Muscle strength recovery prediction method, system and device

By adjusting the number of interpolated muscle strength data to be interpolated and using sEMG signals to correct muscle strength signals, the problem of low interpolation accuracy in the prior art is solved, and more accurate muscle strength prediction is achieved.

CN118830850BActive Publication Date: 2024-12-06SICHUAN HEALTH REHABILITATION VOCATIONAL COLLEGE
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411321832.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-23
Publication Date
2024-12-06
Estimated Expiration
2044-09-23

AI Technical Summary

Technical Problem

When constructing the model, the existing muscle strength prediction methods do not match the sampling frequency of the sEMG signal and the muscle strength signal, resulting in low interpolation processing accuracy, which affects the accuracy of the muscle strength prediction results, and does not consider the changes in muscle state and fatigue effects.

Method used

By analyzing the numerical fluctuations of the muscle strength data, adjust the number of to be interpolated, obtain the actual number of to be interpolated in each to be interpolated interval, and construct an adjustment factor based on the sEMG signal to correct the muscle strength signal to establish a more accurate muscle strength prediction model.

Benefits of technology

The interpolation accuracy is improved, the changes in muscle state and the effects of fatigue are taken into account, and the accuracy of muscle strength prediction is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118830850B_ABST
    Figure CN118830850B_ABST
Patent Text Reader

Abstract

The present application discloses a method, system and device for predicting muscle strength recovery, which collects sample muscle strength data and sample sEMG signals, obtains the total number of interpolation values; adjusts the total number of interpolation values ​​according to the numerical fluctuation of the sample muscle strength data, and obtains a number of interpolation points in each interpolation interval; obtains the predicted value of each interpolation point in each interpolation interval, constructs an adjustment factor for each interpolation point, and uses the adjustment factor and the sample sEMG signal to correct the predicted value of each interpolation point to obtain the corrected sample muscle strength data; establishes a muscle strength prediction model; collects the electromyographic signal at the muscle strength point required for prediction of the subject, inputs the electromyographic signal at the muscle strength point required for prediction of the subject into the muscle strength prediction model, and obtains the predicted value of the subject's muscle strength. The present application solves the problem that when constructing a muscle strength prediction model, a fixed value is used when interpolating muscle strength data, causing deviations in the construction of a muscle strength prediction model.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of training prediction, and in particular to a method, system and device for predicting muscle strength recovery. Background Art

[0002] In daily life, muscle injuries often occur. Muscle strength is the maximum force that muscles can produce and is an important indicator to measure muscle health.

[0003] In the field of rehabilitation medicine, the study of muscle strength mainly uses surface electromyographic signals, or sEMG, which are the superposition of action potentials of motor units in many muscle fibers in the time domain. In existing studies, sEMG signals are commonly used to evaluate muscle strength. Common muscle strength prediction methods usually first need to build a muscle strength prediction model based on several pre-collected sEMG signals and muscle strength signals, and then input the sEMG signals of the subject's muscles to be tested into the muscle strength prediction model to obtain the muscle strength prediction results.

[0004] However, when constructing a muscle strength prediction model, the sampling frequency of the sEMG signal is different from that of the muscle strength signal, that is, the number of sampling points in the same time is different, and the number of sampling points of the sEMG signal and the muscle strength signal differs greatly. The sEMG signal and the muscle strength signal cannot be directly synchronized and matched, and thus the muscle strength model cannot be directly constructed. Since the sampling frequency of the muscle strength signal is lower than that of the sEMG signal, the muscle strength signal needs to be interpolated. The common interpolation processing method only sets a fixed number of data to be inserted into the two muscle strength signals adjacent in time series according to the difference in the number of sampling points between the muscle strength signal and the sEMG signal, without considering the fluctuation of the muscle strength signal when the muscle is in different states. At the same time, the inserted value is only set according to the value of the muscle strength signal at both ends of the inserted interval, and the influence of the electromyographic signal is not considered. When the measurement time continues to increase and the muscle is in a fatigue state, the sEMG signal will fluctuate abnormally. At this time, setting the inserted value only according to the size of the previous and next values ​​in the muscle strength signal time series will result in a large error, thereby affecting the accuracy of the muscle strength prediction result.

[0005] The above contents are only used to assist in understanding the technical solution of the present application and do not constitute an admission that the above contents are prior art. Summary of the invention

[0006] The main purpose of the present application is to provide a muscle recovery prediction method, system and device, aiming to solve the technical problem that when interpolating muscle force signals to match muscle force signals with electromyographic signals, the inserted value is set only according to the size of the previous and next values ​​in the time series of the muscle force signal without considering the influence of electromyographic signal fluctuations.

[0007] To achieve the above-mentioned purpose, the present application provides a muscle strength recovery prediction method, comprising: collecting sample muscle strength data and sample sEMG signals, obtaining the total number of values ​​to be interpolated in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signal; allocating the total number of values ​​to be interpolated according to the numerical fluctuation of the sample muscle strength data to obtain the actual number of values ​​to be interpolated in each interval to be interpolated, and dividing each interval to be interpolated based on the actual number of values ​​to be interpolated to obtain a number of points to be interpolated in each interval to be interpolated; obtaining a predicted value of each point to be interpolated in each interval to be interpolated, constructing an adjustment factor for each point to be interpolated, and using the adjustment factor and the sample sEMG signal to correct the predicted value of each point to be interpolated to obtain corrected sample muscle strength data; establishing a muscle strength prediction model based on the corrected sample muscle strength data and the sample sEMG signal; collecting the electromyographic signal at the muscle strength point required to be predicted of the subject, inputting the electromyographic signal at the muscle strength point required to be predicted of the subject into the muscle strength prediction model, and obtaining the predicted value of the subject's muscle strength.

[0008] Optionally, the total number of values ​​to be interpolated is allocated according to the numerical fluctuation of the sample muscle strength data to obtain the actual number of values ​​to be interpolated in each interval to be interpolated, including: determining the degree of numerical fluctuation of each interval to be interpolated, and determining the actual number of values ​​to be interpolated in each interval to be interpolated according to the numerical difference of the numerical fluctuation degree of the interval to be interpolated.

[0009] Optionally, determining the degree of fluctuation of the values ​​of each to-be-interpolated interval includes: using the following formula (1) to determine the degree of fluctuation of the values ​​of each to-be-interpolated interval:

[0010]

[0011] In the formula, Indicates The degree of numerical fluctuation of the interpolation interval, Indicates The first muscle strength data in the time series of the interval to be interpolated, Indicates The last muscle strength data in the time series of the interval to be interpolated, Represents the set of numerical fluctuations of all intervals to be interpolated in the sample muscle strength data. It means to find the absolute value. It means to find the maximum value in a set. It means to find the minimum value in a set.

[0012] Optionally, determining the actual number of interpolated values ​​of each interpolated interval according to the numerical difference of the numerical fluctuation degree of the interpolated interval includes: determining the actual number of interpolated values ​​of each interpolated interval according to the numerical difference of the numerical fluctuation degree of the interpolated interval using the following formula (2):

[0013]

[0014] In the formula, Indicates The actual number of interpolation intervals to be interpolated is Indicates The degree of numerical fluctuation of the interpolation interval, represents the number of intervals to be interpolated. Indicates the total number of values ​​to be interpolated.

[0015] Optionally, constructing the adjustment factor for each point to be interpolated includes: determining the electromyographic interval corresponding to each interval to be interpolated; determining the degree of trend difference between each interval to be interpolated and the corresponding electromyographic interval; and constructing the adjustment factor based on the degree of trend difference between each interval to be interpolated and the sampling time of each point to be interpolated in each interval to be interpolated.

[0016] Optionally, the using the adjustment factor and the sample sEMG signal to correct the predicted value of each to-be-interpolated point to obtain the corrected sample muscle strength data includes: using the adjustment factor and the sample sEMG signal to correct the predicted value of each to-be-interpolated point using the following formula (3) to obtain the corrected sample muscle strength data:

[0017]

[0018] In the formula, Indicates The first The predicted value after correction of the interpolation point, Indicates The first interval to be interpolated The predicted value of the interpolation point, Indicates The first interval to be interpolated The adjustment factor of the interpolation point, Indicates The first The sampling time of the interpolation points, Indicates The sampling time of the first data in the time series in the interval to be interpolated, Indicates The sampling time of the last bit of data in the time series in the interval to be interpolated, Indicates the sample muscle strength data The first The Euclidean distance between the interpolation points and the corresponding signal on the sample sEMG signal, Indicates the sample muscle strength data The Euclidean distance between the first data in the time series in the interpolation interval and the corresponding signal on the sample sEMG signal, Indicates the sample muscle strength data The Euclidean distance between the first data in the time series in the interval to be interpolated and the corresponding signal on the sample sEMG signal.

[0019] Optionally, after obtaining the predicted muscle strength value of the subject, the muscle strength recovery prediction method further comprises: using a muscle strength evaluation method to determine the predicted muscle strength value of the subject to obtain the predicted muscle strength recovery status of the subject.

[0020] In addition, to achieve the above objectives, the present application also provides a muscle strength recovery prediction system, comprising:

[0021] The data acquisition module is used to collect sample muscle strength data and sample sEMG signals, and obtain the total number of interpolation values ​​in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signal; the interval division module is used to allocate the total number of interpolation values ​​according to the numerical fluctuation of the sample muscle strength data, obtain the actual number of interpolation values ​​in each interpolation interval, and divide each interpolation interval based on the actual number of interpolation values ​​to obtain a number of interpolation points in each interpolation interval; the value correction module is used to obtain the total number of interpolation values ​​in each interpolation interval; The predicted value of each to-be-interpolated point in the interval is obtained, an adjustment factor of each to-be-interpolated point is constructed, and the predicted value of each to-be-interpolated point is corrected by using the adjustment factor and the sample sEMG signal to obtain the corrected sample muscle strength data; a model construction module is used to establish a muscle strength prediction model based on the corrected sample muscle strength data and the sample sEMG signal; a muscle strength prediction module is used to collect the electromyographic signal of the subject at the muscle strength required to be predicted, and input the electromyographic signal of the subject at the muscle strength required to be predicted into the muscle strength prediction model to obtain the predicted muscle strength value of the subject.

[0022] The present application also provides a muscle strength recovery prediction device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the above-mentioned muscle strength recovery prediction method.

[0023] The present application proposes a muscle strength recovery prediction method, system and device, which adjust the initial number of interpolation values ​​by analyzing the numerical fluctuation of muscle strength data, and obtain the actual number of interpolation values ​​in each interpolation interval. This solves the problem that the common interpolation processing method only sets the sampling frequency difference between the muscle strength signal and the sEMG signal, and inserts a fixed number of data into the time-series adjacent muscle strength signals, resulting in low interpolation accuracy, thereby affecting the low accuracy of the subsequent muscle strength prediction model construction. It realizes the setting of different sampling accuracies when the muscle is in different states or the state changes, thereby improving the accuracy of interpolation.

[0024] In addition, the predicted value of each interpolation point in each interpolation interval is obtained, the adjustment factor of each interpolation point is constructed, and the predicted value of each interpolation point is corrected using the adjustment factor and the sample sEMG signal. This solves the problem that the common interpolation processing method only considers the relationship between the sEMG signal and the muscle force signal based on the numerical value of the muscle force signal, and also does not consider the influence of muscle fatigue when the human body is in a continuous measurement state, thereby resulting in poor interpolation accuracy. This realizes the adjustment of the predicted value of the muscle force signal to be interpolated based on the relationship between the sEMG signal and the muscle force signal, and further adjusts the predicted value according to the characteristics of the fatigue state, thereby ensuring that accurate prediction results are obtained in the subsequent muscle strength recovery prediction process. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 A flowchart of a method for predicting muscle strength recovery according to one embodiment of the present application;

[0026] Figure 2 is a structural block diagram of a muscle strength recovery prediction system according to one embodiment of the present application;

[0027] Figure 3 It is a schematic diagram of the structure of a muscle strength recovery prediction device according to an embodiment of the present application.

[0028] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0029] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0030] The present application provides a solution to set different sampling precisions when the muscle is in different states or when the state changes, so as to improve the accuracy of interpolation. At the same time, based on the relationship between the sEMG signal and the muscle strength signal, the predicted value of the muscle strength signal to be interpolated and obtained only based on the muscle strength signal is adjusted. The predicted value is further adjusted according to the characteristics of the fatigue state, so as to ensure that a more accurate prediction result is obtained in the subsequent muscle strength recovery prediction process.

[0031] The following is a detailed introduction to this application plan.

[0032] Figure 1 A flowchart of a method for predicting muscle strength recovery according to one embodiment of the present application, referring to Figure 1 The muscle strength recovery prediction method may be performed by a muscle strength recovery prediction device having data processing capability, and the muscle strength recovery prediction method may include:

[0033] Step S10, collecting sample muscle strength data and sample sEMG signals, and obtaining the total number of to-be-interpolated values ​​in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signals.

[0034] Among them, muscle force data is the mechanical force generated by the muscle. It can be measured and recorded by mechanical sensors such as dynamometers, isokinetic muscle testers, etc. The force generated during muscle contraction can directly reflect the intensity and force output of muscle contraction. sEMG signals are surface electromyographic signals, which can be recorded non-invasively by placing electrodes on the surface of the skin. sEMG signals can reflect the voltage changes of muscle electrical activity, that is, sEMG signals are a series of voltage waveforms that change over time, which can specifically reflect the frequency, amplitude and time domain characteristics of muscle activity.

[0035] In the specific implementation process, sEMG signals and muscle strength data of several patients are collected when they perform muscle strength rehabilitation training of prescribed movements. For example, any patient can be recorded as a target subject, and the muscle strength data and sEMG signals generated when the target subject performs muscle strength rehabilitation training of prescribed movements are used as sample muscle strength data and sample sEMG signals, and the sample muscle strength data and sample sEMG signals are used as examples for explanation.

[0036] For example, the specific steps of collecting sample muscle force data and sample sEMG signals include:

[0037] When the target subject performs muscle strength rehabilitation training, he / she should continuously exert force for 15 seconds according to the prescribed movements, rest for 5 seconds after the prescribed movements, and repeat the above operations. The number of repetitions can be set to 6. When the target subject performs the prescribed movements or rehabilitation training, a force sensor can be used to collect muscle strength data at the target muscle to obtain sample muscle strength data. At the same time, the electrodes are pasted on the target muscle surface of the target subject according to the direction of the muscle fibers to ensure good contact between the electrodes and the skin. Conductive paste or gel can be used to improve the signal quality, and the electrodes are used to collect sEMG signals at the target muscle of the target subject to obtain sample sEMG signals.

[0038] In one embodiment, after collecting the sample muscle strength data and the sample sEMG signal of the target subject, the sample sEMG signal may be subjected to denoising processing.

[0039] Among them, due to the influence of the acquisition environment, detection equipment, and muscle characteristics, the sample sEMG signal initially acquired contains interferences such as noise, artifacts, and baseline drift, and the sample sEMG signal needs to be denoised first to improve the quality and reliability of the sample sEMG signal. Exemplarily, a wavelet transform denoising algorithm can be used to denoise the sample sEMG signal.

[0040] In one embodiment, in step S10, obtaining the total number of to-be-interpolated values ​​in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signal may specifically include:

[0041] S11, determining the number of sampling points of the sample muscle force data based on the sampling frequency of the sample muscle force data, and determining the number of sampling points of the sample sEMG signal based on the sampling frequency of the sample sEMG signal;

[0042] S12. Record the difference between the number of sampling points of the sample sEMG signal and the number of sampling points of the sample muscle strength data as the total number of sample muscle strength data to be interpolated.

[0043] The number of sampling points is equal to the sampling frequency multiplied by the sampling time, and the sampling time is the duration of the target subject's muscle strength rehabilitation training.

[0044] Step S20, allocating the total number of interpolation values ​​according to the numerical fluctuation of the sample muscle strength data to obtain the actual number of interpolation values ​​in each interpolation interval, and dividing each interpolation interval based on the actual number of interpolation values ​​to obtain a number of interpolation points in each interpolation interval.

[0045] In one embodiment, in this step, the initial number of interpolation values ​​is adjusted according to the numerical fluctuation of the sample muscle strength data, and the actual number of interpolation values ​​for each interpolation interval is obtained, which may specifically include:

[0046] S21, determining the degree of fluctuation of the values ​​of each interval to be interpolated;

[0047] S22. Allocate the total number of values ​​to be interpolated according to the numerical differences of the numerical fluctuation degrees of the intervals to be interpolated, and determine the actual number of values ​​to be interpolated of each interval to be interpolated.

[0048] Among them, two data adjacent in time series in the sample muscle strength data are taken as an interval to be interpolated.

[0049] Specifically, Take the interpolation interval as an example and determine the The degree of fluctuation of the value in the interpolation interval The calculation formula is:

[0050]

[0051] In the formula, Indicates The first muscle strength data in the time series of the interval to be interpolated, Indicates The last muscle strength data in the time series of the interval to be interpolated, Represents the set of numerical fluctuations of all intervals to be interpolated in the sample muscle strength data. It means to find the absolute value. It means to find the maximum value in a set. It means to find the minimum value in a set.

[0052] It should be noted that when When it is 0, adjust The value of is 0.1. The numerical fluctuation degree of each interval to be interpolated is scaled in proportion so that the sum of the numerical fluctuation degrees of all intervals to be interpolated is 1, and then the scaled numerical fluctuation degree is multiplied by the total number of interpolated values ​​to be obtained to obtain the actual number of interpolated values ​​in each interval to be interpolated, ensuring that the sum of the actual number of interpolated values ​​in each interval to be interpolated is still the total number of interpolated values.

[0053] Specifically, Take the interpolation interval as an example and determine the The actual number of interpolation intervals to be interpolated The calculation formula is:

[0054]

[0055] In the formula, Indicates The degree of numerical fluctuation of the interpolation interval, represents the number of intervals to be interpolated. Indicates the total number of values ​​to be interpolated.

[0056] It should be noted that, since the degree of change of the first and last data of different intervals to be interpolated is different, the actual muscle states corresponding to different intervals to be interpolated are also different. For example, when the muscle state changes from exertion to relaxation or from relaxation to exertion, the muscle strength will fluctuate greatly. When the muscle is in the same state, that is, when the muscle is continuously in a state of exertion or continuous relaxation, the change in muscle strength is small. Based on this, in an embodiment of the present application, different numbers of interpolation values ​​are set for the interval to be interpolated at a state change and the interval to be interpolated at a state that remains unchanged, thereby improving the accuracy of the interpolation and making the interpolated sample muscle strength data more consistent with the actual muscle movement state.

[0057] Step S30, obtaining the predicted value of each interpolation point in each interpolation interval, constructing an adjustment factor for each interpolation point, and correcting the predicted value of each interpolation point using the adjustment factor and the sample sEMG signal to obtain corrected sample muscle strength data.

[0058] In the specific implementation process, the least squares method can be used to fit the sample muscle strength data to obtain the predicted value of each interpolation point in each interpolation interval.

[0059] In one embodiment, in step S30, constructing the adjustment factor of each to-be-interpolated point may specifically include:

[0060] S31, determining the myoelectric interval corresponding to each interval to be interpolated;

[0061] S32, determining the trend difference between each interpolation interval and the corresponding electromyographic interval;

[0062] S33: constructing an adjustment factor based on the trend difference degree of each interpolation interval and the sampling time of each interpolation point in each interpolation interval.

[0063] The myoelectric intervals corresponding to the sampling times of the intervals to be interpolated on the sample sEMG signal may be recorded as the myoelectric intervals corresponding to the intervals to be interpolated.

[0064] In the specific implementation process, firstly, according to the characteristic that the sample sEMG signal is generated before muscle contraction movement, a time error value is preset, and the sample muscle strength data is time-shifted based on the preset time error value, that is, the sampling time corresponding to each value on the sample muscle strength data is added with the preset time error value, and then the electromyographic intervals corresponding to the sampling times of the first data and the last data in the timing of each interpolation interval are obtained on the sample sEMG signal.

[0065] It should be noted that when the muscle is in a resting state, the muscle strength is small, and the sEMG signal amplitude is small at this time; when the muscle is in a state of continuous exertion, the muscle strength value continues to increase, and the sEMG signal amplitude increases and lasts for a period of time; when the muscle is in a resting state again, the muscle strength value gradually decreases, and the sEMG signal amplitude returns to near the origin; this shows that as the subjects perform muscle strength rehabilitation training according to the prescribed movements, the sample sEMG signal and the sample muscle strength data at the same muscle have similar change trends. Based on this, we can first obtain the degree of trend difference between the sample sEMG signal and the sample muscle strength data, and then adjust the pre-fitted values ​​in the sample muscle strength data according to the degree of trend difference.

[0066] In one embodiment, determining the trend difference between each to-be-interpolated interval and the corresponding electromyographic interval in step S32 may specifically include:

[0067] The DTW algorithm is used to determine the DTW distance between each interval to be interpolated and the corresponding electromyographic interval, and the determined DTW distance is recorded as the trend difference degree between each interval to be interpolated and the corresponding electromyographic interval.

[0068] The trend difference between each interval to be interpolated and the corresponding electromyographic interval is linearly normalized, and the sampling time of the interpolation points of all the intervals to be interpolated is linearly normalized.

[0069] It should be noted that linear normalization is performed on the values ​​in the set, that is, mapping the values ​​in the set to the range of (0, 1) according to their magnitude.

[0070] In one embodiment, in step S33, constructing the adjustment factor based on the trend difference degree of each to-be-interpolated interval and the sampling time of each to-be-interpolated point in each to-be-interpolated interval may specifically include:

[0071] The ratio of the trend difference degree of each interpolation point in each interpolation interval to the sampling time is recorded as the adjustment factor of each interpolation point.

[0072] It should be noted that the degree of trend difference of each interpolation point in each interpolation interval is the degree of trend difference of the interpolation interval where the interpolation point is located. The greater the degree of trend difference of the interpolation point, the more it indicates that there is a deviation in the predicted value of the interpolation point obtained by fitting. However, when the muscle is in a fatigue state, the degree of trend difference of each interpolation point in the interpolation interval where the fatigue state is located is also relatively large. Therefore, it is considered to add sampling time for adjustment. In this way, it can be ensured that the pre-fitted value in the sample muscle strength data is adjusted based on the sample sEMG signal while taking into account the influence of the muscle fatigue state.

[0073] In one embodiment, in step S30, using the adjustment factor and the sample sEMG signal to correct the predicted value of each interpolation point to obtain the corrected sample muscle strength data may specifically include:

[0074] The electromyographic signal corresponding to each interpolation point is obtained on the sample sEMG signal, and the Euclidean distance between each interpolation point and the corresponding electromyographic signal, and the Euclidean distance between the two endpoints of each interpolation interval and the electromyographic signal corresponding to each point are obtained in turn. According to the difference in the obtained Euclidean distances, combined with the sampling time difference between each interpolation point and the two endpoints of the interpolation interval, and the adjustment factor of each interpolation point, the predicted value of each interpolation point is corrected.

[0075] The sampling time of each interpolation point and its corresponding electromyographic signal is the same.

[0076] It should be noted that when the adjustment factor of the point to be interpolated is large, the lower the confidence of the predicted value of the point to be interpolated, the higher the necessity for adjustment. At the same time, since the predicted value of each point to be interpolated is obtained by fitting, if the point to be interpolated is closer to the endpoint of the interval to be interpolated in time series, it means that the confidence of the predicted value of the point to be interpolated is high, otherwise it is low. Since the two endpoints of the interval to be interpolated where each point to be interpolated is located are the muscle strength data actually collected, the fluctuation trends of the sample muscle strength data and the sample sEMG signal are similar, and then the distance between the two endpoints of each interval to be interpolated and the electromyographic signals corresponding to these two endpoints can be used as the adjustment standard for the predicted value of each point to be interpolated in each interval to be interpolated, combined with the sampling time difference between each point to be interpolated and the two endpoints of the interval to be interpolated, and the adjustment factor of each point to be interpolated, the predicted value of the point to be interpolated is adjusted.

[0077] For example, The first Points to be interpolated For example, remember The first muscle strength data in the time series in the interval to be interpolated is the target starting point, The last data bit in the time series in the interval to be interpolated As the target end point, get and sample sEMG signals with Euclidean distance of EMG signals with the same sampling time (That is and the Euclidean distance of the corresponding signal on the sample sEMG signal), obtain and sample sEMG signals with Euclidean distance of EMG signals with the same sampling time (That is and the Euclidean distance of the corresponding signal on the sample sEMG signal), obtain and sEMG signals Euclidean distance of EMG signals with the same sampling time (That is and the Euclidean distance of the corresponding signal on the sample sEMG signal), the following formula can be used to The first Points to be interpolated The predicted value is corrected:

[0078]

[0079] in, Indicates The first The predicted value after correction of the interpolation point, Indicates The first interval to be interpolated The predicted value of the interpolation point, Indicates The first interval to be interpolated The adjustment factor of the interpolation point, Indicates The first The sampling time of the interpolation points, Indicates The sampling time of the first data in the time series in the interval to be interpolated, Indicates The sampling time of the last bit of data in the time series in the interval to be interpolated.

[0080] It should be noted that the smaller the DTW distance, the higher the similarity of the two sequences. First, the predicted value of each interpolation point is corrected based on the trend difference degree of each interpolation interval. When the trend difference degree of the interpolation interval is smaller, it means that the change trends of the muscle force signal and the electromyographic signal in the interpolation interval are similar, and then the confidence of each interpolation prediction value in the interpolation interval is higher. On the contrary, when the trend difference degree of the interpolation interval is larger, it means that there is a difference in the change trend of the muscle force signal and the electromyographic signal in the interpolation interval, and then the confidence of each interpolation prediction value in the interpolation interval is lower. Accordingly, the sample sEMG signal needs to be used to adjust the prediction value. However, when the measurement time continues to increase and the muscle becomes fatigued, the sEMG signal will fluctuate abnormally. Specifically, in the early stage of fatigue, in order to maintain force output, the muscle will recruit more motor units, resulting in an increase in the amplitude of the sEMG signal. As fatigue further develops, due to the decline in muscle control ability, the amplitude of the sEMG signal may become larger, showing greater fluctuations. Therefore, when the muscle is in a fatigue state, the trend difference degree of the interpolation interval will also be greater. Based on this, in an embodiment of the present application, the adjustment factor of each interpolation point is first obtained by combining the trend difference degree and the length of the sampling time, that is, the necessary degree of correction of each interpolation point, and then the predicted value of each interpolation point is corrected in combination with the adjustment factor and the sample sEMG signal. In this way, the corrected predicted value of each interpolation point is closer to the actual muscle strength change than the predicted value obtained by fitting.

[0081] Step S40: establishing a muscle strength prediction model based on the corrected sample muscle strength data and the sample sEMG signal.

[0082] In a specific implementation, first, the sEMG signals and muscle strength data of several patients collected in step S10 during muscle strength rehabilitation training of prescribed movements are processed according to the processing method of the corrected sample muscle strength data and sample sEMG signals in the above embodiment. Secondly, the sEMG signals and muscle strength data are feature extracted and the appropriate features are selected using the Pearson correlation coefficient. Finally, a support vector machine (SVM) is used for model training to obtain a muscle strength prediction model. In this exemplary embodiment, the kernel function selected by the support vector machine can be a radial basis function (RBF).

[0083] The input of the muscle strength prediction model is the electromyographic signal at the point where the subject needs to predict the muscle strength, and the output is the predicted muscle strength value of the subject.

[0084] Step S50, collecting the electromyographic signal of the subject at the muscle strength point where the subject needs to predict the muscle strength, inputting the electromyographic signal of the subject at the muscle strength point where the subject needs to predict the muscle strength into the muscle strength prediction model, and obtaining the predicted muscle strength value of the subject.

[0085] In the specific implementation process, after the muscle strength prediction model is established, the electromyographic signal of the subject's muscle strength required to be predicted is input into the muscle strength prediction model, and the muscle strength prediction model outputs the subject's predicted muscle strength value, and then the muscle strength recovery status is judged according to the muscle strength evaluation method commonly used in rehabilitation medicine.

[0086] The muscle strength recovery prediction method provided by the present application adjusts the initial number of interpolation values ​​by analyzing the numerical fluctuation of muscle strength data to obtain the actual number of interpolation values ​​in each interpolation interval, thereby solving the problem that a common interpolation processing method only sets the sampling frequency difference between muscle strength signals and sEMG signals, inserts a fixed number of data into two muscle strength signals adjacent in time series, resulting in low interpolation accuracy and low accuracy of the constructed muscle strength prediction model. By obtaining the predicted value of each interpolation point in each interpolation interval, an adjustment factor of each interpolation point is constructed, and the predicted value of each interpolation point is corrected using the adjustment factor and the sample sEMG signal, thereby solving the problem that a common interpolation processing method only sets the sampling frequency difference between muscle strength signals and sEMG signals without considering the relationship between sEMG signals and muscle strength signals, and also does not consider the influence of muscle fatigue when the human body is in a continuous measurement state, resulting in poor interpolation accuracy.

[0087] Based on the above embodiments, Figure 2 FIG. 1 is a structural block diagram of a muscle strength recovery prediction system according to an embodiment of the present application, such as Figure 2 As shown, the muscle strength recovery prediction system may include: a data acquisition module 210, an interval division module 220, a value correction module 230, a model construction module 240, and a muscle strength prediction module 250, wherein:

[0088] The data acquisition module 210 is used to collect sample muscle strength data and sample sEMG signals, and obtain the total number of interpolation values ​​in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signals;

[0089] The interval division module 220 is used to allocate the total number of to-be-interpolated values ​​according to the numerical fluctuation of the sample muscle strength data, obtain the actual number of to-be-interpolated values ​​of each to-be-interpolated interval, and divide each to-be-interpolated interval based on the actual number of to-be-interpolated values ​​to obtain a number of to-be-interpolated points of each to-be-interpolated interval;

[0090] The numerical correction module 230 is used to obtain the predicted value of each to-be-interpolated point in each to-be-interpolated interval, construct an adjustment factor for each to-be-interpolated point, and use the adjustment factor and the sample sEMG signal to correct the predicted value of each to-be-interpolated point to obtain the corrected sample muscle strength data;

[0091] The model building module 240 is used to establish a muscle strength prediction model based on the corrected sample muscle strength data and the sample sEMG signal;

[0092] The muscle strength prediction module 250 is used to collect the electromyographic signal of the subject at the muscle strength required to be predicted, input the electromyographic signal of the subject at the muscle strength required to be predicted into the muscle strength prediction model, and obtain the predicted muscle strength value of the subject.

[0093] In an exemplary embodiment, the interval division module 220 may also be used to determine the degree of value fluctuation of each interval to be interpolated, and determine the actual number of values ​​to be interpolated of each interval to be interpolated according to the numerical difference of the degree of value fluctuation of the interval to be interpolated.

[0094] In an exemplary embodiment, the interval division module 220 may use the following formula (1) to determine the degree of fluctuation of the value of each to-be-interpolated interval:

[0095]

[0096] In formula (1), Indicates The degree of numerical fluctuation of the interpolation interval, Indicates The first muscle strength data in the time series of the interval to be interpolated, Indicates The last muscle strength data in the time series of the interval to be interpolated, Represents the set of numerical fluctuations of all intervals to be interpolated in the sample muscle strength data. It means to find the absolute value. It means to find the maximum value in a set. It means to find the minimum value in a set.

[0097] In an exemplary embodiment, the interval division module 220 may use the following formula (2) to determine the actual number of values ​​to be interpolated in each interval to be interpolated:

[0098]

[0099] In the formula, Indicates The actual number of interpolation intervals to be interpolated is Indicates The degree of numerical fluctuation of the interpolation interval, represents the number of intervals to be interpolated. Indicates the total number of values ​​to be interpolated.

[0100] In an exemplary embodiment, the numerical correction module 230 can also be used to determine the electromyographic interval corresponding to each interval to be interpolated; determine the degree of trend difference between each interval to be interpolated and the corresponding electromyographic interval; and construct an adjustment factor based on the degree of trend difference between each interval to be interpolated and the sampling time of each point to be interpolated in each interval to be interpolated.

[0101] In an exemplary embodiment, the numerical correction module 230 can also be used to correct the predicted value of each interpolation point using the adjustment factor and the sample sEMG signal using the following formula (3) to obtain the corrected sample muscle strength data:

[0102]

[0103] In the formula, Indicates The first The predicted value after correction of the interpolation point, Indicates The first interval to be interpolated The predicted value of the interpolation point, Indicates The first interval to be interpolated The adjustment factor of the interpolation point, Indicates The first The sampling time of the interpolation points, Indicates The sampling time of the first data in the time series in the interval to be interpolated, Indicates The sampling time of the last bit of data in the time series in the interval to be interpolated, Indicates The first The interpolation point and the sample sEMG signal are The first The Euclidean distance of the electromyographic signals of the interpolation points with the same sampling time, Indicates The first bit of data in the time sequence of the interpolation interval and the first bit of data in the sample sEMG signal The Euclidean distance of the electromyographic signals with the same sampling time of the first data in the time series in the interval to be interpolated, Indicates The first bit of data in the time sequence of the interpolation interval and the first bit of data in the sample sEMG signal The Euclidean distance of the electromyographic signals with the same sampling time of the last data in the time series in the interpolation interval.

[0104] In an exemplary embodiment, the muscle strength recovery prediction system may further include an acquisition module, where the acquisition module is used to determine the predicted muscle strength value of the subject using a muscle strength evaluation method to obtain the predicted muscle strength recovery status of the subject.

[0105] Those skilled in the art should understand that the division of the various modules in the embodiment is merely a division of logical functions, and in actual application, all or part of them can be integrated into one or more actual carriers, and these modules can be implemented entirely in the form of software called through a processing unit, or entirely in the form of hardware, or in the form of a combination of software and hardware. It should be noted that each module in a muscle strength recovery prediction system in this embodiment corresponds one-to-one to each step in a muscle strength recovery prediction method in the aforementioned embodiment. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned muscle strength recovery prediction method, which will not be repeated here.

[0106] Based on the above embodiments, Figure 3 FIG. 1 is a schematic diagram of a muscle strength recovery prediction device according to an embodiment of the present application. Figure 3 As shown, the electronic device may include: a processor (processor) 310, a communication interface (Communications Interface) 320, a memory (memory) 330 and a communication bus 340, wherein the processor 310, the communication interface 320 and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call the logic instructions in the memory 330 to execute a muscle strength recovery prediction method, which includes: collecting sample muscle strength data and sample sEMG signals, and obtaining the initial number of interpolation values ​​for each interpolation interval in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signal; adjusting the initial number of interpolation values ​​according to the numerical fluctuation of the sample muscle strength data to obtain the actual number of interpolation values ​​for each interpolation interval, and dividing each interpolation interval based on the actual number of interpolation values ​​to obtain A plurality of points to be interpolated in each interval to be interpolated; obtaining a predicted value of each point to be interpolated in each interval to be interpolated, constructing an adjustment factor for each point to be interpolated, and using the adjustment factor and the sample sEMG signal to correct the predicted value of each point to be interpolated, to obtain corrected sample muscle strength data; establishing a muscle strength prediction model based on the corrected sample muscle strength data and the sample sEMG signal; collecting the electromyographic signal at the muscle strength point where the subject needs to predict, inputting the electromyographic signal at the muscle strength point where the subject needs to predict into the muscle strength prediction model, to obtain the predicted muscle strength value of the subject.

[0107] In addition, the logic instructions in the above-mentioned memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0108] On the basis of the above embodiments, on the other hand, the present invention further provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the muscle strength recovery prediction method provided by the above methods, which includes: collecting sample muscle strength data and sample sEMG signals, and obtaining the initial number of interpolation intervals in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signals; adjusting the initial number of interpolation intervals according to the numerical fluctuation of the sample muscle strength data, and obtaining the number of interpolation intervals in each interval to be interpolated. The method comprises the steps of: obtaining an actual number of interpolated values ​​in the value interval, dividing each interpolated value interval based on the actual number of interpolated values, and obtaining a plurality of interpolated points in each interpolated value interval; obtaining a predicted value of each interpolated point in each interpolated value interval, constructing an adjustment factor for each interpolated point, and correcting the predicted value of each interpolated point using the adjustment factor and the sample sEMG signal to obtain corrected sample muscle strength data; establishing a muscle strength prediction model based on the corrected sample muscle strength data and the sample sEMG signal; collecting the electromyographic signal at the muscle strength location where the subject needs to predict, inputting the electromyographic signal at the muscle strength location where the subject needs to predict into the muscle strength prediction model, and obtaining the predicted muscle strength value of the subject.

[0109] On the basis of the above embodiments, on another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the processor executes the muscle strength recovery prediction method provided by the above methods, the method comprising: collecting sample muscle strength data and sample sEMG signals, obtaining the initial number of interpolation values ​​for each interpolation interval in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signals; adjusting the initial number of interpolation values ​​according to the numerical fluctuation of the sample muscle strength data to obtain the actual number of interpolation values ​​for each interpolation interval, and based on The actual number of values ​​to be interpolated is used to divide each interval to be interpolated, so as to obtain a number of points to be interpolated in each interval to be interpolated; the predicted value of each point to be interpolated in each interval to be interpolated is obtained, an adjustment factor of each point to be interpolated is constructed, and the predicted value of each point to be interpolated is corrected by using the adjustment factor and the sample sEMG signal to obtain corrected sample muscle strength data; a muscle strength prediction model is established based on the corrected sample muscle strength data and the sample sEMG signal; the electromyographic signal of the subject at the muscle strength required to be predicted is collected, and the electromyographic signal at the muscle strength required to be predicted of the subject is input into the muscle strength prediction model to obtain the predicted muscle strength value of the subject.

Claims

1. A method for predicting muscle strength recovery, It is characterized in that The muscle strength recovery prediction method comprises the following steps: Collecting sample muscle strength data and sample sEMG signals, and obtaining the total number of to-be-interpolated values ​​in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signals; The total number of to-be-interpolated values ​​is allocated according to the numerical fluctuation of the sample muscle strength data to obtain the actual number of to-be-interpolated values ​​in each to-be-interpolated interval, and each to-be-interpolated interval is divided based on the actual number of to-be-interpolated values ​​to obtain a number of to-be-interpolated points in each to-be-interpolated interval; Obtaining a predicted value of each to-be-interpolated point in each to-be-interpolated interval, constructing an adjustment factor for each to-be-interpolated point, and correcting the predicted value of each to-be-interpolated point using the adjustment factor and the sample sEMG signal to obtain corrected sample muscle strength data; Establishing a muscle strength prediction model based on the corrected sample muscle strength data and the sample sEMG signal; Collecting the electromyographic signal of the subject at the position where the muscle strength needs to be predicted, inputting the electromyographic signal of the subject at the position where the muscle strength needs to be predicted into the muscle strength prediction model, and obtaining the predicted muscle strength value of the subject; The method of allocating the total number of values ​​to be interpolated according to the numerical fluctuation of the sample muscle strength data to obtain the actual number of values ​​to be interpolated in each interpolation interval includes: Determine the degree of numerical fluctuation of each interval to be interpolated, and use the following formula (2) to determine the actual number of interpolated values ​​in each interval to be interpolated based on the numerical difference of the numerical fluctuation of the interval to be interpolated: In the formula, Indicates The actual number of interpolation intervals to be interpolated is Indicates The degree of numerical fluctuation of the interpolation interval, represents the number of intervals to be interpolated. Indicates the total number of values ​​to be interpolated; The method of using the adjustment factor and the sample sEMG signal to correct the predicted value of each interpolation point to obtain corrected sample muscle strength data includes: The adjustment factor and the sample sEMG signal are used to correct the predicted value of each interpolation point using the following formula (3) to obtain the corrected sample muscle strength data: In the formula, Indicates The first The predicted value after correction of the interpolation point, Indicates The first interval to be interpolated The predicted value of the interpolation point, Indicates The first interval to be interpolated The adjustment factor of the interpolation point, Indicates The first The sampling time of the interpolation points, Indicates The sampling time of the first data in the time series in the interval to be interpolated, Indicates The sampling time of the last bit of data in the time series in the interval to be interpolated, Indicates the sample muscle strength data The first The Euclidean distance between the interpolation points and the corresponding signal on the sample sEMG signal, Indicates the sample muscle strength data The Euclidean distance between the first data in the time series in the interpolation interval and the corresponding signal on the sample sEMG signal, Indicates the sample muscle strength data The Euclidean distance between the last bit of data in the time series in the interval to be interpolated and the corresponding signal on the sample sEMG signal.

2. The muscle strength recovery prediction method according to claim 1, It is characterized in that Determining the degree of fluctuation of the values ​​of each to-be-interpolated interval includes: The following formula (1) is used to determine the degree of numerical fluctuation of each interpolation interval: In the formula, Indicates The degree of numerical fluctuation of the interpolation interval, Indicates The first muscle strength data in the time series of the interval to be interpolated, Indicates The last muscle strength data in the time series of the interval to be interpolated, Represents the set of numerical fluctuations of all intervals to be interpolated in the sample muscle strength data. It means to find the absolute value. It means to find the maximum value in a set. It means to find the minimum value in a set.

3. The muscle strength recovery prediction method according to claim 1, It is characterized in that The step of constructing the adjustment factor for each point to be interpolated includes: Determine the myoelectric interval corresponding to each interval to be interpolated; Determine the trend difference between each interpolation interval and the corresponding electromyographic interval; The adjustment factor is constructed based on the trend difference degree of each interpolation interval and the sampling time of each interpolation point in each interpolation interval.

4. The muscle strength recovery prediction method according to claim 1, It is characterized in that After obtaining the predicted muscle strength value of the subject, the muscle strength recovery prediction method further includes: The muscle strength evaluation method is used to determine the predicted muscle strength value of the subject, and the predicted muscle strength recovery status of the subject is obtained.

5. A muscle strength recovery prediction system, It is characterized in that The muscle strength recovery prediction system comprises: A data acquisition module, used for acquiring sample muscle strength data and sample sEMG signals, and obtaining the total number of to-be-interpolated values ​​in the sample muscle strength data according to the sampling frequency of the sample muscle strength data and the sampling frequency of the sample sEMG signals; An interval division module is used to allocate the total number of to-be-interpolated values ​​according to the numerical fluctuation of the sample muscle strength data to obtain the actual number of to-be-interpolated values ​​of each to-be-interpolated interval, and divide each to-be-interpolated interval based on the actual number of to-be-interpolated values ​​to obtain a number of to-be-interpolated points of each to-be-interpolated interval; A numerical correction module, used for obtaining the predicted value of each to-be-interpolated point in each to-be-interpolated interval, constructing an adjustment factor for each to-be-interpolated point, and correcting the predicted value of each to-be-interpolated point using the adjustment factor and the sample sEMG signal to obtain corrected sample muscle strength data; A model building module, used for establishing a muscle strength prediction model based on the corrected sample muscle strength data and the sample sEMG signal; A muscle strength prediction module is used to collect the electromyographic signal of the subject at the muscle strength prediction location where the subject needs to predict the muscle strength, input the electromyographic signal of the subject at the muscle strength prediction location where the subject needs to predict the muscle strength into the muscle strength prediction model, and obtain the predicted muscle strength value of the subject; The interval division module is also used to determine the degree of fluctuation of the values ​​of each interval to be interpolated, and determine the actual number of values ​​to be interpolated in each interval to be interpolated using the following formula (2) according to the numerical difference of the degree of fluctuation of the values ​​of the interval to be interpolated: In the formula, Indicates The actual number of interpolation intervals to be interpolated is Indicates The degree of numerical fluctuation of the interpolation interval, represents the number of intervals to be interpolated. Indicates the total number of values ​​to be interpolated; The numerical correction module is also used to: use the adjustment factor and the sample sEMG signal to correct the predicted value of each interpolation point using the following formula (3) to obtain the corrected sample muscle strength data: In the formula, Indicates The first The predicted value after correction of the interpolation point, Indicates The first interval to be interpolated The predicted value of the interpolation point, Indicates The first interval to be interpolated The adjustment factor of the interpolation point, Indicates The first The sampling time of the interpolation points, Indicates The sampling time of the first data in the time series in the interval to be interpolated, Indicates The sampling time of the last bit of data in the time series in the interval to be interpolated, Indicates the sample muscle strength data The first The Euclidean distance between the interpolation points and the corresponding signal on the sample sEMG signal, Indicates the sample muscle strength data The Euclidean distance between the first data in the time series in the interpolation interval and the corresponding signal on the sample sEMG signal, Indicates the sample muscle strength data The Euclidean distance between the last bit of data in the time series in the interval to be interpolated and the corresponding signal on the sample sEMG signal.

6. A muscle strength recovery prediction device, It is characterized in that The muscle strength recovery prediction device comprises: At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the muscle strength recovery prediction method as described in any one of claims 1 to 4.

Citation Information

Patent Citations

  • High-precision wrist sphygmomanometer

    CN113017587A

  • Perception system for measuring human muscle work

    CN118383769A