A waist muscle stimulation method and system based on wearable device

Abstract

The present invention relates to the technical field of lumbar muscle stimulation, and in particular to a lumbar muscle stimulation method and system based on a wearable device, which specifically includes: constructing a local electromyographic signal sequence by performing frame processing on the lumbar electromyographic signal of a patient, decomposing it in combination with a short-time Fourier transform to construct a muscle stimulation index; constructing a muscle movement index by analyzing the change characteristics of the lumbar electromyographic signal of the patient, and combining the muscle stimulation index; constructing a movement continuity index by analyzing the change of the patient's movement state over time to reflect the fatigue degree of the patient's lumbar muscles; based on the movement continuity index of each signal frame, combining an LSTM neural network to classify the lumbar muscle state of the patient in each signal frame, and then directional stimulation is performed on the patient's lumbar muscles in a targeted manner. Thus, the lumbar muscle stimulation of the wearable device is realized, the accuracy of the lumbar muscle state classification result is improved, and it is helpful to relieve the patient's pain and make the patient recover faster.

Description

Technical Field

[0001] The present invention relates to the technical field of waist muscle stimulation, and in particular to a waist muscle stimulation method and system based on a wearable device. Background Art

[0002] For patients with lumbar injuries, athletes, sedentary office workers, etc., it is usually necessary to stimulate the lumbar muscles to gradually restore or maintain a healthy muscle state. The principle of traditional lumbar muscle stimulation methods is that the user's brain transmits bioelectric signals to muscle nerves. After the muscle nerves receive the bioelectric signals, they control the muscle contraction and complete a series of muscle movements. With the development of science and technology, electrical muscle stimulation technology (EMS) has gradually appeared in the market. Its principle is to simulate the brain to generate bioelectric signals, "cheat" the muscle nerves, and make the muscles produce "passive contraction", which also achieves the stimulation of the lumbar muscles.

[0003] Electromyographic signals are caused by the electrical activity of skeletal muscles, and most of them come from subjective will. That is, skeletal muscles will determine the discharge frequency and the number of mobilized muscle fibers according to the difficulty of the task. The discharge frequency of muscle fibers may be inconsistent in different states, and under different physiological states, multiple muscle fibers may need to be mobilized to complete the task, resulting in mutual interference between the discharge frequencies of different muscle fibers. Therefore, it is impossible to accurately determine whether the patient is in a state of exercise through electromyographic signals, and thus it is impossible to provide targeted directional stimulation to the patient, resulting in a longer recovery time for the patient.

[0004] In summary, the present invention collects surface electromyographic signals of waist muscles, decomposes the electromyographic signals through short-time Fourier transform, constructs a muscle stimulation index, and constructs a muscle movement index based on the analysis of the changing characteristics of the electromyographic signals under the motion state. The motion continuity index is constructed by analyzing the temporal changes of the muscle movement index. As the input of the LSTM neural network, it can more accurately classify whether the patient is in a state of motion, and then perform targeted stimulation on the patient's waist muscles. Summary of the invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a waist muscle stimulation method and system based on a wearable device, and the technical solutions adopted are as follows:

[0006] In a first aspect, an embodiment of the present invention provides a waist muscle stimulation method based on a wearable device, the method comprising the following steps:

[0007] Collect electromyographic signals from the surface of waist muscles;

[0008] The electromyographic signal is processed into frames to obtain electromyographic signals of each frame as each signal frame; the local electromyographic signal sequence of each signal frame is obtained according to the time sequence of each signal frame; the amplitude and frequency of each frequency domain component of each local electromyographic signal sequence are obtained by short-time Fourier transform; the amplitude-frequency change rate between adjacent frequency domain components of each local electromyographic signal sequence is obtained according to the amplitude and frequency of each frequency domain component; the muscle stimulation index of each signal frame is obtained according to the amplitude-frequency change rate between adjacent frequency domain components; the local electromyographic signal stimulation index sequence of each signal frame is obtained according to the muscle stimulation index of each signal frame; the muscle movement index of each signal frame is obtained according to the local electromyographic signal stimulation index sequence; the movement continuity index of each signal frame is obtained according to the muscle movement index of each signal frame and adjacent signal frames;

[0009] The patient's waist muscle status is obtained based on the motion continuity index combined with the LSTM neural network. Different waist wearable devices are processed differently according to different waist muscle status.

[0010] Preferably, the local electromyographic signal sequence of each signal frame is obtained according to the time sequence of each signal frame, specifically:

[0011] The sequence consisting of the t signal frame and the m signal frames before and after it arranged in time order is taken as the local electromyographic signal sequence of the t signal frame.

[0012] Preferably, the step of obtaining the amplitude-frequency change rate between adjacent frequency domain components of each local electromyographic signal sequence according to the amplitude and frequency of each frequency domain component specifically includes:

[0013] In each local electromyographic signal sequence, calculate the absolute value of the frequency difference and the absolute value of the amplitude difference between the ath and a+1th frequency domain components; calculate the ratio of the absolute value of the frequency difference to the frequency of the ath frequency domain component; calculate the ratio of the absolute value of the amplitude difference to the amplitude of the ath frequency domain component; calculate the average of the frequency ratio and the amplitude ratio; and use the average as the amplitude-frequency change rate between the ath and a+1th frequency domain components.

[0014] Preferably, obtaining the muscle stimulation index of each signal frame according to the amplitude-frequency change rate between adjacent frequency domain components specifically includes:

[0015] In each local electromyographic signal sequence, calculate the calculation result of an exponential function with a natural constant as the base and the inverse of the amplitude-frequency change rate between adjacent frequency domain components as the exponent; calculate the reciprocal of the sum of the calculation result and 1; calculate the sum of all the reciprocals; and use the sum as the muscle stimulation index of the signal frame corresponding to the local electromyographic signal sequence.

[0016] Preferably, obtaining the local electromyographic signal stimulation index sequence of each signal frame according to the muscle stimulation index of each signal frame specifically includes:

[0017] The sequence composed of muscle stimulation indexes of all signal frames in the local electromyographic signal sequence of each signal frame is used as the local electromyographic signal stimulation index sequence of each signal frame.

[0018] Preferably, obtaining the muscle movement index of each signal frame according to each local electromyographic signal stimulation index sequence specifically includes:

[0019] Calculate the sum of the reciprocals of the squares of the second-order differences of all elements in the local electromyographic signal stimulation index sequence; calculate the product of the sum of the reciprocals and the Hurst index of the local electrical signal stimulation index sequence; and use the product as the muscle movement index of the signal frame corresponding to the local electrical signal stimulation index sequence.

[0020] Preferably, obtaining the motion continuity index of each signal frame according to the muscle motion index of each signal frame and adjacent signal frames specifically includes:

[0021] A sequence consisting of the t signal frame and its first m signal frames in ascending time order is used as a previous adjacent frame sequence of the t signal frame. In the previous adjacent sequence of the t signal frame, the ratio of the sequence number of the c-th signal frame to the number of elements in the previous adjacent sequence is calculated, which is recorded as the first ratio; the absolute value of the difference between the muscle movement index of the c-th signal frame and the c+1-th signal frame is calculated; the ratio of the muscle movement index of the c-th signal frame to the absolute value of the difference is calculated, which is recorded as the second ratio; the product of the first ratio and the second ratio is calculated; the sum of all the products in the previous adjacent frame sequence is calculated; and the sum is used as the motion continuity index of the t signal frame.

[0022] Preferably, the patient's waist muscle state is obtained according to the motion continuity index combined with the LSTM neural network, specifically:

[0023] The motion continuity index of each signal frame and the amplitude and frequency of each frequency domain component of the local electromyographic signal sequence are used as the input of the LSTM neural network to classify the patient's lumbar muscle status. The output of the neural network is the classification result of the lumbar muscle status of each signal frame, and the classification results include motion state and non-motion state.

[0024] Preferably, the different processing of the waist wearable device according to different waist muscle states is specifically as follows:

[0025] When the waist muscles are in motion, the waist wearable device does not perform electrical stimulation; when the waist muscles are not in motion, the waist wearable device performs electrical stimulation.

[0026] In a second aspect, an embodiment of the present invention also provides a waist muscle stimulation system based on a wearable device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of any one of the above methods when executing the computer program.

[0027] The embodiments of the present invention have at least the following beneficial effects:

[0028] The present invention processes the patient's waist electromyographic signal by frame processing, constructs a local electromyographic signal sequence, takes the local electromyographic signal sequence as the input of short-time Fourier transform, decomposes the local electromyographic signal sequence to obtain various frequency domain components, and constructs a muscle stimulation index by analyzing the amplitude and frequency changes of each frequency domain component, reflecting the degree of stimulation of the patient's waist cells; by analyzing the change characteristics of the electromyographic signal when the patient's waist is painful and when in motion, a muscle movement index is constructed in combination with the muscle stimulation index to reflect the possibility that the patient is in a state of motion; by analyzing the change of the patient's motion state over time, a motion continuity index is constructed to reflect the fatigue degree of the patient's waist muscles; based on the motion continuity index of each signal frame, it is used as the input of an LSTM neural network, and the actual motion state of the patient's waist muscles in each signal frame is more accurately classified, and then the patient's waist muscles are stimulated in a targeted manner, which helps to relieve the patient's pain and make the patient recover faster. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0030] Figure 1 A flowchart of a waist muscle stimulation method based on a wearable device provided in one embodiment of the present invention;

[0031] Figure 2 Schematic diagram of the steps of the waist muscle stimulation method using a wearable device. DETAILED DESCRIPTION

[0032] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, characteristics and effects of a waist muscle stimulation method and system based on a wearable device proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0033] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0034] The following is a detailed description of a method and system for stimulating waist muscles based on a wearable device provided by the present invention in conjunction with the accompanying drawings.

[0035] See also Figure 1 , which shows a flowchart of a waist muscle stimulation method based on a wearable device provided by an embodiment of the present invention, the method comprising the following steps:

[0036] Step S001, collecting surface electromyographic signals of waist muscles.

[0037] Since there are many types of waist muscle stimulation devices and people who wear waist muscle stimulation devices, this embodiment takes the waist injury patient wearing waist electrical stimulator as an example for analysis. All movements and postures of the human body are controlled by the nervous system. The movement consciousness generated by the brain is transmitted to the corresponding muscles through nerve signals, and the human body movement is completed by controlling the specific muscle contraction. The electromyographic signal is the electrophysiological signal generated by the muscle contraction. Based on this, this embodiment collects the electromyographic signal on the surface of the waist muscle through the EMG sensor in the waist electrical stimulator. The collection frequency is recorded as f, and the value in this embodiment is 1000Hz. The collection time is recorded as T, and the empirical value of 100s is taken in this embodiment. It should be noted that the values ​​of f and T can be set by the implementer, and this embodiment does not make specific restrictions.

[0038] Since the EMG signals of waist muscles are easily disturbed by the environment during the collection process, there may be noise in the EMG signals obtained, which need to be denoised. In order to retain the shape of the original signal as much as possible, this embodiment denoises the collected EMG signals of waist muscles through a 20Hz to 500Hz Butterworth bandpass filter to eliminate environmental noise. The Butterworth bandpass filter denoising is a well-known technology, and the specific process is not repeated in this embodiment.

[0039] Step S002, decomposing the electromyographic signal by short-time Fourier transform to construct a muscle stimulation index, based on which the changing characteristics of the electromyographic signal under motion state are analyzed to construct a muscle movement index, and by analyzing the temporal changes of the muscle movement index to construct a movement continuity index.

[0040] 1. Decompose the electromyographic signal and construct a muscle stimulation index. When the patient has waist pain or is in motion, the waist muscle cells are stimulated to a high degree and the muscle cells contract to a large extent, causing the cells to change their potential and generate bioelectricity. At this time, the amplitude of the electromyographic signal collected by the EMG sensor is large. Based on this, this embodiment constructs a muscle stimulation index to reflect the exercise intensity of the human body at each time point. The construction process is as follows:

[0041] The electromyographic signals of the waist surface muscles obtained through the above steps are divided into continuous non-overlapping frames with 20 ms as one frame of electromyographic signals. The electromyographic signal data obtained are taken as signal frames, and the t-th signal frame is taken as the center. The t-th signal frame is taken before and after the t-th signal frame according to the time sequence of data acquisition. signal frames, and the obtained The sequence composed of the signal frame and the t-th signal frame is recorded as the local electromyographic signal sequence of the t-th signal frame, and the length of the local electromyographic signal sequence is In this embodiment The size of takes the empirical value of 10, takes the local electromyographic signal sequence of the t-th signal frame as the input of the short-time Fourier transform, and obtains the amplitude, frequency, etc. of the local electromyographic signal sequence in each frequency domain component, wherein the short-time Fourier transform is a well-known technology, and the specific process is not repeated in this embodiment. Based on this, the muscle stimulation index can be constructed, and its calculation formula is as follows:

[0042] ;

[0043] ;

[0044] in It represents the amplitude-frequency change rate of the ath frequency domain component and the a+1th frequency domain component after the local electromyographic signal sequence of the tth signal frame is decomposed by short-time Fourier transform, , They respectively represent the amplitudes of the ath and a+1th frequency domain components of the local electromyographic signal sequence of the tth signal frame after short-time Fourier transform decomposition, , They represent the frequencies of the ath and a+1th frequency domain components of the local electromyographic signal sequence of the tth signal frame after short-time Fourier transform decomposition, represents the muscle stimulation index of the t-th signal frame, It represents the number of frequency domain components after the local electromyographic signal sequence of the t-th signal frame is decomposed by short-time Fourier transform. Represents an exponential function with the natural constant e as the base.

[0045] The larger the number of frequency domain components, The larger the value is, the more complex the changes in the myoelectric signals of the waist muscles are, and the more high-frequency details they have. That is, the more likely the waist cells of the patient are in a stimulated state at this time, and the greater the difference in amplitude and frequency of adjacent frequency domain components, that is, , The larger the value is, the greater the difference between each frequency domain component and the main frequency domain component. At this time, the more unstable the electromyographic signal of the waist muscle is, that is, the patient's waist muscle cells are in an unstable active state, and the greater the degree of stimulation may be. At the same time, since the electromyographic signal is relatively weak, its amplitude is usually between a few microvolts and tens of microvolts, and the gradient value of the Sigmoid function near 0 is large, accordingly, this embodiment uses the Sigmoid function to process the amplitude-frequency change rate to obtain the enhanced amplitude-frequency change rate, that is, , can more sensitively identify the changes in the amplitude and frequency of the adjacent frequency domain components of the electromyographic signal, so the larger the muscle stimulation index calculated at this time; the smaller the number of frequency domain components, that is, The smaller it is, the simpler the change of the myoelectric signal of the waist muscle is, that is, the more likely the waist cells of the patient are in an unstimulated state, and the smaller the amplitude of each frequency domain component is, , The smaller it is, the more stable the frequency domain components are relative to the main frequency domain components. At this time, the more stable the electromyographic signals of the lumbar muscles are, the less stimulated the patient's lumbar cells are, and therefore the smaller the calculated muscle stimulation index is.

[0046] 2. Construct a muscle movement index based on the muscle stimulation index. The muscle stimulation index obtained through the above steps can only reflect the degree of stimulation of the patient's waist muscles. When the patient feels pain or the patient is in motion, the waist muscles can be stimulated. At this time, the calculated muscle stimulation index is large, which cannot accurately reflect whether the patient is in motion. When the patient is in motion, different muscle tissues coordinate with each other, so that the frequency and amplitude of the electromyographic signal show regular changes, and the patient has not fully recovered, and usually will not and cannot make fast and explosive movements, such as short-distance sprints, etc., usually slower movements, such as jogging, etc. At this time, the changes in the waist electromyographic signal are relatively smooth, and the patient's waist is in pain. The patient cannot independently control the contraction of the waist muscles, that is, the frequency and amplitude of the electromyographic signal show an irregular change pattern. Based on this, this embodiment constructs a muscle movement index based on the muscle stimulation index to reflect the possibility that the patient is in a painful state, and its construction process is as follows:

[0047] The sequence composed of the muscle stimulation index of each signal frame in the local electromyographic signal sequence of the t-th signal frame is recorded as the local electromyographic signal stimulation index sequence, and the sequence length is , the local electromyographic signal stimulation index sequence of the t-th signal frame is used as the input for the Hurst index calculation, and the Hurst index of the sequence is obtained, which is recorded as the Hurst index of the t-th signal frame, wherein the Hurst index is a well-known technology, and the specific process is not repeated here. Based on this, the muscle movement index can be calculated, and its calculation formula is as follows:

[0048] ;

[0049] in represents the muscle movement index of the t-th signal frame, represents the Hurst exponent of the t-th signal frame, represents the number of nearest neighbor signal frames selected before and after the tth signal frame in time order. In this embodiment, The size of takes the empirical value of 10, , , They represent the values ​​of the bth, b+1th, and b+2th elements in the local electromyographic signal stimulation index sequence of the tth signal frame. is the second-order difference of the b-th element in the sequence.

[0050] When the patient is in motion, the frequency and amplitude of the EMG signals of the patient's waist muscles show regular changes, and the changes of the EMG signals of the waist are relatively smooth. Therefore, in the local EMG signal stimulation index sequence of the t-th signal frame, if the second-order difference of each element is smaller, that is, The smaller it is, the smoother the change of muscle stimulation index in time is, and the larger the Hurst index calculated by the local electromyographic signal stimulation index sequence of the tth signal frame is, that is, The larger the value is, the stronger the periodicity of the sequence is, that is, the stronger the regularity of the electromyographic signal is, and the more likely the patient is in motion at this time, so the larger the calculated muscle movement index is.

[0051] 3. Construct a motion continuity index based on the muscle movement index. When the patient's waist muscles are in motion for a long time, it will cause muscle fatigue, soreness, pain, and there is a risk of muscle damage. At this time, in order to enable the patient to recover better, there is no need to perform external electrical stimulation on the patient's waist muscles through a waist electrical stimulator; when the patient's waist muscles are in a resting state or a painful state, in order to enable the patient to recover as soon as possible, it is necessary to perform a certain degree of external electrical stimulation on the patient's waist muscles through a waist electrical stimulator to increase cell activity and promote cell metabolism. The electrical stimulation of the nerves can activate the nervous system to a certain extent and produce analgesic effects, which helps to relieve pain. The sequence of the tth signal frame and its first m signal frames in ascending time order is used as the previous adjacent frame sequence of the tth signal frame. Based on this, this embodiment constructs a motion continuity index based on the muscle movement index, and its calculation formula is as follows:

[0052] ;

[0053] in represents the motion continuity index of the t-th signal frame, represents the number of nearest neighbor signal frames selected before the tth signal frame in chronological order. In this embodiment, The empirical value of is 10, c is the sequence number of each signal frame in the adjacent frame sequence before the tth signal frame, , represents the muscle movement index of the cth signal frame and the c+1th signal frame in the previous adjacent frame sequence. is the first ratio, is the second ratio.

[0054] The closer the t-th signal frame is to the c-th signal frame in the previous adjacent frame sequence in time, that is, The larger the value is, the greater the correlation between the motion state of the tth signal frame and the cth signal frame in the previous adjacent frame sequence. At the same time, if each signal frame is more likely to be in motion, and the difference in motion between each signal frame and its adjacent signal frames is smaller, that is, The bigger, The smaller it is, the more it indicates that the patient is in a state of continuous motion at this time, so the calculated motion continuity index is larger; the farther the t-th signal frame is in time distance from the c-th signal frame in the previous adjacent frame sequence, that is, The smaller it is, the smaller the correlation between the motion state of the tth signal frame and the cth signal frame in the previous adjacent frame sequence is. At the same time, if each signal frame is more likely to be in a non-motion state, and the difference in the motion degree between each signal frame and its adjacent signal frames is greater, that is, The smaller, The larger the value is, the more it indicates that the patient is in a state of non-continuous movement, so the calculated movement continuity index is smaller.

[0055] Step S003, using the motion continuity index of each signal frame as the input of the LSTM neural network to more accurately classify whether the patient is in a motion state, and then perform targeted stimulation on the patient's waist muscles.

[0056] The motion continuity index of each signal frame obtained by the above steps, for some signal frames ahead in time, because there is no signal frame data long enough in front of it, the motion continuity index of this part of the signal frame cannot be calculated and obtained normally, so in this embodiment, the obtained plate structure factor is obtained by linear interpolation to obtain the motion continuity index of this part of the signal frame, wherein linear interpolation is a known technology, and the specific process is not repeated. The motion continuity index can reflect the degree of continuity of the patient in the state of motion at each time point. The larger the motion continuity index, the more it indicates that the patient is in a state of continuous motion, that is, the patient stimulates the waist muscle cells through active exercise, promotes muscle relaxation and contraction, and helps the body to recover. At this time, there is no need for the outside world to passively stimulate the waist muscles; the smaller the motion continuity index, the more it indicates that the patient is in a non-motion state, that is, the patient does not stimulate the waist muscle cells by active exercise. In order to make the patient recover as soon as possible, it is necessary for the outside world to passively stimulate the waist muscles, thereby increasing the cell activity, promoting cell metabolism, and helping the patient's body to recover as soon as possible. Based on the above analysis, the motion continuity index, amplitude, and frequency of each signal frame can be used as the input of the LSTM long short-term memory network, with Adam as the optimization algorithm and cross entropy as the loss function. The output of the network is the classification result of whether each signal frame in the electromyographic signal is in a motion state. The LSTM long short-term memory network is a well-known technology, and the specific process will not be repeated here. When the classification result is in a motion state, the patient is not passively electrically stimulated. When the classification result is a non-motion state, the patient's waist muscles are passively directional stimulated by a waist electrical stimulator, thereby helping the patient relieve waist pain, promoting local blood flow, and delivering oxygen and nutrients to the waist muscles, which is beneficial to restoring damaged muscle and tissue function, thereby enabling the patient to recover faster. The schematic diagram of the steps of the above method is shown in the figure. Figure 2 shown.

[0057] Based on the same inventive concept as the above method, an embodiment of the present invention also provides a waist muscle stimulation system based on a wearable device, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-mentioned waist muscle stimulation methods based on a wearable device.

[0058] In summary, the embodiment of the present invention provides a waist muscle stimulation method based on a wearable device, which processes the patient's waist electromyographic signal by framing and constructing a local electromyographic signal sequence, which is used as the input of the short-time Fourier transform, and decomposes the local electromyographic signal sequence to obtain each frequency domain component, and constructs a muscle stimulation index by analyzing the amplitude and frequency changes of each frequency domain component to reflect the degree of stimulation of the patient's waist cells; by analyzing the change characteristics of the electromyographic signal when the patient has waist pain and when in motion, a muscle movement index is constructed in combination with the muscle stimulation index to reflect the possibility that the patient is in a state of motion; by analyzing the temporal changes of the patient's movement state, a movement continuity index is constructed to reflect the fatigue degree of the patient's waist muscles; based on the movement continuity index of each signal frame, it is used as the input of the LSTM neural network to more accurately classify the actual movement state of the patient's waist muscles in each signal frame, and then the patient's waist muscles are stimulated in a targeted manner, which helps to relieve the patient's pain and make the patient recover faster.

[0059] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0060] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0061] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A waist muscle stimulation method based on a wearable device, characterized in that: The method comprises the following steps: Collect electromyographic signals from the surface of waist muscles; The electromyographic signal is processed into frames to obtain electromyographic signals of each frame as each signal frame; the local electromyographic signal sequence of each signal frame is obtained according to the time sequence of each signal frame; the amplitude and frequency of each frequency domain component of each local electromyographic signal sequence are obtained by short-time Fourier transform; the amplitude-frequency change rate between adjacent frequency domain components of each local electromyographic signal sequence is obtained according to the amplitude and frequency of each frequency domain component; the muscle stimulation index of each signal frame is obtained according to the amplitude-frequency change rate between adjacent frequency domain components; the local electromyographic signal stimulation index sequence of each signal frame is obtained according to the muscle stimulation index of each signal frame; the muscle movement index of each signal frame is obtained according to the local electromyographic signal stimulation index sequence; the movement continuity index of each signal frame is obtained according to the muscle movement index of each signal frame and adjacent signal frames; The patient's waist muscle status is obtained based on the motion continuity index combined with the LSTM neural network; different waist wearable devices are processed differently according to different waist muscle status.

2. A waist muscle stimulation method based on a wearable device as claimed in claim 1, characterized in that: The local electromyographic signal sequence of each signal frame is obtained according to the time sequence of each signal frame, specifically: The sequence consisting of the t signal frame and the m signal frames before and after it arranged in time order is taken as the local electromyographic signal sequence of the t signal frame.

3. A waist muscle stimulation method based on a wearable device as claimed in claim 1, characterized in that: The step of obtaining the amplitude-frequency change rate between adjacent frequency domain components of each local electromyographic signal sequence according to the amplitude and frequency of each frequency domain component specifically includes: In each local electromyographic signal sequence, calculate the absolute value of the frequency difference and the absolute value of the amplitude difference between the ath and a+1th frequency domain components; calculate the ratio of the absolute value of the frequency difference to the frequency of the ath frequency domain component; calculate the ratio of the absolute value of the amplitude difference to the amplitude of the ath frequency domain component; calculate the average of the frequency ratio and the amplitude ratio; and use the average as the amplitude-frequency change rate between the ath and a+1th frequency domain components.

4. A waist muscle stimulation method based on a wearable device as claimed in claim 1, characterized in that: The step of obtaining the muscle stimulation index of each signal frame according to the amplitude-frequency change rate between adjacent frequency domain components specifically includes: In each local electromyographic signal sequence, calculate the calculation result of an exponential function with a natural constant as the base and the inverse of the amplitude-frequency change rate between adjacent frequency domain components as the exponent; calculate the reciprocal of the sum of the calculation result and 1; calculate the sum of all the reciprocals; and use the sum as the muscle stimulation index of the signal frame corresponding to the local electromyographic signal sequence.

5. A waist muscle stimulation method based on a wearable device as claimed in claim 1, characterized in that: The method of obtaining a local electromyographic signal stimulation index sequence of each signal frame according to the muscle stimulation index of each signal frame specifically includes: The sequence composed of muscle stimulation indexes of all signal frames in the local electromyographic signal sequence of each signal frame is used as the local electromyographic signal stimulation index sequence of each signal frame.

6. A waist muscle stimulation method based on a wearable device as claimed in claim 1, characterized in that: The step of obtaining the muscle movement index of each signal frame according to each local electromyographic signal stimulation index sequence specifically includes: Calculate the sum of the reciprocals of the squares of the second-order differences of all elements in the local electromyographic signal stimulation index sequence; calculate the product of the sum of the reciprocals and the Hurst index of the local electrical signal stimulation index sequence; and use the product as the muscle movement index of the signal frame corresponding to the local electrical signal stimulation index sequence.

7. A waist muscle stimulation method based on a wearable device as claimed in claim 1, characterized in that: The step of obtaining the motion continuity index of each signal frame according to the muscle motion index of each signal frame and adjacent signal frames specifically includes: A sequence consisting of the t signal frame and its first m signal frames in ascending time order is used as a previous adjacent frame sequence of the t signal frame. In the previous adjacent sequence of the t signal frame, the ratio of the sequence number of the c-th signal frame to the number of elements in the previous adjacent sequence is calculated, which is recorded as the first ratio; the absolute value of the difference between the muscle movement index of the c-th signal frame and the c+1-th signal frame is calculated; the ratio of the muscle movement index of the c-th signal frame to the absolute value of the difference is calculated, which is recorded as the second ratio; the product of the first ratio and the second ratio is calculated; the sum of all the products in the previous adjacent frame sequence is calculated; and the sum is used as the motion continuity index of the t signal frame.

8. A waist muscle stimulation method based on a wearable device as claimed in claim 1, characterized in that: The patient's waist muscle state is obtained according to the movement continuity index combined with the LSTM neural network, specifically: The motion continuity index of each signal frame and the amplitude and frequency of each frequency domain component of the local electromyographic signal sequence are used as the input of the LSTM neural network to classify the patient's lumbar muscle status. The output of the neural network is the classification result of the lumbar muscle status of each signal frame, and the classification results include motion state and non-motion state.

9. A waist muscle stimulation method based on a wearable device as claimed in claim 1, characterized in that: The different processing of the waist wearable device according to different waist muscle states is specifically as follows: When the waist muscles are in motion, the waist wearable device does not perform electrical stimulation; when the waist muscles are not in motion, the waist wearable device performs electrical stimulation.

10. A waist muscle stimulation system based on a wearable device, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images