fuzzy levels for the hierarchical mapping of BCW capability factor weight values, as well as interval boundaries and core values ​​for calculating membership degrees.

[0039] S402: Based on the fuzzy level mapping matrix, collect action amplitude parameters, action frequency parameters and rest interval parameters, calculate the mean and volatility of the time series samples, compare the mean with the level center difference and perform proportional adjustment to obtain the parameter adjustment coefficient set; Based on the generated fuzzy level mapping results, especially the calculated clarity level evaluation value First, the rehabilitation equipment collects three key parameters of the patient's current "seated leg raise" movement: amplitude of movement, frequency of movement, and rest interval. The most recent data is then continuously recorded. The training data, the sequence of motion amplitude (unit: degrees) is as follows: The sequence of action frequencies (unit: times / minute) is as follows: The rest interval (unit: seconds) sequence is as follows: Next, the mean and volatility of these three time series samples are calculated, with the mean amplitude being... Degree, frequency mean Repetitions / minute, average rest interval is The volatility is calculated as the ratio of the standard deviation to the mean, but this is not used directly here. Instead, the mean is used for the next comparison. Subsequently, the calculated mean of the parameters is compared with the preset target value for the level center, which is based on the clarity level evaluation value. It is obtained by interpolation from a continuous objective function that defines the ideal training parameters corresponding to different ability levels, for example, level. Corresponding amplitude Degree, level Corresponding amplitude Degree, then level The target range is Similarly, the set frequency target is... Repetitions / minute, with a target rest interval of [number]. Seconds, the difference between the current mean and the target value is calculated, and the magnitude of the difference is... Degree, frequency difference is times / minute, rest interval difference is Finally, adjust the difference proportionally by setting an adjustment ratio coefficient. Its settings are based on the clinical recommendations of rehabilitation therapists, and the initial adjustments are set as follows: Multiply the difference by the coefficient to obtain the amplitude adjustment coefficient. The frequency adjustment coefficient is The rest interval adjustment factor is Finally, the set of parameter adjustment coefficients is obtained. .

[0040] S403: Based on the parameter adjustment coefficient set, perform weighted calculations on the movement amplitude parameter, movement frequency parameter, and rest interval parameter, aggregate them to form a configuration parameter vector, perform proportional allocation calculations based on the weight distribution relationship, and generate a personalized exercise intensity configuration plan; Adjust the coefficient set based on the generated parameters. It also retrieves the average exercise parameters of the current patient, i.e., the range of motion. Degree, frequency of action Repetitions / minute, rest interval In seconds, a weighted calculation is performed on these three parameters. This weighted calculation essentially involves applying adjustment coefficients for correction. Specifically, the mean of the current parameters is added to the corresponding adjustment coefficient, but in the opposite direction—adjusting from the current value towards the target value. Therefore, the new magnitude of the action is... Degree, new action frequency is refreshes / minute, new rest interval is... In seconds, the newly calculated parameter values ​​are aggregated to form a configuration parameter vector. Subsequently, the vector is calculated for final proportional allocation based on a preset weight distribution relationship. This weight distribution relationship defines the adjustment proportion of multiple parameters in a single comprehensive intensity adjustment, and its setting is based on biomechanical analysis and is set as amplitude weight. Frequency weight Rest interval weight This step is used to proportionally allocate the total adjustment amount when an overall intensity increase or decrease is needed (e.g., when the BCW factor itself undergoes a large jump). However, in this adjustment based on subtle feedback, this proportional allocation calculation is used to verify whether the adjusted parameter combination exceeds the existing safety range. Since this adjustment is a fine-tuning, the result of the previous step is directly used, and the result is rounded to conform to actual operation. The final personalized exercise intensity configuration scheme is as follows: the range of motion is set to... Degree, action frequency set to Repetitions / minute, rest interval set to Second.

[0041] Please see Figure 6 The specific steps of S5 are as follows: S501: Based on the adjusted motion amplitude parameters, motion frequency parameters, and rest interval parameters in the personalized exercise intensity configuration scheme, establish a parameter configuration set, execute instruction sequence index mapping, collect displacement and angle signals, calculate the rate of change and perform frequency decomposition, and generate a motion execution feature matrix. Based on the generated personalized exercise intensity configuration scheme, the scheme specifies the adjusted range of motion parameters as follows: Degree, action frequency parameter is The parameters are: times / minute, and rest interval. First, combine these three core parameters to create a parameter configuration set. Next, the instruction sequence index mapping is performed. This process searches a pre-defined rehabilitation movement instruction library, which maps different combinations of parameters to unique execution codes. For example, the "seated leg raise" movement is mapped to... Degree range, The configuration of the frequency per minute is mapped to the instruction index STL-L2-B, thereby setting the current parameter configuration set. The index is precisely matched, and then, as the patient begins to perform the action according to the instruction, inertial measurement unit (IMU) sensors fixed to the knee and ankle are used to... The sampling frequency is used to collect displacement and angle signals in real time during the motion process, in a continuous... In a single leg raise movement lasting a few seconds, a total of [number] data were collected. The frame data includes the knee flexion and extension angle sequence and the ankle spatial displacement sequence. Next, the rate of change of the acquired signal is calculated, which is done by calculating the difference between two consecutive sampling points and then dividing by the sampling time interval. (seconds) to obtain angular velocity and linear velocity, for example, if in The knee joint angle is Degree, in The time angle is If the degree is given, then the angular velocity at that moment is... After performing this operation on the entire time series signal at degrees / second, the resulting angular velocity sequence is then subjected to frequency decomposition. This process does not directly call a specific algorithm name, but is described as follows: decomposing the velocity signal into multiple sinusoidal wave components of different frequencies and calculating the energy proportion of each frequency component, thereby identifying the low-frequency components that dominate the motion (e.g., ...). (corresponding to the overall frequency of the leg-lifting movement) and high-frequency components characterizing movement stability (e.g., Signals within the range, originating from muscle tremors), ultimately identify the key characteristics of this movement, specifically including the maximum angle value achieved (e.g., (degrees), peak angular velocity (e.g.) degrees / second), dominant motion frequency ( ) and the proportion of high-frequency band energy (e.g. Data such as ) are integrated into one Vectors that are continuous This process is repeated for the next step to generate a... The action execution feature matrix.

[0042] S502: Call the action execution feature matrix, perform matching degree calculation on displacement and angle change, calculate the action difference value and compare it with the action completion threshold, record the frame index below the threshold, perform temporal aggregation and normalization processing, and generate the action completion evaluation dataset. Call the generated Action execution feature matrix, which records in detail... The kinematic characteristics of the repeated actions are shown in Table 5. First, for each row of the matrix representing a single action, the key changes in displacement and angle are extracted. For example, for the first action, the measured maximum angle is... The degree, and the target value set in the personalized exercise intensity configuration plan ( The matching degree is calculated by dividing the absolute difference between the measured value and the target value by the target value, and then using... Subtracting that ratio, i.e. This is the matching degree of that single parameter. The matching degrees of multiple key parameters such as angle and angular velocity are weighted and summed according to their influence on the quality of movement. The weighting coefficients are preset by rehabilitation experts based on clinical importance. For example, the weight of angle is 100%. Angular velocity weights are The overall matching degree is The calculated overall matching degree is the action completion degree, and its opposite is the action difference value. Subsequently, the calculated completion degree value of each action is compared with a preset action completion degree threshold. The threshold is set with the following reference: it is allowed to not exceed The execution deviation, therefore the threshold is set to In the successive comparisons, records were made of completion scores lower than [a certain value]. The frame index of the action (simplified here as the sequence number of the action), for example, if the frame index of the action is... Second and third The completion rates of each action are as follows: and Then record the index. Next, time-series aggregation and normalization are performed. Time-series aggregation refers to calculating the time sequence over the entire training group ( In a given set of actions, the percentage of actions that did not meet the standard out of the total number of actions is... Normalization, on the other hand, transforms this non-compliance rate into a standardized completion score, for example, using... Subtracting this ratio, we get Repeat this process for multiple workout sets to obtain a time series consisting of multiple workout completion scores. For example, the scores obtained in three consecutive workout sets would be... Finally, a dataset for evaluating action completion is generated. .

[0043] Table 5. Summary of Action Execution Characteristics Action sequence number Maximum angle (degree) Peak angular velocity (degree / s) Dominant frequency (hertz) High frequency energy proportion (%) 1 86.5 55.1 0.31 8.1 2 87.2 54.8 0.29 7.5 3 84.1 50.3 0.33 11.2 ... ... ... ... ... 10 86.8 55.0 0.30 7.9 As shown in Table 5, this table extracts and displays the key kinematic feature parameters extracted for each repetition of the movement in one training set.

[0044] S503: Based on the action completion assessment dataset and pain feedback intensity signal, extract the completion rate and mean pain within the time window, calculate the correlation coefficient and the difference between the deviation benchmark value, construct the deviation value sequence and perform smoothing and weighted integration to generate exercise adaptation feedback data; Based on the generated action completion evaluation dataset Simultaneously, the pain feedback intensity signal input by the patient via electronic device during the corresponding exercise group was collected. This signal was used... The Numerical Rating Scale (NRS). Represents painlessness. Representing the most intense pain, let's assume the average pain signal intensity corresponding to the three sets of exercises is as follows: First, extract the average completion rate and average pain mean within the specified time window (here, all three sets of exercises). The average completion rate is... The average pain level was Next, the correlation coefficient between the completion sequence and the pain signal sequence is calculated to quantify the linear relationship between them. This calculation involves operations on covariance and standard deviation, yielding a coefficient between... arrive The values ​​between, for example, the calculated result is This indicates that the lower the completion rate of the movement, the stronger the pain. Then, the calculated average completion rate and average pain mean are compared with a preset deviation benchmark value. This benchmark value is set with reference to the ideal rehabilitation state, i.e., a completion rate of... (Near-perfect execution) and pain score: (Almost painless), the calculated differences are the completion deviations. and pain deviation These two deviation values ​​are normalized (e.g., divided by their maximum deviation range) and then combined into a single comprehensive deviation value, for example... The comprehensive deviation values ​​calculated for each evaluation period (e.g., daily) are constructed into a deviation value sequence. For example, the data for three consecutive days would be... The sequence is smoothed by using a window size of [size missing]. The moving average calculation, the smoothed value of the second point. The smoothing value of the third point is Finally, the smoothed sequence is weighted and integrated, giving higher weight to recent data; for example, weighting the most recent deviation value. The previous value is assigned a weight. The final integrated value is This value is the final generated exercise adaptation feedback data.

[0045] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory, characterized in that, Includes the following steps: S1: Acquire joint angle change sequence through joint angle sensor and perform time-series difference processing, monitor pain intensity time-series data, calculate the time difference between the angle peak time point of joint angle change sequence and the pain peak time point of pain intensity time-series data, and generate pain delay response coefficient; S2: Call the pain delay response coefficient, detect the joint support force value and stability force fluctuation amplitude during the weight-bearing phase of the movement, perform weighted ratio calculation to obtain the joint load tolerance value, collect the inhalation duration and exhalation duration of five consecutive respiratory cycles, and input them into the variance analysis algorithm to calculate and generate respiratory coordination index. S3: Based on the respiratory coordination index, threshold segmentation and trend consistency judgment are performed on the joint load tolerance value. When the numerical change direction is the same, positive weighting is performed, and when the direction is opposite, negative weighting is performed to generate BCW ability factor weight value. S4: Call the BCW ability factor weight value input fuzzy control algorithm for hierarchical mapping, collect the motion amplitude parameters, motion frequency parameters and rest interval parameters of the current training cycle, adjust and calculate the three types of parameters according to the hierarchical results, and generate a personalized training intensity configuration scheme.

2. The personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory according to claim 1, characterized in that, The pain delay response coefficient includes the time difference between the peak pain value and the peak joint angle value, the joint angle change rate, and the pain intensity change rate. The joint load tolerance value includes the joint support force fluctuation amplitude, stability force amplitude, and weighted ratio. The respiratory coordination index includes the inspiratory to expiratory time ratio, respiratory rate stability, and periodic variability. The BCW capability factor weight value includes trend consistency weight, weighted direction value, and threshold adjustment coefficient.

3. The personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory according to claim 1, characterized in that, The specific steps of S1 are as follows: S101: Obtain the joint angle change sequence through the joint angle sensor, extract the angle values ​​of adjacent sampling points and calculate the angle difference, compare the angle difference with the time interval to obtain the angular velocity distribution, record the extreme point time index according to the amplitude change, and generate the angle peak time point set; S102: Detect time-series data of pain intensity, extract pain values ​​from adjacent sampling points and calculate the rate of change, compare the rate of change with the fluctuation amplitude threshold to determine local peak candidate points, filter time index points according to time interval and intensity stability range, and generate a set of pain peak time points; S103: Call the set of pain peak time points and the set of angle peak time points, match the corresponding time points and calculate the time difference to form a delayed sequence, take the average value, quantify the average value as a factor to characterize the time response relationship, and generate the pain delayed response coefficient.

4. The personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory according to claim 3, characterized in that, The specific steps of S2 are as follows: S201: Call the pain delay response coefficient, detect the joint support force value during the weight-bearing stage of the action, calculate the instantaneous value sequence of support force, compare it item by item with the pain delay response coefficient matrix, calculate the time node response adjustment factor, and perform weighted superposition to generate the joint support force adjustment value. S202: Based on the joint support force adjustment value, detect the stability force fluctuation amplitude during the load-bearing stage, perform a ratio calculation between the fluctuation amplitude and the support force adjustment value, perform time window weighted smoothing on the comparison value sequence, extract aggregated interval data, and generate joint load tolerance value; S203: Based on the joint load tolerance value, collect the inspiratory and expiratory duration data for five respiratory cycles, obtain the inspiratory-to-expiratory ratio duration sequence, input the sequence into the variance analysis algorithm for calculation, and perform differential normalization between the variance result and the respiratory coordination benchmark value to generate a respiratory coordination index.

5. The personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory according to claim 4, characterized in that, The joint support force adjustment value refers to the corrected support force value calculated based on the instantaneous joint support force value sequence and the pain delay response coefficient matrix. The joint load tolerance value refers to the result of time-window weighted smoothing of the ratio of the stability force fluctuation amplitude to the joint support force adjustment value during the load-bearing stage. The difference normalization refers to the calculation method of standardizing the difference between the ANOVA results and the respiratory coordination baseline.

6. The personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory according to claim 4, characterized in that, The specific steps for S3 are as follows: S301: Based on the respiratory coordination index, calculate the signal fluctuation rate within a continuous time window and record the direction identifier. Perform segmented mapping according to the joint load tolerance threshold range, extract the combination of average amplitude and direction identifier, and obtain the joint load segmented vector set. S302: Based on the joint load segmented vector set, compare the direction identification data of adjacent time periods, call the average amplitude result, calculate the direction matching ratio according to the time series order and convert it into matrix form to obtain the direction consistency judgment matrix. S303: For the directional relationship data in the directional consistency judgment matrix, call the respiratory coordination index and joint load tolerance value for normalization and perform positive and negative weighted operation to generate BCW ability factor weight value.

7. The personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory according to claim 6, characterized in that, The joint load tolerance threshold range is a numerical range of the load that a joint can withstand under specific movement conditions. The direction consistency determination matrix is ​​a matrix data structure corresponding to the direction relationship of joint load changes in adjacent time periods.

8. The personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory according to claim 6, characterized in that, The specific steps of S4 are as follows: S401: Call the BCW capability factor weight value input to the fuzzy control algorithm for hierarchical mapping, perform fuzzification processing on the input weight sequence according to the level interval boundary, calculate and aggregate the membership degree of the weight components, perform classification calculation according to the difference of the level interval boundary, and generate a fuzzy level mapping matrix. S402: Based on the fuzzy level mapping matrix, collect action amplitude parameters, action frequency parameters and rest interval parameters, calculate the mean and volatility of the time series samples, compare the mean with the level center difference and perform proportional adjustment to obtain the parameter adjustment coefficient set; S403: Based on the parameter adjustment coefficient set, perform weighted operations on the movement amplitude parameter, movement frequency parameter and rest interval parameter, aggregate to form a configuration parameter vector, perform proportional allocation calculation according to the weight distribution relationship, and generate a personalized exercise intensity configuration scheme.

9. The personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory according to claim 1, characterized in that, The method also includes step S5: S5: Based on the adjusted movement amplitude parameters, movement frequency parameters and rest interval parameters in the personalized exercise intensity configuration scheme, execute parameter configuration, monitor the actual movement completion degree and pain feedback intensity, perform deviation calculation, and generate exercise adaptation feedback data; The exercise adaptation feedback data includes deviations in actual movement completion, differences in pain feedback, and intensity adaptation adjustments.

10. The personalized rehabilitation exercise method for rheumatoid arthritis patients based on BCW theory according to claim 9, characterized in that, The specific steps of S5 are as follows: S501: Based on the adjusted motion amplitude parameters, motion frequency parameters and rest interval parameters in the personalized exercise intensity configuration scheme, establish a parameter configuration set, execute instruction sequence index mapping, collect displacement and angle signals, calculate the rate of change and perform frequency decomposition, and generate a motion execution feature matrix. S502: Call the action execution feature matrix, perform matching degree calculation on displacement and angle change, calculate the action difference value and compare it with the action completion threshold, record the frame index below the threshold, perform temporal aggregation and normalization processing, and generate the action completion evaluation dataset. S503: Based on the action completion assessment dataset and pain feedback intensity signal, extract the completion rate and mean pain within the time window, calculate the correlation coefficient and the difference between the deviation benchmark value, construct the deviation value sequence and perform smoothing and weighted integration to generate exercise adaptation feedback data.

Images