Closed-loop electrical stimulation control system based on multi-modal vibrations and model prediction

Abstract

The application discloses a closed-loop electric stimulation control system based on multi-modal vibration and model prediction, comprising: initializing the system; obtaining personalized parameters and stimulation parameters and determining a basic electric stimulation intensity in combination with a preset function; collecting high-frequency vibration signals and low-frequency vibration signals in real time, performing signal preprocessing, obtaining piezoelectric film signals and MEMS accelerometer signals and analyzing to determine corresponding muscle movement intensity characteristics; weighting and fusing the muscle movement intensity characteristics to obtain a basic muscle movement intensity; predicting and solving the basic muscle movement intensity to obtain a predicted muscle movement intensity; determining an optimal electric stimulation intensity based on the predicted muscle movement intensity, the basic electric stimulation intensity and the basic muscle movement intensity; applying stimulation to a preset position and repeatedly updating the optimal electric stimulation intensity until the system ends running. The application dynamically optimizes the electric stimulation intensity through multi-modal vibration signal analysis and model prediction, and realizes personalized closed-loop electric stimulation control.

Description

Technical Field

[0001] This application relates to the field of electrical stimulation therapy equipment technology, and more specifically, to a closed-loop electrical stimulation control system based on multimodal vibration and model prediction. Background Technology

[0002] Most existing electrical stimulation therapy devices adopt an open-loop control mode, requiring manual preset and adjustment of the output intensity. This makes it impossible to dynamically adapt to the real-time muscle movement state, which can easily lead to problems such as excessive stimulation intensity causing muscle fatigue and worsening of spasms, or insufficient intensity.

[0003] Existing related patents (such as CN108553757A and US20050283204A1) mostly use a single vibration / accelerometer to collect muscle signals, and the control algorithm only uses proportional-integral adjustment or simple threshold judgment, which has the following defects: the single sensor signal is easily affected by environmental noise and cannot accurately quantify muscle movement intensity; the control algorithm has a lag response and cannot predict changes in muscle state, resulting in low adjustment accuracy; and it only adjusts the electrical stimulation intensity, which cannot adapt to the needs of different muscle contraction stages. In order to overcome the shortcomings of the above-mentioned prior art, this invention improves the accuracy of muscle movement intensity recognition by fusing multimodal vibration signals and uses a model predictive control algorithm to achieve advance prediction and precise adjustment of electrical stimulation output parameters, forming a substantial difference from existing patents and meeting the requirements for invention patent authorization. Summary of the Invention

[0004] In view of the above problems, the purpose of this invention is to provide a closed-loop electrical stimulation control system based on multimodal vibration and model prediction. Through multimodal vibration signal analysis and model prediction, the electrical stimulation intensity is dynamically optimized to achieve personalized closed-loop electrical stimulation control.

[0005] The first aspect of this invention provides a closed-loop electrical stimulation control system based on multimodal vibration and model prediction, comprising: The system initialization module is used to initialize the system. The human-computer interaction module is used to acquire personalized parameters and stimulation parameters and combine them with preset functions to determine the basic electrical stimulation intensity. The signal acquisition and preprocessing module is used to acquire high-frequency vibration signals and low-frequency vibration signals in real time and perform signal preprocessing to obtain piezoelectric film signals and MEMS accelerometer signals; the piezoelectric film signals and MEMS accelerometer signals are analyzed respectively to obtain multiple muscle movement intensity characteristics; The model prediction module is used to normalize and weightedly fuse the multiple muscle movement intensity features to obtain the basic muscle movement intensity; it is used to determine the prediction model using a preset method and combine it with the basic muscle movement intensity to perform prediction and solve to obtain the predicted muscle movement intensity. The optimal electrical stimulation solution module is used to determine the optimal electrical stimulation intensity based on the predicted muscle movement intensity, the baseline electrical stimulation intensity, and the baseline muscle movement intensity. An electrical stimulation output module is used to apply stimulation to a preset position based on the optimal electrical stimulation intensity and the stimulation parameters; The parameter update module is used to update the optimal electrical stimulation intensity and repeat the stimulation until the system finishes running.

[0006] In this solution, the initialization of the system includes: The system initialization includes resetting computational parameters, calibrating sensors, and initializing output parameters. The initialization output parameters include at least setting the electrical stimulation intensity of the stimulation electric field to a preset minimum electrical stimulation intensity.

[0007] In this solution, the step of obtaining personalized parameters and stimulation parameters and determining the basic electrical stimulation intensity in combination with preset functions includes: The personalized parameters include gender, age, and body type; The preset functions include muscle strength training, spasm relief, and neuromuscular rehabilitation. If the preset function is muscle strength training, then the electrical stimulation intensity range is the first preset intensity range; If the preset function is spasm relief, then the electrical stimulation intensity range is the second preset intensity range; If the preset function is neuromuscular rehabilitation, then the electrical stimulation intensity range is the third preset intensity range; The basic electrical stimulation intensity is determined according to the preset standards, based on the personalized parameters and the range of electrical stimulation intensity. The stimulation parameters include: fundamental frequency, difference frequency, difference frequency variation period, and stimulation time.

[0008] In this scheme, the real-time acquisition of high-frequency vibration signals and low-frequency vibration signals, followed by signal preprocessing to obtain piezoelectric thin film signals and MEMS accelerometer signals, includes: The high-frequency vibration signal and the low-frequency vibration signal are filtered by a first preset filter to obtain the filtered high-frequency vibration signal and the filtered low-frequency vibration signal. A preset noise reduction algorithm is used to perform noise reduction processing on the filtered high-frequency vibration signal and the filtered low-frequency vibration signal to obtain the piezoelectric film signal and the MEMS accelerometer signal.

[0009] In this scheme, the piezoelectric film signal and the MEMS accelerometer signal are analyzed respectively to obtain multiple muscle movement intensity characteristics, including: The muscle movement intensity characteristics include the effective value of the piezoelectric thin film signal, the effective value of the MEMS accelerometer signal, the peak value of the piezoelectric thin film signal, the peak value of the MEMS accelerometer signal, the centroid of the piezoelectric thin film signal frequency, and the centroid of the MEMS accelerometer signal frequency. Extract the corresponding RMS values ​​of the piezoelectric thin film signal and the MEMS accelerometer signal based on the piezoelectric thin film signal and the MEMS accelerometer signal, respectively: ; ; in, For piezoelectric film signals, RMS1 is the effective value of the piezoelectric thin film signal, RMS2 is the effective value of the MEMS accelerometer signal, and T is the preset signal sampling period. Extract the corresponding peak values ​​of the piezoelectric thin film signal and the MEMS accelerometer signal based on the piezoelectric thin film signal and the MEMS accelerometer signal, respectively: , t∈[0,T]; , t∈[0,T]; Where Peak1 is the peak value of the piezoelectric thin film signal, and Peak2 is the peak value of the MEMS accelerometer signal; Extract the corresponding centroids of the piezoelectric thin film signal frequency and the MEMS accelerometer signal frequency, respectively, based on the piezoelectric thin film signal and the MEMS accelerometer signal: ; ; Wherein, FC1 is the centroid of the piezoelectric thin film signal frequency, and FC2 is the centroid of the MEMS accelerometer signal frequency. The Fourier transform result of the piezoelectric thin film signal is shown. The result is the Fourier transform of the MEMS accelerometer signal, where [f1,f2] is the frequency range of the piezoelectric thin film signal and [f3,f4] is the frequency range of the MEMS accelerometer signal.

[0010] In this scheme, the normalization and weighted fusion of the multiple muscle movement intensity features to obtain the basic muscle movement intensity includes: Each muscle movement intensity feature is normalized to obtain multiple normalized muscle movement intensity features; The normalized multiple muscle movement intensity features are then weighted and fused: ; Among them, I m(t) represents the baseline muscle activity intensity, where α1, α2, α3, α4, α5, and α6 are preset feature weights, and RMS is used. g1 The normalized effective value of the piezoelectric film signal, RMS g2 Peak is the normalized RMS value of the MEMS accelerometer signal. g1 Peak is the normalized peak value of the piezoelectric thin film signal. g2 The normalized peak value of the MEMS accelerometer signal, FC g1 The centroid of the normalized piezoelectric thin film signal frequency, FC g2 The centroid of the normalized MEMS accelerometer signal frequency.

[0011] In this scheme, the step of determining the prediction model using a preset method and combining it with the basic muscle movement intensity to perform prediction and solution, thereby obtaining the predicted muscle movement intensity, includes: The prediction model is determined using a pre-defined method; The personalized correction coefficient is determined based on the personalized parameters; Based on the baseline muscle activity intensity and the prediction model, the predicted muscle activity intensity is determined as follows: ; Among them, I m (t|t-1) represents the predicted muscle activity intensity, where a, b, c, and d are preset model parameters, and K is a personalized correction coefficient; m (t-1) represents the baseline muscle activity intensity in the previous second, I m (t-2) represents the basic muscle movement intensity in the first two seconds, and u(t-1) represents the electrical stimulation intensity of the electric field in the first second.

[0012] In this scheme, determining the optimal electrical stimulation intensity based on the predicted muscle movement intensity, the baseline electrical stimulation intensity, and the baseline muscle movement intensity includes: The optimal electrical stimulation intensity is determined based on the predicted muscle activity intensity, the baseline electrical stimulation intensity, and the baseline muscle activity intensity. ; in, For optimal electrical stimulation intensity, Based on the basic electrical stimulation intensity, Basic muscle activity intensity, To predict muscle activity intensity.

[0013] In this scheme, applying stimulation to the preset location based on the optimal electrical stimulation intensity and the stimulation parameters includes: The optimal electrical stimulation intensity is taken as the electrical stimulation intensity of the stimulation electric field; The stimulation electric field is output based on the electrical stimulation intensity of the stimulation electric field and the stimulation parameters; The stimulation electric field includes: a composite wave interference electric field, a low-frequency modulated intermediate frequency electric field, and a low-frequency electric field.

[0014] In this scheme, updating the optimal electrical stimulation intensity and repeatedly applying stimulation until the system operation ends includes: Relevant data are repeatedly collected and calculated at preset time intervals to update the optimal electrical stimulation intensity; Stimulation was repeatedly applied based on the updated optimal electrical stimulation intensity; The system terminates operation when the runtime equals the stimulation time.

[0015] This invention discloses a closed-loop electrical stimulation control system based on multimodal vibration and model prediction, comprising: initializing the system; acquiring personalized parameters and stimulation parameters and determining the basic electrical stimulation intensity in combination with preset functions; acquiring high-frequency and low-frequency vibration signals in real time, performing signal preprocessing to obtain piezoelectric film signals and MEMS accelerometer signals and analyzing them to determine the corresponding muscle movement intensity characteristics; weighted fusion of the muscle movement intensity characteristics to obtain the basic muscle movement intensity; predicting and solving the basic muscle movement intensity to obtain the predicted muscle movement intensity; determining the optimal electrical stimulation intensity based on the predicted muscle movement intensity, the basic electrical stimulation intensity, and the basic muscle movement intensity; applying stimulation to preset positions and repeatedly updating the optimal electrical stimulation intensity until the system terminates operation. This invention dynamically optimizes the electrical stimulation intensity through multimodal vibration signal analysis and model prediction, achieving personalized closed-loop electrical stimulation control. Attached Figure Description

[0016] Figure 1 A block diagram of the closed-loop electrical stimulation control system based on multimodal vibration and model prediction provided by the present invention is shown. Figure 2 A flowchart of the closed-loop electrical stimulation control method based on multimodal vibration and model prediction provided by the present invention is shown. Detailed Implementation

[0017] To better understand the above-mentioned objectives, features, and advantages of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0018] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and therefore the scope of protection of the invention is not limited to the specific embodiments disclosed below.

[0019] Figure 1 A block diagram of the closed-loop electrical stimulation control system based on multimodal vibration and model prediction provided by the present invention is shown.

[0020] like Figure 1 As shown, this invention discloses a closed-loop electrical stimulation control system based on multimodal vibration and model prediction, comprising: The system initialization module is used to initialize the system. The human-computer interaction module is used to acquire personalized parameters and stimulation parameters and combine them with preset functions to determine the basic electrical stimulation intensity. The signal acquisition and preprocessing module is used to acquire high-frequency vibration signals and low-frequency vibration signals in real time and perform signal preprocessing to obtain piezoelectric film signals and MEMS accelerometer signals; the piezoelectric film signals and MEMS accelerometer signals are analyzed respectively to obtain multiple muscle movement intensity characteristics; The model prediction module is used to normalize and weightedly fuse multiple muscle movement intensity features to obtain the basic muscle movement intensity; it is used to determine the prediction model using a preset method and combine it with the basic muscle movement intensity to perform prediction and solve the problem to obtain the predicted muscle movement intensity. The optimal electrical stimulation solution module is used to determine the optimal electrical stimulation intensity based on the predicted muscle movement intensity, the baseline electrical stimulation intensity, and the baseline muscle movement intensity. An electrical stimulation output module is used to apply stimulation to a preset location based on optimal electrical stimulation intensity and stimulation parameters. The parameter update module is used to update the optimal electrical stimulation intensity and repeat the stimulation until the system finishes running.

[0021] According to an embodiment of the present invention, system initialization is first performed, including sensor calibration, resetting operational parameters, and initializing output parameters. The electrical stimulation intensity of the stimulation electric field is set to a preset minimum electrical stimulation intensity. Personalized parameters and preset functions are acquired, and the corresponding basic electrical stimulation intensity and stimulation parameters are determined according to different preset functions. The stimulation parameters include: fundamental frequency, difference frequency, difference frequency variation period, and running time. Based on the current electrical stimulation intensity and stimulation parameters, electrical stimulation is output and high-frequency vibration signals and low-frequency vibration signals are acquired. The high-frequency vibration signals and low-frequency vibration signals are filtered and noise-reduced to obtain piezoelectric film signals and MEMS accelerometer signals. The piezoelectric film signals and MEMS accelerometer signals are analyzed to obtain multiple muscle movement intensity features. The obtained multiple muscle movement intensity features are normalized and weighted fused to obtain... The baseline muscle activity intensity is used to solve for the subsequent prediction of muscle activity intensity. A prediction model is determined using a preset method, and the predicted muscle activity intensity is determined based on the prediction model and the baseline muscle activity intensity. The difference between the predicted muscle activity intensity and the baseline electrical stimulation intensity is calculated and substituted into the objective function to obtain the initial electrical stimulation intensity. The initial electrical stimulation intensity is adjusted according to personalized parameters to obtain the optimal electrical stimulation intensity. The optimal electrical stimulation intensity is used as the electrical stimulation intensity of the stimulation electric field, and the corresponding stimulation electric field is output based on the stimulation parameters. Relevant data is repeatedly collected and processed at preset intervals to update the optimal electrical stimulation intensity. Stimulation is applied based on the updated parameters until the running time equals the stimulation time, at which point the system terminates, realizing closed-loop adaptive adjustment of the system's electrical stimulation intensity. The flowchart of the closed-loop electrical stimulation control method based on multimodal vibration and model prediction is shown below. Figure 2 As shown.

[0022] According to an embodiment of the present invention, initializing the system includes: System initialization includes resetting operational parameters, calibrating sensors, and initializing output parameters; The initialization output parameters should include setting the electrical stimulation intensity of the stimulation electric field to the preset minimum electrical stimulation intensity.

[0023] It should be noted that resetting the computational parameters specifically involves clearing and resetting all computational parameters required by the system for weighted fusion of individual muscle movement intensity, solving for predicted muscle movement intensity, and calculating optimal electrical stimulation intensity. This avoids residual computational parameters from the previous run interfering with the current closed-loop control process. Sensor calibration specifically involves performing hardware path self-checks on the sensors used in the system to confirm that the signal acquisition channels are connected normally, and then calibrating and correcting the sensor signal acquisition sensitivity and output linearity. Initializing output parameters also includes initializing the stimulation parameters corresponding to the output electric field to ensure the safety, standardization, and controllability of system startup and initial stimulation.

[0024] The preset minimum electrical stimulation intensity can be set by those skilled in the art according to the actual situation.

[0025] According to an embodiment of the present invention, obtaining personalized parameters and stimulation parameters and determining the basic electrical stimulation intensity in combination with preset functions includes: Personalized parameters include gender, age, and body type; The preset functions include muscle strength training, spasticity relief, and neuromuscular rehabilitation; If the preset function is muscle strength training, the electrical stimulation intensity range is the first preset intensity range; If the preset function is spasm relief, the electrical stimulation intensity range is the second preset intensity range; If the preset function is neuromuscular rehabilitation, the electrical stimulation intensity range is the third preset intensity range; The basic electrical stimulation intensity is determined according to preset standards, based on personalized parameters and the range of electrical stimulation intensity. The stimulation parameters include: fundamental frequency, difference frequency, difference frequency variation period, and stimulation time.

[0026] It should be noted that before collecting vibration information after obtaining the stimulation parameters, the system first outputs the corresponding stimulation electric field based on the current electrical stimulation intensity, i.e., the preset minimum electrical stimulation intensity, combined with the stimulation parameters. The purpose is to use the minimum electrical stimulation intensity as the initial output benchmark before real-time detection of muscle state and solution of optimal electrical stimulation intensity, and to start vibration from the lowest safe level. This ensures that there is a stable initial input reference in the first closed-loop cycle while ensuring safety, providing a reliable benchmark for subsequent feature extraction, model prediction, and intensity optimization calculation, and avoiding the algorithm's inability to converge and prediction distortion due to the lack of initial output.

[0027] Personalized parameters, stimulation parameters, and preset functions are all obtained through user input. Personalized parameters include gender, age, and body type (divided by body mass index: underweight ≤18.5, normal 18.6-23.9, overweight ≥24.0). These three types of parameters directly affect muscle endurance, contraction characteristics, and electrical stimulation sensitivity, providing a personalized basis for subsequent prediction model solutions and the calculation of optimal electrical stimulation intensity. The system has built-in preset reasonable value ranges for different stimulation parameters, only acquiring values ​​within the preset reasonable value ranges. Stimulation parameters exceeding the preset reasonable value ranges are prompted to the user to re-enter, ensuring the safety of stimulation.

[0028] The first preset intensity range is 6-8, the second preset intensity range is 2-4, and the third preset intensity range is 4-6. The preset standards are tables of fine-tuning ranges for different genders, ages, and body types. Based on the personalized parameters of this run, the tables are looked up and the corresponding electrical stimulation intensity ranges are fine-tuned. For example, if the preset function is muscle strength training, the gender is male, the first preset intensity range is uniformly increased by 0.5; if the age is 51, the first preset intensity range is uniformly decreased by 0.5; and if the body type is overweight, the first preset intensity range is uniformly increased by 0.3. The final first preset intensity range is 6.3-8.3. The median value of the first preset intensity range is selected as the base electrical stimulation intensity. This standardizes, quantifies, and programmably executes the rules for determining the base electrical stimulation intensity, providing reasonable and individualized benchmark target values ​​for subsequent model prediction, optimal electrical stimulation intensity solution, and closed-loop dynamic adjustment, thereby improving the accuracy, safety, and adaptability of electrical stimulation control.

[0029] The preset standard, the first preset strength range, the second preset strength range, and the third preset strength range can be specifically set by those skilled in the art according to the actual situation.

[0030] According to an embodiment of the present invention, high-frequency vibration signals and low-frequency vibration signals are acquired in real time and preprocessed to obtain piezoelectric thin film signals and MEMS accelerometer signals, including: The high-frequency vibration signal and the low-frequency vibration signal are filtered by the first preset filter to obtain the filtered high-frequency vibration signal and the filtered low-frequency vibration signal. A preset noise reduction algorithm is used to denoise the filtered high-frequency vibration signal and the filtered low-frequency vibration signal to obtain the piezoelectric film signal and the MEMS accelerometer signal.

[0031] It should be noted that the high-frequency vibration signal range is 10-100Hz, acquired using a piezoelectric film vibration sensor, while the low-frequency vibration signal range is 0.5-10Hz, acquired using a miniature MEMS accelerometer. A multimodal sensor array is composed of a piezoelectric film vibration sensor (model: PVDF-LT100) and a miniature MEMS accelerometer (model: ADXL345) and attached to a preset position. The two sensors need to be spaced 5-8mm apart. The two signals complement each other, improving signal recognition accuracy and anti-interference ability. This differs from existing single-sensor solutions that cover the vibration range of the entire muscle movement cycle, ensuring the comprehensiveness of the acquired signals. During signal acquisition, the sampling frequency is set to 1000Hz, the sampling accuracy is 12 bits, and the input impedance is ≥1MΩ to ensure the authenticity and integrity of the acquired signals and avoid signal distortion.

[0032] The first preset filter can be a Butterworth bandpass filter, which filters the high-frequency vibration signal and the low-frequency vibration signal respectively. Specifically: ; ; Where G1(s) is the transfer function of the Butterworth bandpass filter corresponding to the high-frequency vibration signal, G2(s) is the transfer function of the Butterworth bandpass filter corresponding to the low-frequency vibration signal, s is a Laplace complex variable, and w z1 w is the filter cutoff angular frequency corresponding to the high-frequency vibration signal. z2 ω is the cutoff angular frequency of the filter corresponding to the high-frequency vibration signal, n is the filter order, and j is the number of filter segments.

[0033] In this invention, n can be selected as n=4 order to balance filtering effect and computational efficiency, and to avoid computational lag caused by excessively high order. In this invention, j can be selected as j=2 to ensure smooth filtering curve and reduce signal distortion.

[0034] The filter cutoff angular frequency corresponding to the high-frequency vibration signal passes through the high-frequency cutoff frequency f. z1 Calculations show that the filter cutoff angular frequency corresponding to the low-frequency vibration signal passes through the low-frequency cutoff frequency f. z2 The calculation yielded: ; ; Among them, the high-frequency cutoff frequency f z1 The low-frequency cutoff frequency is 100Hz. z2 It is 0.5Hz.

[0035] The filtered high-frequency vibration signal and the filtered low-frequency vibration signal are obtained based on the transfer function of the corresponding Butterworth bandpass filter: ; ; Where, x h1 (t) represents the filtered high-frequency vibration signal, x h2 (t) represents the filtered low-frequency vibration signal, L -1 x1(s) is the inverse Laplace transform of x1(t), and x2(s) is the Laplace transform of x2(t).

[0036] The preset noise reduction algorithm can use wavelet threshold noise reduction algorithm to remove residual noise (such as electromagnetic interference noise) in the filtered high-frequency cutoff signal and the filtered low-frequency cutoff signal, so as to further improve the signal purity and obtain the piezoelectric film signal and MEMS accelerometer signal.

[0037] The filter order and the number of filter segments can be adjusted by those skilled in the art according to the actual situation.

[0038] The first preset filter, the preset noise reduction algorithm, and the preset position can be specifically set by those skilled in the art according to the actual situation.

[0039] According to embodiments of the present invention, the piezoelectric thin film signal and the MEMS accelerometer signal are analyzed respectively to obtain multiple muscle movement intensity characteristics, including: Muscle movement intensity characteristics include the effective value of the piezoelectric film signal, the effective value of the MEMS accelerometer signal, the peak value of the piezoelectric film signal, the peak value of the MEMS accelerometer signal, the centroid of the piezoelectric film signal frequency, and the centroid of the MEMS accelerometer signal frequency. Extract the RMS values ​​of the piezoelectric thin film signal and the MEMS accelerometer signal respectively based on the piezoelectric thin film signal and the MEMS accelerometer signal: ; ; in, For piezoelectric film signals, RMS1 is the effective value of the piezoelectric thin film signal, RMS2 is the effective value of the MEMS accelerometer signal, and T is the preset signal sampling period. Extract the corresponding peak values ​​of the piezoelectric thin film signal and the MEMS accelerometer signal based on the piezoelectric thin film signal and the MEMS accelerometer signal, respectively: , t∈[0,T]; , t∈[0,T]; Where Peak1 is the peak value of the piezoelectric thin film signal, and Peak2 is the peak value of the MEMS accelerometer signal; Extract the corresponding piezoelectric thin film signal frequency centroid and MEMS accelerometer signal frequency centroid based on the piezoelectric thin film signal and MEMS accelerometer signal, respectively: ; ; Wherein, FC1 is the RMS value of the piezoelectric thin film signal, and FC2 is the RMS value of the MEMS accelerometer signal. The Fourier transform result of the piezoelectric thin film signal is shown. The result is the Fourier transform of the MEMS accelerometer signal, where [f1,f2] is the frequency range of the piezoelectric thin film signal and [f3,f4] is the frequency range of the MEMS accelerometer signal.

[0040] It should be noted that max{} represents selecting the maximum value; the frequency range of the piezoelectric film signal is [10Hz, 100Hz], and the frequency range of the MEMS accelerometer signal is [0.5Hz, 10Hz], which correspond to the frequency ranges of the high-frequency vibration signal and the low-frequency vibration signal, respectively; the number of Fourier transform points for the piezoelectric film signal and the accelerometer signal is set to 1024 to ensure frequency resolution.

[0041] The preset signal sampling period can be 1 second, meaning that an effective value is calculated every 1 second. The effective values ​​of the two signals represent the average energy of the vibration signal, reflecting the average intensity of muscle movement. Extracting the effective values ​​ensures real-time performance. The signal peak value represents the maximum contraction intensity of the muscle. The peak value can be extracted through a peak detection algorithm, which captures the maximum absolute value of the two signals within 1 second in real time. The signal frequency centroid represents the dominant frequency of muscle vibration. It is implemented through an integral operation module with a calculation accuracy of ≤0.1Hz, reflecting the muscle contraction speed. The larger the frequency centroid, the faster the muscle contraction speed and the higher the exercise intensity. By simultaneously extracting the three dimensions of signal effective value, signal peak value, and signal frequency centroid, the muscle movement state can be comprehensively represented from three levels: average energy, instantaneous maximum contraction, and dominant vibration frequency. This provides comprehensive, refined, and quantifiable underlying feature support for subsequent feature weighted fusion, construction of basic muscle movement intensity indicators, and model predictive control, improving the accuracy and individualized adaptation capability of closed-loop electrical stimulation modulation.

[0042] The preset signal sampling period can be set by those skilled in the art according to the actual situation.

[0043] According to an embodiment of the present invention, multiple muscle movement intensity features are normalized and weighted and fused to obtain the basic muscle movement intensity, including: Each muscle movement intensity feature is normalized to obtain multiple normalized muscle movement intensity features; Weighted fusion of multiple normalized muscle movement intensity features: ; Among them, I m (t) represents the baseline muscle activity intensity, where α1, α2, α3, α4, α5, and α6 are preset feature weights, and RMS is used. g1 The normalized effective value of the piezoelectric film signal, RMS g2 Peak is the normalized RMS value of the MEMS accelerometer signal. g1 Peak is the normalized peak value of the piezoelectric thin film signal. g2 The normalized peak value of the MEMS accelerometer signal, FC g1 The centroid of the normalized piezoelectric thin film signal frequency, FC g2The centroid of the normalized MEMS accelerometer signal frequency.

[0044] It should be noted that range normalization can be used to normalize each muscle movement intensity feature. Range normalization is an existing method and will not be elaborated further. The sum of α1, α2, α3, α4, α5, and α6 is 1, specifically set as α1=0.25, α2=0.25, α3=0.15, α4=0.15, α5=0.10, and α6=0.10. The calculated I... m The value range of (t) is mapped to [0,10]. The larger the value, the higher the muscle movement intensity (0 is complete relaxation, 10 is maximum contraction), ensuring the accuracy of muscle movement intensity quantification. Multiple muscle movement intensity features are weighted and fused to avoid the one-sidedness caused by a single feature dominating the evaluation. By configuring fixed weight values, multi-feature fusion calculation can be completed in a standardized and procedural manner, reducing random error interference. At the same time, the basic muscle movement intensity is limited to the quantification range of 0 to 10, and the strength of muscle movement is intuitively represented by the numerical value. 0 corresponds to the muscle's complete relaxation state, and 10 corresponds to the muscle's maximum contraction state, realizing continuous quantitative grading of muscle movement intensity. This provides a standardized quantitative basis for subsequent objective function solving and dynamic adjustment of optimal electrical stimulation intensity, effectively improving the standardization, accuracy, and feasibility of closed-loop control in muscle movement intensity assessment.

[0045] The preset feature weights can be set by those skilled in the art according to the actual situation.

[0046] According to an embodiment of the present invention, a prediction model is determined using a preset method and combined with the basic muscle activity intensity for prediction solution, to obtain the predicted muscle activity intensity, including: The prediction model is determined using a pre-defined method; Determine the personalized correction coefficient based on the personalized parameters; Predicted muscle activity intensity is determined based on baseline muscle activity intensity and a prediction model: ; Among them, I m (t|t-1) represents the predicted muscle activity intensity, where a, b, c, and d are preset model parameters, and K is a personalized correction coefficient; m (t-1) represents the baseline muscle activity intensity in the previous second, I m (t-2) represents the basic muscle movement intensity in the first two seconds, and u(t-1) represents the electrical stimulation intensity of the electric field in the first second.

[0047] It should be noted that the preset method can be a model prediction and control algorithm, which integrates the patient's personalized parameters, the muscle movement information of the previous second, and preset functions to achieve a dual improvement in prediction accuracy and adaptability. The model prediction and control algorithm is existing technology and will not be explained further. The personalized correction coefficient is determined based on the personalized parameters. Specifically, the corresponding correction coefficient is preset according to the acquired personalized parameters and automatically calculated during system operation. For example, if the gender is male, p=1.1; if the age is over 51 years old, p=0.9; if the body type is overweight, p=1.05. The average of the three correction coefficients is calculated as the corresponding personalized correction coefficient. This allows the prediction model to take into account both the temporal dynamics of muscle movement and individual physiological differences, ensuring that the correction process is fast, efficient, and can be executed in a standardized manner.

[0048] The initial values ​​of the preset model parameters a, b, c, and d can be set as follows: a = 0.65, b = 0.25, c = 0.05, and d = 0.05. Those skilled in the art can dynamically adjust the specific values ​​of the preset model parameters a, b, c, and d according to actual needs.

[0049] When the baseline muscle movement intensity of the previous two seconds is missing, or when both the baseline muscle movement intensity of the previous second and the baseline muscle movement intensity of the previous two seconds are missing, the missing baseline muscle movement intensity is set to the preset minimum baseline muscle movement intensity. If the electrical stimulation intensity of the electric field of the previous second is missing, the electrical stimulation intensity of the electric field of the previous second is set to the preset minimum electrical stimulation intensity. In the case of scenarios where the baseline muscle movement intensity or electrical stimulation intensity data may be missing in the early stage of system startup, the rule of filling the missing initial values ​​with the preset minimum baseline muscle movement intensity and the preset minimum electrical stimulation intensity can avoid the problem that the model cannot start the operation due to missing data, and ensure the continuity and stability of the closed-loop control process in the system initialization phase.

[0050] All calculations in this step are dimensionless.

[0051] The preset model parameters, preset minimum basic muscle exercise intensity, and preset method can be specifically set by those skilled in the art according to the actual situation.

[0052] According to embodiments of the present invention, determining the optimal electrical stimulation intensity based on predicted muscle movement intensity, baseline electrical stimulation intensity, and baseline muscle movement intensity includes: The optimal electrical stimulation intensity is determined based on predicted muscle activity intensity, baseline electrical stimulation intensity, and baseline muscle activity intensity. ; in, For optimal electrical stimulation intensity, Based on the basic electrical stimulation intensity, Basic muscle activity intensity, To predict muscle activity intensity.

[0053] It should be noted that by predicting muscle movement intensity, the optimal electrical stimulation intensity is determined from the baseline electrical stimulation intensity and the baseline muscle movement intensity. This constructs an adaptive control logic that combines advanced prediction and real-time feedback, effectively eliminating the physiological lag problem of pure feedback control. It dynamically adapts to the fatigue and state changes of muscles at different contraction stages, achieving precise and stable adjustment of electrical stimulation intensity and improving the accuracy and robustness of stimulation control.

[0054] When determining the optimal electrical stimulation intensity, the constraints can be adjusted in conjunction with personalized parameters. For example, if the age in the personalized parameters is 51 years or older, the maximum change between the current optimal electrical stimulation intensity and the electrical stimulation intensity of the previous stimulation will be controlled within a preset range (e.g., 0-0.3mA) to improve stimulation safety.

[0055] All calculations in this step are dimensionless.

[0056] The preset range of variation can be set by those skilled in the art according to the actual situation.

[0057] According to an embodiment of the present invention, applying stimulation to a preset location based on optimal electrical stimulation intensity and stimulation parameters includes: The optimal electrical stimulation intensity is taken as the electrical stimulation intensity of the stimulation electric field; The electric stimulation field is output based on the intensity of the electric stimulation and the stimulation parameters. The stimulating electric fields include: composite wave interference electric field, low-frequency modulated intermediate frequency electric field, and low-frequency electric field.

[0058] It should be noted that using the optimal electrical stimulation intensity obtained from the closed-loop solution directly as the output benchmark of the stimulation electric field enables precise matching of the electrical stimulation output with the real-time muscle movement state, achieving dynamic tracking and control. By combining the preset stimulation parameters to generate electric field signals, it is ensured that the output electric field meets personalized configuration requirements in terms of intensity, frequency, and period timing. Through the coordinated output of multiple fields, including composite wave interference electric field, low-frequency modulated mid-frequency electric field, and low-frequency electric field, it is possible to provide combined stimulation to muscles and nerves from different dimensions of action, adapting to different functional needs such as muscle strength training, spasm relief, and neuromuscular rehabilitation.

[0059] According to an embodiment of the present invention, updating the optimal electrical stimulation intensity and repeatedly applying stimulation until the system operation ends includes: Relevant data are repeatedly collected and calculated at preset time intervals to update the optimal electrical stimulation intensity; Stimulation was repeatedly applied based on the updated optimal electrical stimulation intensity; The system terminates when the runtime equals the stimulation time.

[0060] It should be noted that the relevant data includes high-frequency and low-frequency vibration signals. The latest high-frequency and low-frequency vibration signals after the last stimulation are re-acquired and processed and calculated again to obtain a new electrical stimulation intensity, which is then used as the updated optimal electrical stimulation intensity. Stimulation is applied based on this updated optimal electrical stimulation intensity, achieving a closed-loop cycle. The system completes one cycle every 0.5 seconds. Each prediction integrates the baseline muscle movement intensity of the previous second with preset personalized parameters, dynamically correcting the prediction model and optimal electrical stimulation parameters to ensure that the prediction accuracy dynamically adapts to muscle movement state and individual differences. The entire closed-loop design not only ensures that the treatment process is implemented according to the preset duration but also significantly improves the real-time performance, accuracy, and individual adaptability of electrical stimulation control through high-frequency iteration and dynamic correction mechanisms, further enhancing the stability of the stimulation effect and the safety of use.

[0061] The preset time interval can be set by those skilled in the art according to the actual situation.

[0062] All information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.), and signals (including but not limited to signals transmitted between user terminals and other devices) involved in this application have been authorized by the user or fully authorized by all parties, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. For example, the "runtime" mentioned in this disclosure was obtained under full authorization.

[0063] This invention discloses a closed-loop electrical stimulation control system based on multimodal vibration and model prediction, comprising: initializing the system; acquiring personalized parameters and stimulation parameters and determining the basic electrical stimulation intensity in combination with preset functions; acquiring high-frequency and low-frequency vibration signals in real time, performing signal preprocessing to obtain piezoelectric film signals and MEMS accelerometer signals and analyzing them to determine the corresponding muscle movement intensity characteristics; weighted fusion of the muscle movement intensity characteristics to obtain the basic muscle movement intensity; predicting and solving the basic muscle movement intensity to obtain the predicted muscle movement intensity; determining the optimal electrical stimulation intensity based on the predicted muscle movement intensity, the basic electrical stimulation intensity, and the basic muscle movement intensity; applying stimulation to preset positions and repeatedly updating the optimal electrical stimulation intensity until the system terminates operation. This invention dynamically optimizes the electrical stimulation intensity through multimodal vibration signal analysis and model prediction, achieving personalized closed-loop electrical stimulation control.

[0064] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms.

[0065] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units. They may be located in one place or distributed across multiple network units. Some or all of the units may be selected to achieve the purpose of this embodiment according to actual needs.

[0066] In addition, in the various embodiments of the present invention, each functional unit can be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit; the integrated unit can be implemented in hardware or in the form of hardware plus software functional units.

[0067] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0068] Alternatively, if the integrated units of this invention are implemented as software functional modules and sold or used as independent products, they can also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of this invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks.

Claims

1. A closed-loop electrical stimulation control system based on multimodal vibration and model prediction, characterized in that, include: The system initialization module is used to initialize the system. The human-computer interaction module is used to acquire personalized parameters and stimulation parameters and combine them with preset functions to determine the basic electrical stimulation intensity. The signal acquisition and preprocessing module is used to acquire high-frequency vibration signals and low-frequency vibration signals in real time and perform signal preprocessing to obtain piezoelectric film signals and MEMS accelerometer signals; the piezoelectric film signals and MEMS accelerometer signals are analyzed respectively to obtain multiple muscle movement intensity characteristics; The model prediction module is used to normalize and weightedly fuse the multiple muscle movement intensity features to obtain the basic muscle movement intensity; it is used to determine the prediction model using a preset method and combine it with the basic muscle movement intensity to perform prediction and solve to obtain the predicted muscle movement intensity. The optimal electrical stimulation solution module is used to determine the optimal electrical stimulation intensity based on the predicted muscle movement intensity, the baseline electrical stimulation intensity, and the baseline muscle movement intensity. An electrical stimulation output module is used to apply stimulation to a preset position based on the optimal electrical stimulation intensity and the stimulation parameters; The parameter update module is used to update the optimal electrical stimulation intensity and repeat the stimulation until the system finishes running.

2. The closed-loop electrical stimulation control system based on multimodal vibration and model prediction according to claim 1, characterized in that, The initialization of the system includes: The system initialization includes resetting computational parameters, calibrating sensors, and initializing output parameters. The initialization output parameters include at least setting the electrical stimulation intensity of the stimulation electric field to a preset minimum electrical stimulation intensity.

3. The closed-loop electrical stimulation control system based on multimodal vibration and model prediction according to claim 1, characterized in that, The process of obtaining personalized parameters and stimulation parameters and determining the basic electrical stimulation intensity in combination with preset functions includes: The personalized parameters include gender, age, and body type; The preset functions include muscle strength training, spasm relief, and neuromuscular rehabilitation. If the preset function is muscle strength training, then the electrical stimulation intensity range is the first preset intensity range; If the preset function is spasm relief, then the electrical stimulation intensity range is the second preset intensity range; If the preset function is neuromuscular rehabilitation, then the electrical stimulation intensity range is the third preset intensity range; The basic electrical stimulation intensity is determined according to the preset standards, based on the personalized parameters and the range of electrical stimulation intensity. The stimulation parameters include: fundamental frequency, difference frequency, difference frequency variation period, and stimulation time.

4. The closed-loop electrical stimulation control system based on multimodal vibration and model prediction according to claim 1, characterized in that, The real-time acquisition of high-frequency and low-frequency vibration signals, followed by signal preprocessing to obtain piezoelectric thin film signals and MEMS accelerometer signals, includes: The high-frequency vibration signal and the low-frequency vibration signal are filtered by a first preset filter to obtain the filtered high-frequency vibration signal and the filtered low-frequency vibration signal. A preset noise reduction algorithm is used to perform noise reduction processing on the filtered high-frequency vibration signal and the filtered low-frequency vibration signal to obtain the piezoelectric film signal and the MEMS accelerometer signal.

5. The closed-loop electrical stimulation control system based on multimodal vibration and model prediction according to claim 1, characterized in that, The piezoelectric film signal and the MEMS accelerometer signal were analyzed respectively to obtain multiple muscle movement intensity characteristics, including: The muscle movement intensity characteristics include the effective value of the piezoelectric thin film signal, the effective value of the MEMS accelerometer signal, the peak value of the piezoelectric thin film signal, the peak value of the MEMS accelerometer signal, the centroid of the piezoelectric thin film signal frequency, and the centroid of the MEMS accelerometer signal frequency. Extract the corresponding RMS values ​​of the piezoelectric thin film signal and the MEMS accelerometer signal based on the piezoelectric thin film signal and the MEMS accelerometer signal, respectively: ； ； in, For piezoelectric film signals, RMS1 is the effective value of the piezoelectric thin film signal, RMS2 is the effective value of the MEMS accelerometer signal, and T is the preset signal sampling period. Extract the corresponding peak values ​​of the piezoelectric thin film signal and the MEMS accelerometer signal based on the piezoelectric thin film signal and the MEMS accelerometer signal, respectively: ,t∈[0,T]; ,t∈[0,T]; Where Peak1 is the peak value of the piezoelectric thin film signal, and Peak2 is the peak value of the MEMS accelerometer signal; Extract the corresponding centroids of the piezoelectric thin film signal frequency and the MEMS accelerometer signal frequency, respectively, based on the piezoelectric thin film signal and the MEMS accelerometer signal: ； ； Wherein, FC1 is the centroid of the piezoelectric thin film signal frequency, and FC2 is the centroid of the MEMS accelerometer signal frequency. The Fourier transform result of the piezoelectric thin film signal is shown. The result is the Fourier transform of the MEMS accelerometer signal, where [f1,f2] is the frequency range of the piezoelectric thin film signal and [f3,f4] is the frequency range of the MEMS accelerometer signal.

6. The closed-loop electrical stimulation control system based on multimodal vibration and model prediction according to claim 5, characterized in that, The process of normalizing and weighting the multiple muscle movement intensity features to obtain the basic muscle movement intensity includes: Each muscle movement intensity feature is normalized to obtain multiple normalized muscle movement intensity features; The normalized multiple muscle movement intensity features are then weighted and fused: ； Among them, I m (t) represents the baseline muscle activity intensity, where α1, α2, α3, α4, α5, and α6 are preset feature weights, and RMS is used. g1 The normalized effective value of the piezoelectric film signal, RMS g2 Peak is the normalized RMS value of the MEMS accelerometer signal. g1 Peak is the normalized peak value of the piezoelectric thin film signal. g2 The normalized peak value of the MEMS accelerometer signal, FC g1 The centroid of the normalized piezoelectric thin film signal frequency, FC g2 The centroid of the normalized MEMS accelerometer signal frequency.

7. The closed-loop electrical stimulation control system based on multimodal vibration and model prediction according to claim 1, characterized in that, The step of determining a prediction model using a preset method and combining it with the basic muscle activity intensity to perform prediction and solution, to obtain the predicted muscle activity intensity, includes: The prediction model is determined using a pre-defined method; The personalized correction coefficient is determined based on the personalized parameters; Based on the baseline muscle activity intensity and the prediction model, the predicted muscle activity intensity is determined as follows: ； Among them, I m (t|t-1) represents the predicted muscle activity intensity, where a, b, c, and d are preset model parameters, and K is a personalized correction coefficient; m (t-1) represents the baseline muscle activity intensity in the previous second, I m (t-2) represents the basic muscle movement intensity in the first two seconds, and u(t-1) represents the electrical stimulation intensity of the electric field in the first second.

8. The closed-loop electrical stimulation control system based on multimodal vibration and model prediction according to claim 1, characterized in that, The determination of the optimal electrical stimulation intensity based on the predicted muscle movement intensity, the baseline electrical stimulation intensity, and the baseline muscle movement intensity includes: The optimal electrical stimulation intensity is determined based on the predicted muscle activity intensity, the baseline electrical stimulation intensity, and the baseline muscle activity intensity. ； in, For optimal electrical stimulation intensity, Based on the basic electrical stimulation intensity, Basic muscle activity intensity, To predict muscle activity intensity.

9. The closed-loop electrical stimulation control system based on multimodal vibration and model prediction according to claim 2, characterized in that, The step of applying stimulation to a preset location based on the optimal electrical stimulation intensity and the stimulation parameters includes: The optimal electrical stimulation intensity is taken as the electrical stimulation intensity of the stimulation electric field; The stimulation electric field is output based on the electrical stimulation intensity of the stimulation electric field and the stimulation parameters; The stimulation electric field includes: a composite wave interference electric field, a low-frequency modulated intermediate frequency electric field, and a low-frequency electric field.

10. The closed-loop electrical stimulation control system based on multimodal vibration and model prediction according to claim 3, characterized in that, The process of updating the optimal electrical stimulation intensity and repeating the stimulation until the system operation ends includes: Relevant data are repeatedly collected and calculated at preset time intervals to update the optimal electrical stimulation intensity; Stimulation was repeatedly applied based on the updated optimal electrical stimulation intensity; The system terminates operation when the runtime equals the stimulation time.

Images