Molar signal monitoring system and method based on optimized variational mode decomposition

Abstract

The invention discloses a molar signal monitoring system and method based on optimized variational mode decomposition, and relates to the technical field of bio-electricity signal detection and processing. The system comprises a signal acquisition module used for acquiring temporalis surface electromyogram signals; the signal preprocessing module is used for performing preprocessing and envelope calibration on the temporalis surface electromyogram signals; the GWO-VMD algorithm module is used for performing multi-modal characteristic decomposition on the target temporalis surface electromyogram signals by using a GWO-VMD algorithm; the feature extraction module is used for extracting time domain features and frequency domain features from the multiple modal components; the classification and recognition module is used for classifying the feature vectors; and the real-time feedback and evaluation module is used for carrying out molar state evaluation according to the state classification result, and generating and sending a molar state monitoring result. According to the invention, different actions such as tooth grinding, tooth clenching, speaking, chewing and the like can be accurately identified, so that the signals of tooth grinding and tooth clenching are extracted for quantitative calculation and evaluation, and higher monitoring precision and real-time performance are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of bioelectrical signal detection and processing. Specifically, it relates to a molar signal monitoring system and method based on optimized variational mode decomposition. Background Art

[0002] Bruxism is a common oral disease, manifested as involuntary rhythmic grinding or clenching between the upper and lower jaw teeth. Bruxism may lead to a series of oral health problems such as tooth wear, temporomandibular joint disorders, and tooth restoration failures, and even affect the sleep quality of patients. However, the current diagnosis and monitoring of bruxism still face great challenges. The existing technologies mainly include polysomnography, piezoelectric sensing detectors, bruxism detection tablets, bruxism sound monitoring, and surface electromyogram (sEMG) recording.

[0003] However, the existing bruxism monitoring technologies generally have the following defects:

[0004] Inaccurate signal recognition: The existing technologies cannot effectively distinguish the signals of actions such as bruxism, tooth clenching, speaking, and chewing, resulting in a low recognition accuracy of bruxism signals.

[0005] Signal interference: Many technologies are prone to interference from external noises or other physiological activities during signal acquisition. Especially methods such as bruxism sound monitoring and the existing surface electromyogram, which are likely to affect the accuracy of signals.

[0006] Lack of quantitative evaluation: Most of the existing methods lack an effective quantitative evaluation mechanism and cannot accurately analyze the severity, type, etc. of bruxism, affecting the accuracy of diagnosis.

[0007] Poor real-time performance: Many technologies rely on post-data analysis or long-term monitoring and cannot provide real-time feedback. Especially for nocturnal bruxism, which is a sudden behavior, it is difficult to diagnose in a timely manner.

[0008] Uncomfortable wearing: Most technologies require wearing devices such as piezoelectric sensors and bruxism detection tablets. These devices may affect the comfort of patients, resulting in a poor wearing experience.

[0009] Therefore, there is an urgent need for a technical solution that can effectively improve the monitoring accuracy of bruxism, provide real-time feedback, and have the ability of quantitative evaluation. Summary of the Invention

[0010] In order to overcome the above problems or at least partially solve the above problems, the present invention provides a molar signal monitoring system and method based on optimized variational mode decomposition, which can accurately identify different actions such as bruxism, tooth clenching, speaking, and chewing, so as to extract the signals of bruxism and tooth clenching for quantitative calculation and evaluation, and has higher monitoring accuracy and real-time performance.

[0011] To solve the above technical problems, the technical solution adopted by the present invention is as follows:

[0012] In a first aspect, the present invention provides a molar signal monitoring system based on optimized variational mode decomposition, including a signal acquisition module, a signal preprocessing module, a GWO-VMD algorithm module, a feature extraction module, a classification and recognition module, and a real-time feedback and evaluation module, where:

[0013] The signal acquisition module is used to acquire the surface electromyogram signal of the temporalis muscle;

[0014] The signal preprocessing module is used to preprocess and envelope calibrate the surface electromyogram signal of the temporalis muscle to obtain the target surface electromyogram signal of the temporalis muscle;

[0015] The GWO-VMD algorithm module is used to perform multi-modal feature decomposition on the target surface electromyogram signal of the temporalis muscle by using the GWO-VMD algorithm to obtain multiple modal components;

[0016] The feature extraction module is used to extract time-domain features and frequency-domain features from multiple modal components to construct a feature vector;

[0017] The classification and recognition module is used to classify the feature vector by using the support vector machine algorithm to generate a classification result of the temporalis muscle state;

[0018] The real-time feedback and evaluation module is used to evaluate the molar state according to the state classification result and generate and send the monitoring result of the molar state.

[0019] Through the cooperation of multiple modules such as the signal acquisition module, the signal preprocessing module, the GWO-VMD algorithm module, the feature extraction module, the classification and recognition module, and the real-time feedback and evaluation module, this system can accurately identify different actions such as molar grinding, teeth clenching, speaking, and chewing by collecting the surface electromyogram signal (sEMG) of the temporalis muscle, and then extract the signals of molar grinding and teeth clenching for quantitative calculation and evaluation, with higher monitoring accuracy and real-time performance.

[0020] Based on the first aspect, further, the above signal acquisition module includes an ADS1299 chip and an ESP32 microcontroller. The ADS1299 chip is used to acquire the surface electromyogram signal of the temporalis muscle, and the ESP32 microcontroller is used to manage the sampling, data transmission, and control of the ADS1299 chip.

[0021] Based on the first aspect, further, the above signal acquisition module further includes a hardware filter, and the hardware filter is connected to the ADS1299 chip.

[0022] Based on the first aspect, further, the above signal preprocessing module includes a denoising sub-module, a normalization processing sub-module, and an envelope calibration sub-module, where:

[0023] The denoising sub-module is used to filter and denoise the surface electromyogram signal of the temporalis muscle, removing power frequency interference, low-frequency artifacts, high-frequency noise, as well as baseline drift and abnormal spikes;

[0024] The normalization processing sub-module is used to eliminate the signal amplitude difference and dimensional influence of the surface electromyogram signal of the temporalis muscle, and normalize all signals to a unified range;

[0025] The envelope calibration sub-module is used to extract the envelope of the surface electromyogram signal of the temporalis muscle by the moving average method, and use the adaptive threshold method to identify the activity start and end points of the surface electromyogram signal of the temporalis muscle, so as to achieve envelope calibration.

[0026] Based on the first aspect, further, the above GWO-VMD algorithm module combines the Grey Wolf Optimization algorithm GWO and the Variational Mode Decomposition algorithm VMD to optimize the key parameters of VMD, and decomposes the target surface electromyogram signal of the temporalis muscle based on the optimized key parameters of VMD; the key parameters include the number of modes K and the penalty factor α.

[0027] Based on the first aspect, further, the above time-domain features include root mean square, average absolute value, zero crossing rate, differential mean square value, and average amplitude change, and the energy, dynamics, and frequency characteristics of the target surface electromyogram signal of the temporalis muscle are described based on the time-domain features.

[0028] In the second aspect, the present invention provides a molar signal monitoring method based on optimized variational mode decomposition, including the following steps:

[0029] Collect the surface electromyogram signal of the temporalis muscle;

[0030] Preprocess and calibrate the envelope of the surface electromyogram signal of the temporalis muscle to obtain the target surface electromyogram signal of the temporalis muscle;

[0031] Use the GWO-VMD algorithm to perform multi-modal feature decomposition on the target surface electromyogram signal of the temporalis muscle to obtain multiple modal components;

[0032] Extract time-domain features and frequency-domain features from multiple modal components to construct a feature vector;

[0033] Use the support vector machine algorithm to classify the feature vector to generate a temporalis muscle state classification result;

[0034] Evaluate the molar state according to the state classification result, and generate and send the molar state monitoring result.

[0035] By collecting the surface electromyogram (sEMG) signals of the temporalis muscle and combining the improved Grey Wolf Optimization (GWO) algorithm to optimize the parameters of Variational Mode Decomposition (VMD), different actions such as grinding teeth, clenching teeth, speaking, and chewing can be accurately identified. Thus, the signals of grinding teeth and clenching teeth can be extracted for quantitative calculation and evaluation, with higher monitoring accuracy and real-time performance.

[0036] Based on the second aspect, further, the method for multi-modal feature decomposition of the target temporalis muscle surface electromyogram signal using the GWO-VMD algorithm includes the following steps:

[0037] Combine the Grey Wolf Optimization algorithm GWO and the Variational Mode Decomposition algorithm VMD to optimize the key parameters of VMD;

[0038] Based on the optimized key parameters of VMD, perform multi-modal feature decomposition on the target temporalis muscle surface electromyogram signal.

[0039] In a third aspect, the present application provides an electronic device, which includes a memory for storing one or more programs; a processor; when the one or more programs are executed by the processor, the method according to any one of the above second aspects is implemented.

[0040] In a fourth aspect, the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the method according to any one of the above second aspects is implemented.

[0041] The present invention has at least the following advantages or beneficial effects:

[0042] 1. Improve the accuracy of grinding teeth diagnosis: Through the GWO-VMD optimization algorithm, the surface electromyogram signals of the temporalis muscle can be accurately decomposed and analyzed, thus achieving high-precision identification of actions such as grinding teeth and clenching teeth. Traditional methods often lead to diagnostic errors due to the inability to effectively distinguish the signals of different actions, and this technology effectively makes up for this deficiency.

[0043] 2. Real-time monitoring and instant feedback: The system can achieve real-time monitoring of abnormal actions such as grinding teeth, and based on the collected signals, it can provide real-time feedback on quantitative indicators such as the intensity and frequency of grinding teeth, helping users detect and intervene in grinding teeth symptoms early.

[0044] 3. Provide a detailed evaluation of the grinding teeth state: Through the quantitative analysis of multiple indicators such as the single contraction intensity, duration, and frequency of grinding teeth, this technology can accurately evaluate the severity of grinding teeth. Compared with the qualitative analysis of the prior art, the quantitative indicators of this technology provide more reference data for clinicians to help formulate personalized treatment plans.

[0045] 4. User Experience Optimization: By adopting a compact and efficient hardware design (such as ADS1299 and ESP32 microcontrollers), the device is small in size and low in power consumption. While ensuring high-precision data acquisition, it can reduce the burden on users when wearing.

[0046] 5. Multimodal Feature Analysis: The system can not only analyze grinding teeth status and clenching teeth, but also analyze various different temporalis muscle activities such as chewing and speaking, further improving the comprehensiveness and accuracy of diagnosis. The signal decomposition and analysis of each functional state provide strong support for subsequent treatment intervention and prevention. Brief Description of the Drawings

[0047] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required in the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and therefore should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other relevant drawings can also be obtained based on these drawings.

[0048] Figure 1 It is the principle block diagram of a molar signal monitoring system based on optimized variational mode decomposition according to an embodiment of the present invention;

[0049] Figure 2 It is the schematic diagram of the signal acquisition module in a molar signal monitoring system based on optimized variational mode decomposition according to an embodiment of the present invention;

[0050] Figure 3 It is the schematic diagram of the GWO optimization convergence curve in an embodiment of the present invention;

[0051] Figure 4 It is the schematic diagram of the surface electromyogram signal in the clenching teeth state and its IMFs decomposed by VMD in an embodiment of the present invention;

[0052] Figure 5 It is the schematic diagram of the surface electromyogram signal in the grinding teeth state and its IMFs decomposed by VMD in an embodiment of the present invention;

[0053] Figure 6 It is the schematic diagram of the surface electromyogram signal in the chewing state and its IMFs decomposed by VMD in an embodiment of the present invention;

[0054] Figure 7 It is the schematic diagram of the surface electromyogram signal in the speaking state and its IMFs decomposed by VMD in an embodiment of the present invention;

[0055] Figure 8 It is the schematic diagram of the surface electromyogram signal in the relaxation state and its IMFs decomposed by VMD in an embodiment of the present invention;

[0056] Figure 9This is a flowchart of a molar signal monitoring method based on optimized variational mode decomposition according to an embodiment of the present invention. Detailed implementation manners

[0057] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and illustrated herein generally may be arranged and designed in a variety of different configurations.

[0058] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0059] It should be noted that like reference numerals and letters denote like items in the following drawings, and thus, once an item is defined in one drawing, it is not necessary to further define and explain it in subsequent drawings.

[0060] It should be noted that in this document, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual such relationship or order between these entities or operations. Moreover, the term "comprising" or any other variation thereof is intended to cover a non-exclusive inclusion, such that a process, method, article, or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or device. Without further limitation, an element defined by the phrase "comprising a..." does not exclude the presence of additional identical elements in the process, method, article, or device comprising the element.

[0061] In the description of the embodiments of the present invention, "a plurality of" represents at least two.

[0062] Embodiment:

[0063] As Figure 1 shown, in a first aspect, an embodiment of the present invention provides a molar signal monitoring system based on optimized variational mode decomposition, including a signal acquisition module, a signal preprocessing module, a GWO-VMD algorithm module, a feature extraction module, a classification and recognition module, and a real-time feedback and evaluation module, where:

[0064] The signal acquisition module is configured to acquire surface electromyogram signals of the temporalis muscle;

[0065] A signal preprocessing module for preprocessing and envelope calibration of the surface electromyogram signal of the temporalis muscle to obtain the target surface electromyogram signal of the temporalis muscle;

[0066] A GWO-VMD algorithm module for performing multi-modal feature decomposition on the target surface electromyogram signal of the temporalis muscle by using the GWO-VMD algorithm to obtain multiple modal components;

[0067] A feature extraction module for extracting time-domain features and frequency-domain features from multiple modal components to construct a feature vector. After the signal is decomposed by VMD, the system realizes feature extraction by extracting time-domain features from each IMF. Five time-domain features include root mean square (RMS), mean absolute value (MAV), zero crossing rate (ZCR), differential average square deviation (DASD), and average amplitude change (AAC), which can effectively describe the energy, dynamics, and frequency characteristics of the signal. Through the extraction of the above five features, a five-dimensional feature vector is extracted from each IMF. If the signal generates M IMFs after GWO-VMD decomposition, the final feature vector constructed is:

[0068] F = [F1, F1, …, F M

[0069] where F i is the five-dimensional feature vector of the i-th IMF, and F is a feature matrix with a total dimension of 5×M. The feature vectors extracted from the multi-modal components cover the energy, dynamics, and frequency characteristics of the signal, providing sufficient prior information for the classification model.

[0070] A classification and recognition module for classifying the feature vector by using the support vector machine algorithm to generate a classification result of the temporalis muscle state; after the signal features are extracted, the support vector machine (SVM) is used for classification. SVM is a supervised learning algorithm commonly used in classification tasks, especially suitable for dealing with small samples and high-dimensional data, and can effectively distinguish different types of electromyogram signals. By training the classification model, the system can classify the extracted features, accurately distinguish different states such as grinding teeth, clenching teeth, speaking, chewing, etc., and monitor their occurrence in real time.

[0071] Classification effect:

[0072] ​To verify the effectiveness of using GWO to optimize VMD parameters for classifying the functional state of the temporalis muscle in the present invention, three groups of comparative experiments were designed and carried out: (1) direct feature extraction and classification without VMD preprocessing; (2) feature extraction and classification based on VMD with fixed parameters; (3) feature extraction and classification based on VMD optimized by GWO. The main objective of the experiment was to evaluate the impact of different signal processing methods on the classification performance, so as to verify the advantages of GWO-optimized VMD parameters in feature extraction and classification tasks. The classification experiment results are shown in Table 1.

[0073] Table 1: Comparison table of classification results of three signal processing methods

[0074]

[0075] In Experiment 1, 5 feature values (such as RMS, MAV, ZCR, DASD, AAC) were directly extracted from the original sEMG signal and input into a support vector machine (SVM) model for classification. The experimental results showed that the classification accuracy was 89.5%, and the precision, recall rate, and F1 value were all around 90%. However, there were significant cross-errors between the chewing and speaking states, and the classification effect was not good. This indicates that the original signal without VMD decomposition contains more noise and invalid information, resulting in insufficient discrimination in feature extraction, thus limiting the classification performance.

[0076] In Experiment 2, the sEMG signal was decomposed using VMD with fixed parameters (K = 5, α = 2000) to obtain 10 groups of intrinsic mode functions (IMFs), and 5 feature values of each IMF were extracted. After recombining the feature vectors, they were input into the SVM model for classification. The experimental results showed that the classification accuracy was improved to 95.25%, and the precision, recall rate, and F1 value all exceeded 95%. However, there were still some cross-errors between the chewing and speaking states. This indicates that although the VMD decomposition with fixed parameters can effectively improve the classification performance, due to the parameter selection relying on experience and not being able to fully adapt to the optimal feature frequency band of the signal, there is still room for improvement in the classification performance.

[0077] In Experiment 3, the sEMG signal was decomposed by optimizing the VMD parameters with GWO (K = 10, α = 142) to obtain 10 groups of IMFs, and the same feature values were extracted and then input into the SVM model for classification. The experimental results showed that the classification accuracy reached 98.25%, and the precision, recall rate, and F1 value were all above 98%. Compared with Experiment 1 and Experiment 2, the VMD method optimized by GWO significantly improved the classification performance, especially the cross-errors between the chewing and speaking states were significantly reduced. This indicates that the VMD parameters optimized by GWO can more accurately extract the feature frequency band of the signal, thus achieving accurate classification of the five functional states of the temporalis muscle.

[0078] Based on the comprehensive results of the three groups of experiments, it can be seen that the VMD method optimized by GWO significantly improves the accuracy of sEMG signal feature extraction and the classification performance. The IMFs decomposed with the optimized parameters (K = 10, a = 142) can better highlight the signal feature frequency bands of different temporal muscle functional states, thus improving the classification accuracy. In contrast, the fixed-parameter VMD decomposition method has good classification ability to a certain extent, but its performance depends on the parameter setting experience. The direct feature extraction method without VMD decomposition leads to a significant decline in classification performance due to the failure to effectively remove noise. The experimental results verify the effectiveness of the GWO-optimized VMD method and prove its advantages in the classification task of temporal muscle functional states. This method not only improves the separation degree of signal features but also enhances the generalization ability of the SVM classifier, providing important technical support for subsequent electromyogram signal analysis, muscle function monitoring, and motion recognition research.

[0079] The real-time feedback and evaluation module is used to evaluate the bruxism state according to the state classification results and generate and send the monitoring results of the bruxism state.

[0080] This system can monitor and feedback the bruxism state in real time, provide instant warnings, and quantitatively evaluate the severity and type of bruxism, providing an auxiliary basis for clinical diagnosis.

[0081] The collected surface electromyogram signal sequence of bruxism x = {x1, x2, …, x n}, where x i is the sEMG sampling value, with the unit being dimensionless, and 0 ≤ x i ≤ 1023. Define the intensity measure of a single bruxism: That is, the average integral electromyogram of the envelope of a single muscle contraction, which is used to measure the strength of bruxism and is calculated as follows:

[0082]

[0083] Furthermore, various indicators for measuring the bruxism state can be calculated. As follows

[0084] The average value of the intensity of a single muscle contraction:

[0085] The maximum value of the intensity of a single muscle contraction:

[0086] The minimum value of the intensity of a single muscle contraction:

[0087] The time of a single bruxism: T single = (end - start + 1)Δt (seconds)

[0088] The longest time of a single bruxism: T max = max({T single}) (seconds)

[0089] Shortest time for single molar grinding: T min = min({T single})(seconds)

[0090] Cumulative time of molar grinding: T total = sum({T single})(seconds)

[0091] Average time for single molar grinding: T avg = mean({T single})(seconds)

[0092] Cumulative number of molar grindings: N total (times)

[0093] Average frequency of molar grinding:

[0094] To some extent, these index parameters can measure the intensity, frequency, and cycle of molar grinding, reflect the severity of bruxism, and provide a reference for the diagnosis of bruxism.

[0095] Through the cooperation of multiple modules such as the signal acquisition module, signal preprocessing module, GWO-VMD algorithm module, feature extraction module, classification and recognition module, and real-time feedback and evaluation module, this system can accurately identify different actions such as molar grinding, teeth clenching, speaking, and chewing by collecting the surface electromyogram (sEMG) of the temporalis muscle, and combine the improved Grey Wolf Optimization (GWO) algorithm to optimize the parameters of Variational Mode Decomposition (VMD). Thus, it can extract the signals of molar grinding and teeth clenching for quantitative calculation and evaluation, with higher monitoring accuracy and real-time performance.

[0096] This system can improve the accuracy of bruxism diagnosis: Through the GWO-VMD optimization algorithm, it can accurately decompose and analyze the surface electromyogram signals of the temporalis muscle, thereby achieving high-precision recognition of actions such as bruxism and teeth clenching. Traditional methods often lead to diagnostic errors due to the inability to effectively distinguish signals of different actions, and this technology effectively makes up for this deficiency. It can also provide real-time monitoring and immediate feedback: The system can achieve real-time monitoring of abnormal actions such as bruxism, and based on the collected signals, it can provide real-time feedback on quantitative indicators such as the intensity and frequency of bruxism, helping users detect and intervene in bruxism symptoms early. At the same time, this system also provides a detailed assessment of the bruxism state: Through the quantitative analysis of multiple indicators such as the single contraction intensity, duration, and frequency of bruxism, this technology can accurately evaluate the severity of bruxism. Compared with the qualitative analysis of existing technologies, the quantitative indicators of this technology provide more reference data for clinicians to help formulate personalized treatment plans. Further, by adopting a compact and efficient hardware design (such as ADS1299 and ESP32 microcontrollers), the device has a small volume and low power consumption, which can ensure high-precision data acquisition while reducing the burden on users when wearing. In this system, based on multi-modal feature analysis, it can not only analyze the bruxism state and teeth clenching, but also analyze various temporalis muscle activities such as chewing and speaking, further improving the comprehensiveness and accuracy of diagnosis. The signal decomposition and analysis of each functional state provide strong support for subsequent treatment intervention and prevention.

[0097] Based on the first aspect, further, the above signal acquisition module includes an ADS1299 chip and an ESP32 microcontroller. The ADS1299 chip is used to collect the surface electromyogram signals of the temporalis muscle, and the ESP32 microcontroller is used to manage the sampling, data transmission, and control of the ADS1299 chip.

[0098] Based on the first aspect, further, the above signal acquisition module further includes a hardware filter, and the hardware filter is connected to the ADS1299 chip.

[0099] In some embodiments of the present invention, as Figure 2 shown, the signal acquisition module is mainly composed of ADS1299 and ESP32, and is mainly used to collect and process the surface electromyogram signals of the temporalis muscle. This module also includes a power supply module, an electrostatic protection module, a hardware filter, and a storage module. The power supply module supplies power to other components, and an electrostatic protection module is set at the signal input end, and this electrostatic protection module is connected to the hardware filter. The specific hardware design includes:

[0100] ADS1299 chip: A 24-bit high-precision analog front-end chip designed specifically for the processing of bioelectrical signals (such as electromyogram signals). Its high signal-to-noise ratio and low power consumption characteristics make it very suitable for capturing weak electromyogram signal fluctuations.

[0101] ESP32 microcontroller: Responsible for managing the sampling, data transmission, and control of the ADS1299. Transmits the acquired data to the host computer in real time via a Bluetooth module for further processing and storage.

[0102] Sampling rate: The system is set to a sampling rate of 1000Hz to ensure that the dynamic characteristics of the EMG signals can be captured, which is suitable for real-time analysis.

[0103] Hardware filter: A 50Hz notch filter (to remove power frequency interference) and a 10 - 500Hz band-pass filter (to extract the effective EMG signal frequency band) are designed to ensure the quality of the acquired signals.

[0104] Based on the first aspect, further, the above signal preprocessing module includes a denoising sub-module, a normalization processing sub-module, and an envelope calibration sub-module, where:

[0105] The denoising sub-module is used to filter and denoise the surface EMG signals of the temporalis muscle, removing power frequency interference, low-frequency artifacts, high-frequency noise, as well as baseline drift and abnormal spikes.

[0106] The normalization processing sub-module is used to eliminate the signal amplitude differences and dimensional effects of the surface EMG signals of the temporalis muscle, and normalize all signals to a unified range.

[0107] The envelope calibration sub-module is used to extract the envelope of the surface EMG signals of the temporalis muscle by the moving average method, and use the adaptive threshold method to identify the start and end points of the activity of the surface EMG signals of the temporalis muscle to achieve envelope calibration.

[0108] In some embodiments of the present invention, in order to improve the quality of the acquired signals and ensure the reliability of subsequent analysis, the present invention performs multi-step signal preprocessing after data acquisition, mainly including:

[0109] Normalization processing: Eliminate signal amplitude differences and dimensional effects, and normalize all signals to a unified range.

[0110] Software denoising: Use a 50Hz notch filter to remove power frequency interference, a 10Hz high-pass filter to remove low-frequency artifacts (such as motion artifacts), and a 500Hz low-pass filter to remove high-frequency noise. In addition, median filtering is also performed using data with a window size of 50 to effectively eliminate baseline drift and abnormal spikes.

[0111] Signal envelope calibration: Extract the signal envelope by the moving average method, and use the adaptive threshold method to identify the start and end points of the signal activity. By calibrating the activity envelope, ensure the clarity and accuracy of the signal segment.

[0112] Based on the first aspect, further, the above GWO-VMD algorithm module combines the Grey Wolf Optimization algorithm (GWO) and the Variational Mode Decomposition algorithm (VMD), optimizes the key parameters of VMD, and decomposes the target surface electromyogram signal of the temporalis muscle based on the optimized key parameters of VMD; the key parameters include the number of modes K and the penalty factor α.

[0113] In some embodiments of the present invention, the GWO-VMD algorithm combines the Grey Wolf Optimization algorithm (GWO) and the Variational Mode Decomposition (VMD), aiming to achieve high-quality decomposition of the signal by optimizing two key parameters of VMD (the number of modes K and the penalty factor α). The core of this method is to use the global search ability of GWO to find the optimal combination of VMD parameters, so as to improve the adaptability and accuracy of signal decomposition.

[0114] Algorithm flow:

[0115] 1. Initialization

[0116] In the initialization stage, set the number of grey wolf populations N, the maximum number of iterations T, and the value ranges of the VMD parameters K and α. Randomly initialize the positions of each grey wolf, and each position corresponds to a combination value of K and α to ensure uniform coverage of the search space. Initially, the fitness values of all grey wolves are set to infinity to provide an initial benchmark for subsequent optimization.

[0117] 2. Fitness function calculation

[0118] In each iteration, for the position of each grey wolf (i.e., the current parameters K and α), use VMD to decompose the target signal to obtain multiple modal components. To quantify the decomposition effect, permutation entropy (PE) and mutual information (MI) are introduced as evaluation indicators:

[0119] Permutation entropy (PE): Measures the complexity of the modal component and is defined as:

[0120]

[0121] where p i represents the frequency of the i-th permutation pattern in the signal time series, and n is the number of possible permutation patterns.

[0122] Mutual information (MI): Measures the correlation between the modal component and the original signal and is defined as:

[0123]

[0124] where p(x,y) is the joint probability distribution, and p(x) and p(y) are the marginal probability distributions.

[0125] To further evaluate the VMD decomposition effect, the ratio of permutation entropy to mutual information is used as the fitness function Fitness to balance the component complexity and the correlation of the original signal. The smaller the fitness value, the better the decomposition effect.

[0126]

[0127] 3. Update the positions of α, β, and δ wolves

[0128] Grey wolves update their positions dynamically by simulating the hunting process. The core is that each grey wolf adjusts its own position based on the positions of the α, β, and δ wolves in the population. The formula is as follows:

[0129]

[0130]

[0131] Among them, and represent the positions of the p-th type of wolf (α, β, δ wolves) and the i-th wolf respectively; is the distance; and are dynamic adjustment coefficients, defined as:

[0132]

[0133] is a random number within the range of [0, 1]; a is the step size attenuation factor, which changes dynamically with the iteration number t. The formula is:

[0134]

[0135] Among them, T is the maximum number of iterations.

[0136] 4. Position update and out-of-bounds processing

[0137] In each iteration, all grey wolves comprehensively adjust their positions according to the positions of the α, β, and δ wolves. To avoid parameter out-of-bounds, a constraint mechanism is used to limit the positions within a predefined range, thus ensuring the physical meaning of K and α.

[0138] 5. Introduction of chaotic mapping

[0139] To improve the population diversity and avoid falling into local optima, GWO-VMD introduces chaotic mapping to generate random numbers and Chaotic mapping enhances the randomness through its nonlinear characteristics, thus improving the global search ability.

[0140] The formula of chaotic mapping is:

[0141] r n+1= μ·r n ·(1 - r n ), μ = 4

[0142] where r n is the current random number and μ is the control factor.

[0143] 6. Iteration termination condition

[0144] The algorithm continuously iterates until the maximum number of iterations T is reached or the population fitness value converges. Finally, the optimized K * and α * parameters are output, and based on this, the parameter optimization of VMD is completed.

[0145] Under the optimization of the GWO-VMD algorithm, the sEMG signal is decomposed into multiple Intrinsic Mode Functions (IMFs).

[0146] Results of the GWO-VMD algorithm:

[0147] 1. GWO optimization performance

[0148] To verify the effect of the Grey Wolf Optimization (GWO) algorithm on optimizing the key parameters of Variational Mode Decomposition (VMD), 20% of the samples were randomly selected from each of the five states of clenching teeth, grinding teeth, chewing, speaking, and relaxation in the experiment, totaling 400 samples as the dataset. By selecting the fitness function to optimize the mode number K and the penalty factor α. During the optimization process, the convergence curve of the fitness value is as Figure 3 shown.

[0149] 1.1 Convergence characteristic analysis

[0150] From the experimental results, the GWO algorithm shows significant convergence characteristics during the optimization process:

[0151] Initial stage: In the first few iterations, the fitness value shows a rapid downward trend, indicating that the GWO algorithm has high global search ability and can quickly approach the optimal solution.

[0152] Mid-to-late stage: As the number of iterations increases, the fitness value tends to be stable and the fluctuation amplitude decreases. This indicates that while searching for the optimal solution, GWO can avoid falling into local optima and shows good convergence stability.

[0153] 1.2 Optimization results and significance

[0154] After optimization by GWO, the optimal parameters of VMD were finally determined as K = 10 and α = 142. This parameter combination can achieve precise decomposition of signals, effectively retain the key characteristics of signals, and simultaneously suppress noise and redundant information. The optimized VMD parameters improve the accuracy of decomposed signals and the physical meaning of components, laying a solid foundation for subsequent feature extraction and classification tasks.

[0155] 1.3 Discussion on Algorithm Advantages

[0156] Compared with traditional parameter empirical setting or random search methods, GWO shows the following advantages in the optimization process:

[0157] Global search ability: By simulating the hunting behavior of wolf packs, GWO can efficiently search for the optimal solution within the global range and avoid falling into local optima.

[0158] Strong stability: In the middle and late iterations, the fitness value fluctuates less, reflecting the strong convergence stability of the algorithm.

[0159] High computational efficiency: The definition of the fitness function combines the balance relationship between permutation entropy (PE) and mutual information (MI), enabling GWO to quickly find the optimum in complex signal decomposition and significantly improving the optimization efficiency.

[0160] 2 VMD Decomposition Effect

[0161] 2 VMD Decomposition Effect

[0162] In this study, to analyze the surface electromyogram signal characteristics of five functional states of the temporalis muscle (grinding teeth, bruxism, chewing, speaking, relaxation), the optimized variational mode decomposition (VMD) method was used to perform multi-modal feature decomposition on the collected original signals. The decomposition results are as Figures 4 - 8 shown, demonstrating significant differences in the frequency distribution and modal energy characteristics of signals in different states.

[0163] 2.1 Decomposition Effect

[0164] 2.1.1 Grinding Teeth State

[0165] The amplitude of the original signal in the grinding teeth state is relatively high and concentrated, reflecting the strong contraction characteristics of the muscle. After decomposition by VMD, it is found that the energy of the low-frequency and medium-frequency modes (IMF 2 to IMF 6) dominates and has significant periodicity. The energy of the high-frequency mode gradually decays, indicating that the main energy of the grinding teeth action is concentrated in the low-frequency range. These characteristics are consistent with the strong and persistent muscle activation pattern of the grinding teeth action.

[0166] 2.1.2 Bruxism State

[0167] The amplitude of the original signal in the bruxism state is small, showing continuous but irregular fluctuations. The decomposition results are as Figure 5As shown, the results show that the intermediate frequency modes (IMF 1 to IMF 5) are dominant, and the energy contribution of the high frequency modes is relatively low. The energy distribution is relatively dispersed, and the spectral characteristics are relatively complex. This phenomenon is related to the irregularity of the grinding action and the involvement of more nerve controls.

[0168] 2.1.3 Chewing state

[0169] The signals in the chewing state show obvious periodicity and rhythm. The results of VMD decomposition are as Figure 6 shown, which indicates that the energy ratios of the low frequency and intermediate frequency modes are significant, and the spectral distribution is stable, indicating that the action is regular and repetitive. The energy of the high frequency mode is weak, highlighting the dominant position of the low frequency component. The rhythm of chewing causes the signal energy to be concentrated, and this characteristic provides an important basis for pattern recognition.

[0170] 2.1.4 Speaking state

[0171] The amplitude of the original signal in the speaking state is small, but the high frequency components increase significantly. The decomposition results are as Figure 7 shown: the middle and high frequency modes (IMF 4 to IMF 10) occupy the main energy, reflecting the characteristics of fast and complex muscle activities; the frequency distribution is wide, indicating that the speaking process requires the coordinated work of multiple groups of muscles to complete complex oral and vocal cord movements.

[0172] 2.1.5 Relaxed state

[0173] The overall energy of the EMG signal in the relaxed state is low, showing a relatively stable trend. It is mainly concentrated in the low frequency part, while the middle and high frequency components are weak, which is in line with the physiological characteristics of low muscle tension and low activation level. The decomposition results are as Figure 8 shown.

[0174] It can be seen from the above decomposition results that the frequency distribution and modal energy characteristics under different functional states of the temporalis muscle are basically consistent with its physiological mechanism, further verifying the applicability and stability of the optimized VMD method in EMG signal analysis. VMD can not only effectively decompose complex signals, but also highlight the characteristic information of different frequency bands in the signal, providing reliable support for subsequent feature extraction and pattern recognition. By decomposing the surface EMG signal of the temporalis muscle into a series of modal components with clear frequency characteristics, this study reveals the regularity and difference of muscle activities in five functional states. These findings not only help to deeply understand the muscle activation patterns under different temporalis muscle states, but also provide a theoretical basis for muscle function monitoring, action recognition and diagnosis of chewing dysfunction.

[0175] As Figure 9 shown, on the second hand, the embodiment of the present invention provides a method for monitoring grinding signals based on optimized variational mode decomposition, including the following steps:

[0176] S1. Collect the surface electromyogram signals of the temporalis muscle;

[0177] S2. Preprocess and envelope calibrate the surface electromyogram signals of the temporalis muscle to obtain the target surface electromyogram signals of the temporalis muscle;

[0178] S3. Use the GWO-VMD algorithm to perform multi-modal feature decomposition on the target surface electromyogram signals of the temporalis muscle to obtain multiple modal components;

[0179] S4. Extract time-domain features and frequency-domain features from multiple modal components to construct a feature vector;

[0180] S5. Use the support vector machine algorithm to classify the feature vector to generate the classification result of the temporalis muscle state;

[0181] S6. Evaluate the bruxism state according to the state classification result, and generate and send the monitoring result of the bruxism state.

[0182] By collecting the surface electromyogram signals (sEMG) of the temporalis muscle and combining the improved grey wolf optimization algorithm (GWO) to optimize the parameters of variational mode decomposition (VMD), different actions such as bruxism, teeth clenching, speaking, and chewing can be accurately identified, so as to extract the signals of bruxism and teeth clenching for quantitative calculation and evaluation, which has higher monitoring accuracy and real-time performance.

[0183] In a third aspect, further, the method for performing multi-modal feature decomposition on the target surface electromyogram signals of the temporalis muscle by using the GWO-VMD algorithm includes the following steps:

[0184] Combine the grey wolf optimization algorithm GWO and the variational mode decomposition algorithm VMD to optimize the key parameters of VMD;

[0185] Perform multi-modal feature decomposition on the target surface electromyogram signals of the temporalis muscle based on the optimized key parameters of VMD.

[0186] In a third aspect, an embodiment of the present application provides an electronic device, which includes a memory for storing one or more programs; a processor. When one or more programs are executed by the processor, the method according to any one of the above second aspects is implemented.

[0187] It further includes a communication interface, and the memory, the processor, and the communication interface are directly or indirectly electrically connected to each other to realize data transmission or interaction. For example, these components can be electrically connected to each other through one or more communication buses or signal lines. The memory can be used to store software programs and modules, and the processor executes the software programs and modules stored in the memory to perform various functional applications and data processing. The communication interface can be used to communicate with other node devices for signaling or data.

[0188] Among them, the memory can be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electric Erasable Programmable Read-Only Memory (EEPROM), etc.

[0189] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc.; it can also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0190] In the embodiments provided in this application, it should be understood that the disclosed methods and systems can also be implemented in other ways. The method and system embodiments described above are only illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of the methods, systems, and computer program products according to multiple embodiments of this application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than marked in the accompanying drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and they can sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.

[0191] In addition, in each embodiment of the present application, the functional modules can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.

[0192] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the method according to any one of the above second aspects is implemented. If the function is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0193] The above are only the preferred embodiments of the present invention and are not used to limit the present invention. For those skilled in the art, the present invention can have various changes and modifications. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

[0194] For those skilled in the art, it is obvious that the present application is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or basic characteristics of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present application is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application. Any reference signs in the claims should not be regarded as limiting the claimed rights.

Claims

1. A molar signal monitoring system based on optimized variational mode decomposition, characterized in that: It includes signal acquisition module, signal preprocessing module, GWO-VMD algorithm module, feature extraction module, classification and recognition module, and real-time feedback and evaluation module, among which: A signal acquisition module, used for collecting surface electromyographic signals of the temporalis muscle; A signal preprocessing module is used to preprocess and calibrate the envelope of the temporalis muscle surface electromyography signal to obtain the target temporalis muscle surface electromyography signal; A GWO-VMD algorithm module is used to perform multi-modal feature decomposition on the surface electromyographic signal of the target temporalis muscle using the GWO-VMD algorithm to obtain multiple modal components; A feature extraction module is used to extract time domain features and frequency domain features from multiple modal components to construct a feature vector; A classification and recognition module is used to classify the feature vector using a support vector machine algorithm to generate a temporalis muscle state classification result; The real-time feedback and evaluation module is used to evaluate the molar status according to the status classification results, and generate and send the molar status monitoring results.

2. A molar signal monitoring system based on optimized variational mode decomposition according to claim 1, characterized in that: The signal acquisition module comprises an ADS1299 chip and an ESP32 microcontroller, wherein the ADS1299 chip is used to collect the surface electromyographic signal of the temporalis muscle, and the ESP32 microcontroller is used to manage the sampling, data transmission and control of the ADS1299 chip.

3. A molar signal monitoring system based on optimized variational mode decomposition according to claim 2, characterized in that: The signal acquisition module also includes a hardware filter, and the hardware filter is connected to the ADS1299 chip.

4. The molar signal monitoring system based on optimized variational mode decomposition according to claim 1, characterized in that: The signal preprocessing module includes a denoising submodule, a normalization processing submodule and an envelope calibration submodule, wherein: The denoising submodule is used to filter and denoise the surface electromyographic signal of the temporalis muscle, remove power frequency interference, low-frequency artifacts, high-frequency noise, baseline drift and abnormal spikes; The normalization processing submodule is used to eliminate the signal amplitude difference and dimension influence of the surface electromyographic signal of the temporalis muscle and normalize all signals to a unified range; The envelope calibration submodule is used to extract the envelope of the temporalis surface electromyographic signal by the moving average method, and to identify the activity start and end points of the temporalis surface electromyographic signal by the adaptive threshold method to achieve envelope calibration.

5. The molar signal monitoring system based on optimized variational mode decomposition according to claim 1, characterized in that: The GWO-VMD algorithm module optimizes the key parameters of VMD by combining the grey wolf optimization algorithm GWO and the variational mode decomposition algorithm VMD, and decomposes the surface electromyographic signal of the target temporalis muscle based on the optimized key parameters of VMD; the key parameters include the mode number K and the penalty factor α.

6. The molar signal monitoring system based on optimized variational mode decomposition according to claim 1, characterized in that: The time domain features include root mean square, mean absolute value, zero crossing rate, difference mean square value and mean amplitude change, and describe the energy, dynamics and frequency characteristics of the surface electromyographic signal of the target temporalis muscle based on the time domain features.

7. A molar signal monitoring method based on optimized variational mode decomposition, characterized in that: The following steps are involved: Collect surface electromyographic signals of temporalis muscle; Preprocessing and envelope calibration of the temporalis muscle surface electromyographic signal to obtain the target temporalis muscle surface electromyographic signal; The GWO-VMD algorithm is used to perform multimodal feature decomposition on the surface electromyographic signal of the target temporalis muscle to obtain multiple modal components. Extract time domain features and frequency domain features from multiple modal components to construct feature vectors; The support vector machine algorithm is used to classify the feature vectors and generate the temporalis muscle status classification results; The molar status is evaluated based on the status classification results, and the molar status monitoring results are generated and sent.

8. The method for monitoring teeth molar signals based on optimized variational mode decomposition according to claim 7, characterized in that: The method for performing multimodal feature decomposition on the surface electromyographic signal of the target temporalis muscle by using the GWO-VMD algorithm comprises the following steps: Combine the Grey Wolf Optimization Algorithm (GWO) and the Variational Mode Decomposition (VMD) algorithm to optimize the key parameters of VMD. The multimodal feature decomposition of the target temporalis muscle surface electromyographic signal is performed based on the optimized key parameters of VMD.

9. An electronic device, characterized in that: include: A memory for storing one or more programs; processor; When the one or more programs are executed by the processor, the method according to any one of claims 7 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 7 to 8 is implemented.

Images