Muscle hardness detection and fascia gun hitting parameter intelligent matching method and system

Abstract

The application relates to the technical field of intelligent fitness equipment and sports rehabilitation, and provides a muscle hardness detection and intelligent matching method and system for percussion parameters of a fascia gun. The method comprises the following steps: obtaining an original hardness detection signal of a target muscle region in a static pressing state; performing noise reduction filtering processing on the original hardness detection signal to extract a characteristic parameter reflecting muscle hardness; inputting the characteristic parameter into a preset intelligent matching algorithm model, the intelligent matching algorithm model being trained in advance through a machine learning method, and outputting a percussion parameter combination suitable for the fascia gun; controlling the fascia gun to perform a percussion operation according to the percussion parameter combination, and collecting feedback signals of the target muscle region in real time during the percussion process, wherein the feedback signals comprise muscle vibration amplitude changes, surface electromyogram signal changes or real-time pressing hardness data; and inputting the feedback signals into the intelligent matching algorithm model to dynamically adjust the percussion parameter combination. The application provides a new idea for the research and development of intelligent muscle relaxation equipment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent fitness equipment and sports rehabilitation, and particularly relates to a muscle hardness detection and intelligent matching method and system for fascia gun hitting parameters. BACKGROUND

[0002] As a device that relieves fatigue and promotes blood circulation by high-frequency vibration hitting muscles, the fascia gun has been widely used in sports rehabilitation, fitness relaxation and other scenarios. The hitting parameters (such as frequency, force, amplitude, etc.) of the existing fascia gun are usually manually adjusted by the user or run based on fixed preset modes (such as "relaxation mode" and "deep hitting mode"). However, such adjustment methods have significant defects:

[0003] Lack of personalized adaptation: the muscle hardness and body basic data (such as age, weight, muscle mass) of different users differ significantly, and the muscle state of the same user changes dynamically in different scenarios (such as before / after exercise, static / fatigue state). In the prior art, the fascia gun cannot accurately perceive the real-time hardness of the muscle, and only relies on the user's subjective experience to adjust the parameters, which easily leads to poor hitting effect (such as insufficient force to relax deeply, and high frequency to cause muscle damage).

[0004] Single detection means: the traditional scheme only obtains limited data through a pressure sensor or simple contact detection, and does not fuse multi-dimensional signals such as pressure, vibration and displacement to comprehensively represent the muscle hardness, elasticity and viscosity characteristics, so the detection accuracy is insufficient.

[0005] Lack of dynamic feedback mechanism: the existing fascia gun cannot collect feedback signals (such as muscle electrical signals and vibration attenuation characteristics) of muscle state changes in real time during the hitting process, and it is difficult to dynamically adjust the parameters according to the real-time response of the muscle, which leads to the disconnection between the parameters and the muscle state during the hitting process, and it is difficult to balance safety and comfort.

[0006] Although there are a few schemes in the prior art that involve sensor-assisted adjustment, they all stop at simple processing of a single signal (such as pressure), and do not form a complete closed loop of "multi-dimensional detection-intelligent modeling-dynamic matching". For example, some devices only control the upper limit of the hitting force through a pressure sensor, without combining vibration, displacement and other signals to quantify muscle hardness characteristics; or rely on a preset empirical formula to match the parameters, without constructing a personalized model through machine learning, and cannot adapt to the complex differences of human muscles. SUMMARY

[0007] The application provides a muscle hardness detection and fascia gun hitting parameter intelligent matching method and system, aiming to solve the problem that in the prior art, there are few schemes involving sensor-assisted adjustment, but all of them are limited to simple processing of a single signal (such as pressure), and have not formed a complete closed loop of “multi-dimensional detection-intelligent modeling-dynamic matching”. For example, some devices only control the upper limit of hitting force through a pressure sensor, without quantifying muscle hardness characteristics in combination with signals such as vibration and displacement; or rely on a pre-set empirical formula to match parameters, without constructing an individualized model through machine learning, and cannot adapt to complex human muscle differences and other problems.

[0008] In a first aspect, the embodiments of the application provide a muscle hardness detection and fascia gun hitting parameter intelligent matching method, applied to a fascia gun; the method comprises:

[0009] A pressure sensor, a vibration sensor or a displacement sensor arranged on the fascia gun head or an independent detection device is used to obtain an original hardness detection signal of a target muscle region in a static pressing state; the original hardness detection signal includes one or more of pressure displacement data, vibration response data and elastic deformation data;

[0010] The original hardness detection signal is subjected to noise reduction filtering processing, and a characteristic parameter reflecting muscle hardness is extracted, the characteristic parameter including one or more of a pressure peak value, a deformation recovery time and a vibration attenuation coefficient;

[0011] The characteristic parameter is input into a pre-set intelligent matching algorithm model, the intelligent matching algorithm model being pre-trained through a machine learning method, so as to output a hitting parameter combination suitable for the fascia gun according to the muscle hardness characteristic parameter, a user's pre-set hitting preference and user's body basic data; wherein the user's body basic data includes one or more of age, weight, height, muscle mass and exercise frequency, and the hitting parameter combination includes one or more of hitting frequency, hitting force, hitting time and amplitude depth;

[0012] The fascia gun is controlled to perform a hitting operation according to the hitting parameter combination, and a feedback signal of the target muscle region is collected in real time during the hitting process, the feedback signal including muscle vibration amplitude change, surface electromyogram signal change or pressing hardness real-time data; the feedback signal is input into the intelligent matching algorithm model, and the hitting parameter combination is dynamically adjusted, so that the hitting parameter combination is matched with the real-time hardness state of the target muscle region.

[0013] In some embodiments, the original hardness detection signal of the target muscle area in the static pressing state is obtained by a pressure sensor, a vibration sensor or a displacement sensor arranged on the head of the fascia gun or a separate detection device, including: the head of the fascia gun is contacted with the target muscle area at a preset initial pressure, and during the process of maintaining the static pressing state, the pressure change data during pressing is collected by the pressure sensor, the vibration response data of the muscle tissue to the static pressure is collected by the vibration sensor, and the displacement change data of the head relative to the muscle surface is collected by the displacement sensor, forming a multi-dimensional original signal containing pressure, vibration and displacement information; wherein the preset initial pressure is a fixed value or a phased increasing pressure value within the safe pressure threshold value that the human muscle can bear.

[0014] In some embodiments, the noise reduction filtering processing of the original hardness detection signal includes: adopting a digital filtering algorithm to eliminate noise of the original hardness detection signal, and the digital filtering algorithm includes one or more combinations of Gaussian filtering, median filtering or adaptive filtering.

[0015] For different noise characteristics of the pressure signal, the vibration signal and the displacement signal, filtering parameters are respectively set to retain low-frequency trend components and high-frequency characteristic components reflecting muscle elasticity in the signal, and remove environmental vibration noise, sensor zero drift noise or human respiratory motion interference signal.

[0016] In some embodiments, the intelligent matching algorithm model is trained in advance by a machine learning method, including: constructing a training data set by using the historical collected muscle hardness detection data, the corresponding human body basic data and the actual hitting effect feedback data, and the training data set contains the mapping relationship between the muscle hardness characteristics of different age, weight and muscle mass populations and the adaptive hitting parameters; an initial model is trained by using a supervised learning algorithm, the supervised learning algorithm includes one or more of a neural network algorithm, a random forest algorithm or a support vector machine algorithm, and the error between the hitting parameter combination output by the model and the actual effective hitting parameter is less than a preset precision threshold by iteratively optimizing the model parameters.

[0017] In some embodiments, the hitting parameter combination of the adaptive fascia gun is output according to the muscle hardness characteristic parameters, the user preset hitting preference and the user body basic data, including: the intelligent matching algorithm model determines a basic range of hitting parameters based on the muscle hardness characteristic parameters first, and then adjusts the basic parameter range by a weight allocation algorithm to generate an individualized hitting parameter combination in combination with the user preset hitting preference and the user body basic data; wherein the weight allocation algorithm dynamically optimizes the influence weight of each input factor according to the historical use data of the user.

[0018] In some embodiments, the control of the fascia gun according to the combination of the hitting parameters includes: driving the motor and the hitting part of the fascia gun according to the hitting frequency, the hitting force, and the amplitude depth in the combination of the hitting parameters; before the hitting operation starts, a pre-hitting is performed at a preset low frequency and low force to obtain an initial feedback signal of the muscle to verify the safety of the parameters; during the hitting process, the motor speed, the battery power, and the mechanical state of the hitting part are monitored in real time, and when an abnormal vibration or overload signal is detected, a parameter protection mechanism is automatically triggered to reduce the hitting force or pause the hitting.

[0019] In some embodiments, the feedback signal of the target muscle area is collected in real time during the hitting process, including: during the intermittent period or the static stage of a single hitting cycle, the real-time pressure change, vibration attenuation characteristics, or muscle electrical signal amplitude change on the muscle surface are collected through the pressure sensor, the vibration sensor, or the additional surface electromyography sensor; the corresponding collection process is triggered synchronously with the hitting action to avoid the interference of the hitting vibration on the collection signal and ensure the accuracy of the feedback signal.

[0020] In some embodiments, the feedback signal is input into the intelligent matching algorithm model to dynamically adjust the combination of the hitting parameters, including: the feature difference value between the feedback signal and the initial detection signal before hitting is calculated, and when the difference value exceeds a preset dynamic adjustment threshold, a parameter adjustment mechanism is triggered; the intelligent matching algorithm model adjusts the hitting frequency, the force, or the amplitude in a one-way or two-way fine-tuning manner according to the real-time feedback of the muscle hardness change trend on the basis of the current combination of the hitting parameters, the adjusted parameter combination needs to meet the human muscle safe hitting parameter range, and the continuity of the hitting operation is maintained during the adjustment process.

[0021] In some embodiments, the feature parameters reflecting the muscle hardness are extracted, including: the pressure peak value when the pressure reaches a stable value and the deformation recovery time required for the muscle deformation to recover to 80% of the initial state after the pressure is removed are extracted from the denoised pressure-displacement signal; the vibration attenuation coefficient required for the vibration amplitude to decay to 30% of the initial value is extracted from the vibration response signal; the extraction process of the feature parameters includes signal peak detection, exponential decay curve fitting, or time window integral calculation to quantify the hardness, elasticity, and viscosity characteristics of the muscle tissue.

[0022] In a second aspect, the present application provides a muscle hardness detection and fascia gun hitting parameter intelligent matching system applied to a fascia gun, the system comprising:

[0023] The signal acquisition unit is configured to acquire an original hardness detection signal of a target muscle region in a static pressing state through a pressure sensor, a vibration sensor or a displacement sensor arranged on a head of the fascia gun or a separate detection device; the original hardness detection signal includes one or more of pressure displacement data, vibration response data and elastic deformation data;

[0024] The filtering processing unit is configured to perform noise reduction filtering processing on the original hardness detection signal to extract a characteristic parameter reflecting muscle hardness, and the characteristic parameter includes one or more of a pressure peak value, a deformation recovery time and a vibration attenuation coefficient;

[0025] The parameter input unit is configured to input the characteristic parameter into a preset intelligent matching algorithm model, and the intelligent matching algorithm model is trained in advance by a machine learning method to output a hitting parameter combination of the fascia gun according to muscle hardness characteristic parameters, user preset hitting preferences and user body basic data; wherein the user body basic data includes one or more of age, weight, height, muscle mass and exercise frequency, and the hitting parameter combination includes one or more of hitting frequency, hitting strength, hitting time and amplitude depth;

[0026] The parameter adjustment unit is configured to control the fascia gun to perform a hitting operation according to the hitting parameter combination, and to acquire a feedback signal of the target muscle region in real time during the hitting process, wherein the feedback signal includes muscle vibration amplitude change, surface electromyogram signal change or pressing hardness real-time data; the feedback signal is input into the intelligent matching algorithm model to dynamically adjust the hitting parameter combination, so that the hitting parameter combination is matched with the real-time hardness state of the target muscle region.

[0027] The muscle hardness detection and fascia gun hitting parameter intelligent matching method and system provided by the embodiment of the application acquire muscle response signals under static pressing through multiple sensors such as pressure, vibration and displacement, extract multiple-dimensional characteristic parameters such as pressure peak value, deformation recovery time and vibration attenuation coefficient, accurately quantify muscle hardness and elastic properties, and break through the limitation of traditional single signal detection. An intelligent matching model based on machine learning is constructed, muscle hardness characteristics, user body data and hitting preferences are fused, personalized parameter combinations are output, and parameters are dynamically adjusted through real-time feedback signals during the hitting process to form a closed-loop control of “detection-modeling-execution-feedback”, and real-time intelligent adaptation of fascia gun parameters and muscle states is realized for the first time. Through pre-hitting verification, abnormal state monitoring and parameter protection mechanism, the hitting safety is ensured while the personalized adjustment is realized, and the defects of “one-size-fits-all” or subjective judgment in the prior art are solved.

[0028] The application generates hitting parameters conforming to individual differences based on multi-dimensional muscle hardness characteristics and user body data, avoids the blindness of traditional manual adjustment, and improves muscle relaxation effect. The feedback signal in the hitting process is used to optimize the parameters in real time, so that the hitting force, frequency and the like always match the real-time state of the muscle (such as automatically increasing the amplitude when the muscle is stiff after exercise, and intelligently reducing the force when the fatigued muscle is pressed too much), which significantly improves the comfort and safety of use. The user does not need to manually adjust the parameters frequently, and the device automatically completes detection, modeling and dynamic control, reduces the use threshold, and adapts to the differentiated needs of different exercise scenarios (such as pre-exercise activation, post-exercise recovery, and daily relaxation). Through pre-hitting verification, abnormal vibration monitoring and parameter protection mechanism, the risk of muscle injury caused by improper parameters is effectively avoided, and the application population of the device is widened (such as the elderly and exercise beginners).

[0029] In summary, the application fills the technical gap in the field of integrated control of "accurate detection-intelligent matching-dynamic feedback" for existing fascia guns, and provides a new idea for the research and development of intelligent muscle relaxation devices.

[0030] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the application. BRIEF DESCRIPTION OF DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings needed in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0032] Figure 1 is a step schematic flow chart of a muscle hardness detection and fascia gun hitting parameter intelligent matching method provided by an embodiment of the application;

[0033] Figure 2 is a structural schematic diagram of a fascia gun provided by an embodiment of the application;

[0034] Figure 3 is a structural schematic block diagram of a muscle hardness detection and fascia gun hitting parameter intelligent matching system provided by an embodiment of the application;

[0035] Figure 4 is a structural schematic block diagram of a fascia gun provided by an embodiment of the application.

[0036] It should be understood that the above general description and the following detailed description are only exemplary and explanatory, and cannot limit the application. DETAILED DESCRIPTION

[0037] With reference to the drawings of the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are some of the embodiments of the present application, but not all of the embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts are within the scope of the present application.

[0038] The flowcharts shown in the drawings are only illustrative, and do not necessarily include all contents and operations / steps, nor are they necessarily executed in the order described. For example, some operations / steps can be further decomposed, combined or partially merged, so the actual execution order can be changed according to the actual situation.

[0039] It should be understood that, in order to clearly describe the technical solutions of the embodiments of the present application, in the embodiments of the present application, the terms "first", "second", etc. are used to distinguish the same items or similar items with basically the same functions and effects. Those skilled in the art can understand that the terms "first", "second", etc. do not limit the quantity and execution order, and the terms "first", "second", etc. also do not necessarily mean different.

[0040] It should be understood that the terms used in the present application are only for the purpose of describing specific embodiments and do not intend to limit the present application. As used in the present application specification and the appended claims, unless otherwise clear from the context, the singular forms "a", "an" and "the" are intended to include the plural forms.

[0041] It should also be understood that the term "and / or" used in the present application specification and the appended claims means any combination of one or more of the associated listed items and all possible combinations, and includes these combinations.

[0042] As a device that relieves fatigue and promotes blood circulation by high-frequency vibration hitting muscles, the fascia gun has been widely used in sports rehabilitation, fitness relaxation and other scenarios. The hitting parameters (such as frequency, force, amplitude, etc.) of the existing fascia gun usually depend on manual adjustment by the user, or run based on fixed preset modes (such as "relaxation mode" and "deep hitting mode"). However, such adjustment methods have significant defects:

[0043] Lack of personalized adaptation: the muscle hardness of different users, body basic data (such as age, weight, muscle mass) differ significantly, and the muscle state of the same user changes dynamically in different scenarios (such as before / after exercise, static / fatigue state). In the prior art, the fascia gun cannot accurately perceive the real-time hardness of the muscle, and only relies on the user's subjective experience to adjust the parameters, which easily leads to poor hitting effect (such as insufficient force to relax deeply, and high frequency to cause muscle damage).

[0044] Single detection means: traditional solutions only obtain limited data through pressure sensors or simple contact detection, without fusing multi-dimensional signals such as pressure, vibration, and displacement to comprehensively represent muscle hardness, elasticity, and viscosity characteristics, resulting in insufficient detection accuracy.

[0045] Lack of dynamic feedback mechanism: existing fascia guns cannot collect real-time feedback signals (such as muscle electrical signals and vibration attenuation characteristics) of muscle state changes during the hitting process, making it difficult to dynamically adjust parameters according to real-time muscle responses, resulting in a disconnection between parameters and muscle states during the hitting process, and making it difficult to balance safety and comfort.

[0046] Although there are a few existing technologies involving sensor-assisted adjustment, they all stop at simple processing of single signals (such as pressure) and do not form a complete closed loop of "multi-dimensional detection - intelligent modeling - dynamic matching". For example, some devices only control the upper limit of hitting force through pressure sensors without combining vibration, displacement, and other signals to quantify muscle hardness characteristics; or rely on pre-set empirical formulas to match parameters without constructing personalized models through machine learning, making it difficult to adapt to complex human muscle differences.

[0047] To solve the above problems, please refer to Figure 1 , the embodiments of the present application provide a muscle hardness detection and fascia gun hitting parameter intelligent matching method, which is applied to a fascia gun as shown in Figure 2 . At the same time, it should be noted that the method provided by the present application involves each information extracted under the authorization of the relevant user and in accordance with the relevant provisions, which does not infringe on the privacy of the user.

[0048] The muscle hardness detection and fascia gun hitting parameter intelligent matching method provided includes steps S101 to S104. Details are as follows:

[0049] Step S101. Through the pressure sensor, vibration sensor, or displacement sensor arranged on the fascia gun head or independent detection device, the original hardness detection signal of the target muscle area in the static pressing state is obtained; the original hardness detection signal includes one or more of pressure displacement data, vibration response data, and elastic deformation data.

[0050] Specifically, through the multi-modal sensor integrated in the fascia gun head or independent detection device, the physical signal of the target muscle area in the static pressing state is obtained, covering the original data of three dimensions of pressure, vibration, and displacement, which is used to comprehensively represent the mechanical properties (hardness, elasticity, and viscosity) of the muscle.

[0051] The sensor configuration includes: a pressure sensor, using a piezoresistive or capacitive sensor, embedded in the contact surface of the gun head, which collects the pressure value changes on the muscle surface in real time when pressed, forming a "pressure-displacement curve" (pressure change data with pressing depth). A vibration sensor, integrated with a three-axis accelerometer or a piezoelectric ceramic sensor, detects muscle vibration response signals (such as vibration frequency and amplitude attenuation process) during or after pressing.

[0052] The detection process includes: after the user triggers the detection mode, the fascia gun contacts the target muscle with a constant initial pressure (such as 5-10N), gradually applies pressure to the preset threshold (such as 50-100N), and synchronously collects pressure-displacement data during the process; after a short pressing state, the vibration sensor captures the muscle vibration attenuation signal, and the displacement sensor records the deformation recovery process.

[0053] Breaking through the limitations of traditional single pressure detection, through the coordinated collection of pressure, vibration, and displacement signals, the physical properties of muscle hardness (pressure peak), elasticity (deformation recovery ability), and viscosity (vibration attenuation characteristics) are comprehensively covered, providing rich data sources for subsequent feature extraction. The sensors are integrated in the gun head or independent devices, without the need for additional accessories, and the user only needs to naturally press to complete the detection, making the operation convenient and consistent with the actual use scenario.

[0054] Step S102. The original hardness detection signal is subjected to noise reduction filtering processing, and a feature parameter reflecting muscle hardness is extracted, which includes one or more of pressure peak, deformation recovery time, and vibration attenuation coefficient.

[0055] Specifically, by performing noise reduction filtering processing on the original signal, environmental noise (such as device vibration and user movement interference) is removed, key feature parameters that can directly reflect muscle hardness are extracted, and a feature vector quantifying muscle state is constructed.

[0056] Noise reduction filtering uses digital signal processing technology to process pressure, vibration, and displacement signals respectively:

[0057] The pressure-displacement signal is filtered by sliding average filtering or median filtering to remove high-frequency noise and retain the trend curve of pressure change with displacement;

[0058] The vibration response signal is filtered by a band-pass filter (such as a Butterworth filter) to remove low-frequency motion noise and high-frequency electromagnetic interference, and to retain the effective frequency band (such as 10-100Hz) of muscle vibration;

[0059] The elastic deformation data is differentiated to calculate the deformation velocity, and the stress-strain curve is fitted with pressure data.

[0060] Characteristic parameter extraction: pressure peak (P_max): the maximum pressure value measured during the pressing process, reflecting the initial muscle stiffness; deformation recovery time (T_rec): the time required for the muscle deformation to recover to 80% of the initial state after the pressure is completely released, representing muscle elasticity; vibration attenuation coefficient (a): by fitting the attenuation curve of the vibration signal (such as the exponential decay model A(t)=A0e^(-αt)), the attenuation factor a is extracted, reflecting muscle viscosity (the larger a is, the faster the attenuation is, and the muscle is harder); elastic modulus (E): according to Hooke's law, the equivalent elastic modulus of the muscle is calculated by the slope (ΔF / Δx) of the pressure-displacement curve, quantifying the hardness.

[0061] By using targeted filtering algorithms to improve signal quality, the reliability of the characteristic parameters is ensured, and detection errors caused by environmental interference are avoided; the extracted characteristic parameters directly correspond to the mechanical properties of the muscle (hardness, elasticity, viscosity), providing input variables with strong explainability for subsequent intelligent matching, which is better than the traditional single pressure signal fuzzy representation.

[0062] Step S103. Input the characteristic parameters into a pre-set intelligent matching algorithm model, which is pre-trained by a machine learning method to output an adapted percussion parameter combination of the fascia gun according to the muscle hardness characteristic parameters, the user's pre-set percussion preference, and the user's body basic data; wherein the user's body basic data includes one or more of age, weight, height, muscle mass, and exercise frequency, and the percussion parameter combination includes one or more of percussion frequency, percussion force, percussion time, and amplitude depth.

[0063] Specifically, by using a pre-trained machine learning model, the muscle characteristic parameters, user basic data (age, weight, etc.), and percussion preference (such as "gentle relaxation" and "deep stimulation") are input to output an adapted percussion parameter combination (frequency, force, amplitude, etc.), realizing personalized parameter configuration.

[0064] Model architecture: supervised learning algorithms such as gradient boosting trees (XGBoost), multi-layer perceptron (MLP), or convolutional neural networks (CNN, suitable for vibration signal features) are used to construct a "feature input-parameter output" mapping relationship; training data sources: collect characteristic parameters, subjective comfort parameters (user feedback optimal experience parameters), and physiological indicators (such as muscle electrical signal change amplitude) of different users (covering age, weight, and muscle mass differences) in different muscle states (before / after exercise, fatigue / relaxation), forming a multi-label training set.

[0065] Input-output definition: Input layer: Muscle characteristic parameters: pressure peak, deformation recovery time, vibration attenuation coefficient, elastic modulus; User basic data: age, weight, height, muscle mass (obtained by body fat scale or bioelectrical impedance method), exercise frequency; Strike preference: mode selected by user interface (such as 1-5 level intensity preference). Output layer: Strike parameter combination: frequency (such as 1200-3200 rpm), force (motor output power), amplitude depth (5-16 mm), single strike time (recommended not to exceed 10 minutes).

[0066] Model optimization: Introduce regularization techniques (such as L2 regularization) to avoid overfitting, optimize hyperparameters through cross-validation; Support online update: When the user accumulates enough usage data, the model can automatically incrementally learn, and optimize the accuracy of personalized matching.

[0067] Break the limitations of traditional "fixed preset mode", combine individual differences (age, muscle mass) and real-time muscle state (characterized by characteristic parameters), generate customized parameters, solve the problem of "insufficient force" or "excessive striking"; In addition to muscle physical properties, long-term physical data of users (such as exercise frequency reflecting muscle habit strength) and subjective preferences are included, balancing scientificity and user experience, improving comfort and safety.

[0068] Step S104 controls the fascia gun to perform a striking operation according to the strike parameter combination, and real-time feedback signals of the target muscle region are collected during the striking process, including muscle vibration amplitude changes, surface electromyographic signal changes, or real-time data of pressing hardness; The feedback signal is input into the intelligent matching algorithm model, and the strike parameter combination is dynamically adjusted to keep the strike parameter combination matched with the real-time hardness state of the target muscle region.

[0069] Specifically, during the fascia gun striking process, real-time feedback signals of muscle state changes (such as electromyographic signals, vibration amplitudes) are collected, input into the intelligent model to dynamically adjust parameters, forming a "detection-striking-feedback-optimization" closed loop to ensure that the parameters match the real-time state of the muscle.

[0070] Real-time feedback signal collection includes: Muscle vibration amplitude change: Detect the response amplitude of the muscle during striking through the gun head vibration sensor, a decrease in amplitude may indicate muscle relaxation, and the force needs to be reduced; Surface electromyographic signal (sEMG): Integrate dry electrode sensors to collect electromyographic signals of the target muscle, and use a decrease (muscle relaxation) or an increase (excessive stimulation) in electromyographic amplitude as a basis for adjustment; Real-time data of pressing hardness: Temporarily trigger the pressure sensor at a short interval (such as every 2 seconds) during striking to quickly detect the current change in muscle hardness.

[0071] The dynamic adjustment strategy includes establishing a feedback threshold: if the electromyographic amplitude decreases by more than 20%, the hitting force is automatically reduced by 10%; if the vibration amplitude continues to rise (muscle tension), the frequency is gradually increased to the user's tolerance limit; the real-time feedback signal is input into a pre-trained model, and the adjusted parameters (such as frequency ± 200 rpm and force ± 5%) are output, and the motor parameters are smoothly adjusted through the PID control algorithm to avoid sudden changes that cause discomfort.

[0072] Unlike traditional "one-time parameter setting", the hitting parameters are dynamically corrected in real time to adapt to the state changes of the muscle during hitting (such as reduced hardness after fatigue relief, requiring reduced force), avoiding disconnection between parameters and actual state; through electromyographic signal monitoring to prevent excessive stimulation (such as triggering a protection mechanism when the electromyographic signal abnormally rises), and dynamically optimizing the force and frequency in combination with the vibration response to improve the relaxation effect while reducing the risk of muscle damage.

[0073] Steps S101-S104 solve the core problem of traditional fascia guns "rough parameter adjustment and lack of real-time adaptation" through a complete closed loop of "multi-dimensional detection (S101) - feature quantization (S102) - intelligent modeling (S103) - dynamic feedback (S104)", realizing the upgrade from "user passive adjustment" to "device active intelligent adaptation", significantly improving the hitting effect, safety and personalized experience, especially suitable for high-precision control scenarios such as sports rehabilitation and fatigue recovery.

[0074] In some embodiments, the original hardness detection signal of the target muscle area in the static pressing state is obtained by setting a pressure sensor, a vibration sensor or a displacement sensor on the fascia gun head or a separate detection device, including: contacting the fascia gun head with the target muscle area at a preset initial pressure, and during the static pressing state, collecting pressure change data during pressing by the pressure sensor, collecting vibration response data of muscle tissue to static pressure by the vibration sensor, and collecting displacement change data of the gun head relative to the muscle surface by the displacement sensor, forming a multi-dimensional original signal containing pressure, vibration and displacement information; wherein the preset initial pressure is a fixed value or a phased incremental pressure value within the safe pressure threshold that the human muscle can withstand.

[0075] In the static pressing state, the pressure, vibration and displacement three types of original signals are collected by the multi-sensor cooperation of the fascia gun head or the independent device, wherein the preset initial pressure (within the safety threshold) is used in the pressing process, supporting fixed pressure or phased incremental pressure, to ensure that the signal covers the response characteristics of the muscle under different stress levels.

[0076] Sensor integration and contact mode: pressure sensor is embedded in the center area of the contact surface of the gun head, vibration sensor (such as MEMS accelerometer) is fixed inside the gun head to reduce motor vibration interference, displacement sensor (such as high-precision potentiometer) is connected to the telescopic rod of the gun head to measure the pressing depth.

[0077] Pressing process: fixed pressure mode: the gun head contacts the muscle with constant initial pressure (such as 30N, which is 1 / 3 of the safety upper limit verified by ergonomics), keeps still for 3-5 seconds, and synchronously collects pressure (stable value), vibration (micro-vibration under static stress), and displacement (pressing depth) data. Incremental mode in stages: starting from 10N, increasing by 10N every 1 second to 50N (upper limit of safety threshold), keeping each pressure stage for 1 second, recording the pressure-displacement curve and corresponding vibration signal at each stage.

[0078] Signal synchronous acquisition: the three types of sensors are triggered by a synchronous clock, and the sampling frequency is unified to 1000Hz, ensuring that the pressure, vibration, and displacement data are aligned on the time axis, facilitating subsequent fusion analysis.

[0079] The preset initial pressure is within the safety threshold (such as not exceeding 50% of the human muscle tolerance limit), avoiding muscle damage caused by pressing, and the incremental mode in stages adapts to the differences in muscle hardness of different users (such as hard muscles requiring greater pressure to trigger deformation). The fixed pressure mode quickly obtains basic hardness data, and the incremental mode in stages captures the nonlinear characteristics of muscles under progressive stress (such as the elastic limit point), providing more comprehensive raw data for subsequent feature extraction.

[0080] In some embodiments, the noise reduction filtering processing of the original hardness detection signal includes: adopting a digital filtering algorithm to eliminate noise of the original hardness detection signal, the digital filtering algorithm including one or more combinations of Gaussian filtering, median filtering, or adaptive filtering; setting filtering parameters respectively for different noise characteristics of pressure signals, vibration signals, and displacement signals, retaining low-frequency trend components and high-frequency characteristic components reflecting muscle elasticity in the signals, and removing environmental vibration noise, sensor zero drift noise, or human respiratory motion interference signals.

[0081] For different noise characteristics of pressure, vibration, and displacement signals, a customized digital filtering algorithm combination is adopted to remove environmental noise (such as motor vibration and respiratory motion) and retain effective signal components reflecting muscle elasticity (high-frequency components) and hardness (low-frequency trend).

[0082] Signal classification filtering processing: pressure signal: first remove burst pulse noise (such as user involuntary shaking) through median filtering (window size 5 points), then use adaptive filtering (such as LMS algorithm) to dynamically eliminate sensor zero drift (baseline offset caused by long-term use), and retain low-frequency pressure change trend (main component reflecting muscle hardness).

[0083] Vibration signal: A 5-200 Hz band-pass Gaussian filter is used to filter out respiratory motion interference below 5 Hz and motor electromagnetic noise above 200 Hz, retain the effective frequency band of muscle vibration (10-100 Hz), and extract the instantaneous amplitude and frequency through Hilbert transform.

[0084] Displacement signal: The displacement curve is smoothed using a moving average filter (window size 10 points), combined with a human kinematics model (such as body fluctuation caused by breathing), and the low-frequency motion interference is predicted and compensated through Kalman filtering to ensure that the displacement data accurately reflects the gun head pressing depth.

[0085] Parameter adaptive adjustment dynamically adjusts the filtering parameters according to the signal noise intensity (such as increasing the median filter window when the noise is large), and triggers the filtering strategy switching through real-time calculation of signal-to-noise ratio (SNR).

[0086] Avoiding feature loss caused by "one-size-fits-all" filtering, such as retaining low-frequency trends (hardness-related) while removing high-frequency jitter in pressure signal, and retaining high-frequency attenuation characteristics reflecting viscosity in vibration signal, improving the reliability of subsequent feature parameters. Adapt to complex use scenarios (such as body shaking when standing), ensure the stability of signal processing in different environments, and lay a foundation for accurate feature extraction.

[0087] In some embodiments, the intelligent matching algorithm model is pre-trained by machine learning method, including: using the historical collected muscle hardness detection data, the corresponding human body basic data, and the actual hitting effect feedback data to construct a training data set, the training data set contains the mapping relationship between muscle hardness characteristics of different age, weight, muscle mass population and adaptive hitting parameters; using a supervised learning algorithm to train the initial model, the supervised learning algorithm includes one or more of neural network algorithm, random forest algorithm or support vector machine algorithm, and by iteratively optimizing the model parameters, the error between the hitting parameter combination output by the model and the actual effective hitting parameter is less than the preset precision threshold.

[0088] Using the historical collected multi-dimensional data set (muscle characteristics, body data, hitting effect) to train the intelligent model, and through the supervised learning algorithm to establish the mapping relationship between "input features and hitting parameters", and iteratively optimize the output parameter error to be lower than the preset threshold (such as frequency error ± 50 rpm, force error ± 5%).

[0089] Training dataset construction: Data sources: Muscle hardness data: obtained through static pressing detection of 1000+ users (age 18-60, muscle mass 15%-40%), containing pressure peak (50-150N), deformation recovery time (0.5-3 seconds) and other characteristics. Basic body data: age, weight, height, muscle mass (obtained by InBody body measurement instrument), exercise frequency (reported by users, divided into 0-3 times / week, 4-7 times / week). Effect feedback data: user subjective score (1-5 points for comfort) after use, muscle electrical signal change rate (objective index of relaxation effect). Data labeling: "user feedback best parameters + muscle electrical signal optimal change" as the label, construct a multi-output regression problem (frequency, intensity, amplitude, etc. as output variables).

[0090] Model training process: an integrated model combining XGBoost and neural network (MLP): XGBoost processes structured features (age, pressure peak), MLP processes signal-derived features (vibration attenuation coefficient), and outputs through weighted averaging fusion. Loss function: mean square error (MSE) + absolute error (MAE), avoid the influence of outliers, 1000 iterations, early stopping mechanism (stop if the validation set error does not decrease for 50 consecutive times).

[0091] Based on real population data training, covering a wide range of individual differences, the model output parameters fit different users' physiological characteristics (such as old people with poor muscle elasticity, automatically reducing amplitude), avoiding the limitations of traditional empirical formulas. XGBoost provides feature importance analysis (such as muscle mass has the highest weight on intensity), which is convenient for subsequent strategy optimization; neural network captures nonlinear relationships (such as the complex mapping of vibration attenuation coefficient and frequency), improving the model's generalization ability.

[0092] In some embodiments, the output of the adaptive muscle hardness characteristic parameters, user preset hitting preferences, and user body basic data includes: the intelligent matching algorithm model first determines the basic range of hitting parameters based on the muscle hardness characteristic parameters, then combines the user preset hitting preferences and the user body basic data, and adjusts the basic parameter range through a weight distribution algorithm to generate a personalized hitting parameter combination; wherein the weight distribution algorithm dynamically optimizes the influence weight of each input factor according to user historical use data.

[0093] In the model output stage, the muscle hardness characteristics are used to determine the parameter basic range (such as increasing the lower limit of intensity if the pressure peak is high), then combined with user preferences (such as "deep mode") and body data (such as body weight affecting the upper limit of amplitude safety), and through dynamic weight distribution to generate personalized parameters.

[0094] Baseline range determination: Establish muscle characteristic-parameter mapping rule library (based on training data statistics): P_max ≤ 80N: force baseline range 20%-50%, frequency 1800-2600 rpm; P_max > 80N: force baseline range 40%-70%, frequency 2200-3200 rpm. Safety boundary constraints: amplitude depth ≤ 16 mm (mechanical limit), single hitting time ≤ 10 minutes (medical recommendation).

[0095] Weight distribution algorithm: Initial weights: muscle characteristics (0.6) > body data (0.3) > hitting preference (0.1); Dynamic adjustment: gradually increase the hitting preference weight to 0.2-0.3 based on user historical usage data (e.g., a user frequently selects "high intensity" mode), forming a personalized adjustment strategy. Example: user selects "gentle mode" (preference weight + 0.1) and body weight is 70 kg (body data weight + 0.05), then the final force = baseline force × (0.6 muscle characteristic weight + 0.35 body / preference weight).

[0096] With physiological characteristics as the core (avoiding excessive reliance on user subjective selection leading to injury), while taking into account individual preferences (such as athletes actively needing stronger stimulation), balancing safety and user experience. Optimize weights through historical data to adapt to "different scenarios for the same user" (such as preferring more intense relaxation after exercise, the model automatically remembers and adjusts), improving long-term use fit.

[0097] In some embodiments, the control of the fascia gun according to the hitting parameter combination to perform a hitting operation includes: driving the motor and hitting components of the fascia gun to move according to the hitting frequency, hitting intensity, and amplitude depth parameters in the hitting parameter combination; before the hitting operation starts, pre-hitting is performed with preset low frequency and low intensity parameters to obtain muscle initial feedback signals to verify parameter safety; during the hitting process, the motor speed, battery power, and mechanical state of the hitting component are monitored in real time, and when abnormal vibration or overload signals are detected, a parameter protection mechanism is automatically triggered to reduce the hitting intensity or pause the hitting.

[0098] Before executing the hitting parameters, safety is verified through pre-hitting; during the hitting process, the motor and mechanical state are monitored in real time, and the protection mechanism is triggered when abnormal, ensuring that the device operates in a safe range.

[0099] Pre-hitting verification: Before formal hitting, run for 5 seconds with low parameters (frequency 1000 rpm, intensity 10%, amplitude 5 mm), collect muscle vibration signals during pre-hitting: if the vibration amplitude exceeds the safety threshold (e.g., 150% of the initial detection value, indicating excessive muscle tension), automatically reduce the intensity by 10% and re-verify; if the signal is normal, switch to the target parameters to start formal hitting.

[0100] Real-time monitoring and protection: motor monitoring: real-time monitoring of rotation speed (error ±2%) through Hall sensors, detection of load current through current sensors, overload determination when exceeding 120% of rated current; mechanical state: accelerometer detects abnormal vibration of gun head (such as sudden change in frequency ±500 rpm), determines that the hitting part is deviated or muscle spasm; protection mechanism: immediately reduce 20% intensity when overloaded, pause hitting and prompt user to adjust position when abnormal vibration occurs, restart with 50% intensity after recovery.

[0101] Pre-hitting mechanism eliminates the risk of parameter mismatch (such as user selecting high-intensity mode by mistake), real-time monitoring prevents motor overheating, mechanical failure or muscle overstimulation, especially suitable for sensitive scenarios such as sports rehabilitation. Through dynamic adjustment rather than direct shutdown, it reduces user interruption, while ensuring the timeliness of abnormal handling and improving device reliability.

[0102] In some embodiments, the feedback signal of the target muscle area is collected in real time during the hitting process, including: during the intermittent period of the fascia gun or the static stage of a single hitting cycle, the real-time pressure change, vibration attenuation characteristics or muscle electrical signal amplitude change of the muscle surface are collected through the pressure sensor, vibration sensor or additional surface electromyography sensor; the corresponding collection process is triggered synchronously with the hitting action to avoid interference of hitting vibration on the collected signal and ensure the accuracy of the feedback signal.

[0103] During the intermittent period (such as 0.5 seconds pause every 5 seconds of hitting) or the static stage of a single cycle, the muscle feedback signal is collected synchronously to avoid interference of hitting vibration and ensure signal authenticity.

[0104] Collection timing control: intermittent collection: periodic interruption strategy is adopted, such as "5 seconds of hitting + 0.5 seconds of static" cycle, static period triggers sensor to collect pressure (real-time hardness), vibration (attenuation characteristics); periodic collection: for high-frequency hitting (such as 3200 rpm, cycle about 18.75 ms), in the static stage (about 2 ms) when the hitting part rebounds to the highest position, quickly sample the vibration sensor (avoid hitting impact noise).

[0105] Sensor expansion: additional surface electromyography sensor (dry electrode type, attached to the edge of the gun head), collects electromyography signal (sampling frequency 1000 Hz) during the static period, extracts root mean square value (RMS) through band-pass filtering (10-500 Hz) to reflect muscle activation level.

[0106] Avoid strong noise periods of hitting vibration, ensure that pressure, myoelectricity and other signals are not contaminated by mechanical movement, for example, the measurement error of myoelectricity RMS is reduced from ± 15% of non-intermittent period to ± 5%, and the credibility of feedback data is improved. The period is suitable for high-frequency hitting scene, and the real-time feedback of every 100 ms level (such as 20 times / second detection) is realized, and the rapid change of muscle state (such as relaxation immediately after hitting) is captured in time.

[0107] In some embodiments, the feedback signal is input into the intelligent matching algorithm model, and the hitting parameter combination is dynamically adjusted, including: calculating the feature difference value of the feedback signal and the initial detection signal before hitting, when the difference value exceeds the preset dynamic adjustment threshold, triggering the parameter adjustment mechanism; the intelligent matching algorithm model adjusts the hitting frequency, intensity or amplitude according to the real-time feedback muscle hardness change trend, and adjusts the hitting frequency, intensity or amplitude according to the preset step size on the basis of the current hitting parameter combination, and the adjusted parameter combination needs to meet the human muscle safe hitting parameter range, and the continuity of the hitting operation is maintained during the adjustment process.

[0108] By calculating the feature difference (such as pressure peak value drop amplitude) of the feedback signal and the initial signal, the parameter fine-tuning is triggered, and the adjustment process follows the safety range constraint and maintains the continuity of the hitting (without interrupting the operation).

[0109] Difference value calculation and threshold determination: define the key difference index: hardness change rate: (current pressure peak value - initial peak value) / initial peak value, absolute value>15% trigger adjustment; myoelectricity attenuation rate: (initial myoelectricity RMS - current RMS) / initial RMS, >20% (relaxation) or <-10% (tension) trigger adjustment. Threshold classification: slight difference (10%-15%) fine-tune frequency (±100 rpm), moderate difference (15%-25%) adjust intensity (±10%), severe difference (>25%) adjust frequency and amplitude at the same time (±2mm).

[0110] Parameter fine-tuning rules: one-way adjustment: when the hardness decreases (relaxation), fine-tune in the order of “amplitude↓→intensity↓→frequency↓” to avoid excessive relaxation; when the hardness rises (tension), fine-tune in the order of “frequency↑→intensity↑→amplitude↑” (not exceeding the upper limit of safety). Continuity guarantee: the adjustment step size is ≤5% of the basic parameter (such as intensity ±5% each time), which realizes smooth transition through PWM control motor, and avoids the user from perceiving parameter mutation.

[0111] Real-time tracking of the impact effect (such as automatically reducing the force after the hardness decreases to avoid energy waste), solving the problem of traditional equipment "fixed parameters leading to later effect decline", for example, during the muscle relaxation process after exercise, the parameters can be continuously optimized with the change of hardness. By presetting the adjustment sequence and step limit, it is ensured that the parameters are always within the medical safety range (such as the amplitude not exceeding 16mm), while avoiding frequent large adjustments affecting user experience.

[0112] In some embodiments, the feature parameters reflecting muscle hardness are extracted, including: extracting the pressure peak value when the pressure reaches a stable value from the denoised pressure-displacement signal, and the deformation recovery time required for the muscle deformation to recover to 80% of the initial state after the pressure is removed; extracting the vibration attenuation coefficient required for the vibration amplitude to decay to 30% of the initial value from the vibration response signal; the extraction process of the feature parameters includes signal peak detection, exponential decay curve fitting or time window integral calculation, to quantify the hardness, elasticity and viscosity characteristics of muscle tissue.

[0113] Three core features are extracted from the denoised signal: pressure peak value (hardness), deformation recovery time (elasticity), and vibration attenuation coefficient (viscosity), which quantify muscle mechanical properties through peak detection and curve fitting algorithms.

[0114] Pressure-displacement signal feature extraction: pressure peak value (P_max): take the maximum value in the pressure stable stage (mean value within 2-4 seconds after pressing), which reflects the maximum resistance of the muscle to pressing; deformation recovery time (T_rec): after the pressure is removed, the time for the displacement curve to fall from the maximum value to 80% of the initial position, which is calculated by threshold detection of the signal falling edge (such as the time difference corresponding to 80% of the initial displacement).

[0115] Vibration response signal feature extraction: vibration attenuation coefficient (a): exponential decay fitting (A(t)=A0*e^(-αt)) is performed on the vibration amplitude curve, and a is solved by least squares method. The larger the a value, the faster the vibration attenuation (the higher the muscle viscosity and the greater the hardness).

[0116] Signal processing tools: peak detection: local maximum value algorithm (set the maximum value within a 50ms time window) is used to avoid noise misjudgment; curve fitting: use the curve_fit function of Python SciPy library, and the fitting error R 2 >0.95 is considered as an effective feature.

[0117] The three types of features correspond to the hardness (P max), elasticity (T rec), and viscosity (a) of the muscle, respectively, forming a complete mechanical property description, which is more comprehensive than the traditional single pressure value (for example, two muscles have the same pressure peak value, but T rec is different, which indicates the difference in elasticity and requires different striking frequencies). The unified extraction method ensures that the detection data of different users and different times are comparable, facilitating model training and cross-scene analysis, such as the change in T rec before and after exercise, which can directly reflect the relaxation effect.

[0118] Referring to Figure 3 as shown, Figure 3 is a structural schematic diagram of a muscle hardness detection and fascia gun striking parameter intelligent matching system 200 provided by the embodiments of the present application. The muscle hardness detection and fascia gun striking parameter intelligent matching system 200 is used to execute the steps of the muscle hardness detection and fascia gun striking parameter intelligent matching method shown in each of the above embodiments. The muscle hardness detection and fascia gun striking parameter intelligent matching system 200 can be a single server or a server cluster, or the muscle hardness detection and fascia gun striking parameter intelligent matching system 200 can be a terminal, which can be a handheld terminal, a notebook computer, a wearable device, or a robot, etc.

[0119] As Figure 3 shown, the muscle hardness detection and fascia gun striking parameter intelligent matching system 200 includes:

[0120] The signal acquisition unit 201 is configured to acquire an original hardness detection signal of a target muscle region in a static pressing state through a pressure sensor, a vibration sensor, or a displacement sensor arranged on a fascia gun head or an independent detection device; the original hardness detection signal includes one or more of pressure displacement data, vibration response data, and elastic deformation data;

[0121] The filtering processing unit 202 is configured to perform noise reduction filtering processing on the original hardness detection signal to extract a feature parameter reflecting muscle hardness, the feature parameter including one or more of a pressure peak value, a deformation recovery time, and a vibration attenuation coefficient;

[0122] The parameter input unit 203 is configured to input the feature parameter into a preset intelligent matching algorithm model, the intelligent matching algorithm model being pre-trained by a machine learning method to output a striking parameter combination of the fascia gun according to the muscle hardness feature parameter, a user's preset striking preference, and user's body basic data; wherein the user's body basic data includes one or more of age, weight, height, muscle mass, and exercise frequency, and the striking parameter combination includes one or more of striking frequency, striking force, striking time, and amplitude depth;

[0123] The parameter adjustment unit 204 is configured to control the fascia gun to perform a hitting operation according to the hitting parameter combination, and collect a feedback signal of the target muscle region in real time during the hitting operation, the feedback signal including a muscle vibration amplitude change, a surface electromyogram signal change, or real-time data of a pressing hardness; input the feedback signal into the intelligent matching algorithm model, dynamically adjust the hitting parameter combination, and keep the hitting parameter combination matched with the real-time hardness state of the target muscle region.

[0124] In some embodiments, the original hardness detection signal of the target muscle region in a static pressing state is obtained by a pressure sensor, a vibration sensor, or a displacement sensor arranged on the fascia gun head or a separate detection device, including: contacting the fascia gun head with the target muscle region at a preset initial pressing force, collecting pressure change data during pressing by the pressure sensor, collecting vibration response data of the muscle tissue to the static pressure by the vibration sensor, collecting displacement change data of the gun head relative to the muscle surface by the displacement sensor, and forming a multi-dimensional original signal containing pressure, vibration, and displacement information; wherein the preset initial pressing force is a fixed value or a phased incremental pressure value within a safe pressure threshold range that can be borne by human muscle.

[0125] In some embodiments, the noise reduction filtering processing of the original hardness detection signal includes: performing noise elimination on the original hardness detection signal by using a digital filtering algorithm, the digital filtering algorithm including one or more combinations of Gaussian filtering, median filtering, or adaptive filtering; setting filtering parameters for different noise characteristics of the pressure signal, the vibration signal, and the displacement signal, respectively, retaining low-frequency trend components and high-frequency characteristic components reflecting muscle elasticity in the signal, and removing environmental vibration noise, sensor zero drift noise, or human respiratory motion interference signals.

[0126] In some embodiments, the intelligent matching algorithm model is trained in advance by a machine learning method, including: constructing a training data set by using historically collected muscle hardness detection data, body basic data of a corresponding human body, and actual hitting effect feedback data, the training data set containing a mapping relationship between muscle hardness characteristics of different age, weight, and muscle mass populations and adaptive hitting parameters; training an initial model by using a supervised learning algorithm, the supervised learning algorithm including one or more of a neural network algorithm, a random forest algorithm, or a support vector machine algorithm, and iteratively optimizing model parameters so that an error between hitting parameter combinations output by the model and actual effective hitting parameters is less than a preset precision threshold.

[0127] In some embodiments, the outputting of the percussion parameter combination of the muscle mass gun according to the muscle hardness characteristic parameter, the user's preset percussion preference and the user's body basic data comprises: the intelligent matching algorithm model determines a basic range of percussion parameters based on the muscle hardness characteristic parameter, and then adjusts the basic parameter range by a weight distribution algorithm to generate a personalized percussion parameter combination in combination with the user's preset percussion preference and the user's body basic data; wherein the weight distribution algorithm dynamically optimizes the influence weight of each input factor according to the user's historical use data.

[0128] In some embodiments, the controlling of the muscle mass gun to perform the percussion operation according to the percussion parameter combination comprises: driving the muscle mass gun motor and percussion component to move according to the percussion frequency, percussion strength and amplitude depth parameters in the percussion parameter combination; before the percussion operation starts, pre-percussion is performed at a preset low frequency and low strength parameter to obtain an initial feedback signal of the muscle mass to verify the parameter safety; during the percussion process, the muscle mass gun motor speed, battery power and mechanical state of the percussion component are monitored in real time, and when an abnormal vibration or overload signal is detected, a parameter protection mechanism is automatically triggered to reduce the percussion strength or pause the percussion.

[0129] In some embodiments, the real-time collection of the feedback signal of the target muscle region during the percussion process comprises: during the muscle mass gun percussion interval or the static stage of a single percussion cycle, the real-time pressure change, vibration attenuation characteristic or electromyographic signal amplitude change of the muscle surface are collected by the pressure sensor, vibration sensor or additional surface electromyographic sensor; the corresponding collection process is triggered synchronously with the percussion action to avoid the interference of the percussion vibration on the collection signal and ensure the accuracy of the feedback signal.

[0130] In some embodiments, the inputting of the feedback signal into the intelligent matching algorithm model to dynamically adjust the percussion parameter combination comprises: calculating the characteristic difference value of the feedback signal and the initial detection signal before percussion, and triggering a parameter adjustment mechanism when the difference value exceeds a preset dynamic adjustment threshold; the intelligent matching algorithm model adjusts the percussion frequency, strength or amplitude by a preset step in one direction or in both directions according to the real-time feedback muscle hardness change trend on the basis of the current percussion parameter combination, and the adjusted parameter combination needs to meet the human muscle safe percussion parameter range, and the continuity of the percussion operation is maintained during the adjustment process.

[0131] In some embodiments, the extracting the characteristic parameter reflecting muscle hardness comprises: extracting, from the denoised pressure-displacement signal, a pressure peak value when the pressure reaches a stable value, and a deformation recovery time required for muscle deformation to recover to 80% of the initial state after the pressure is removed; and extracting, from the vibration response signal, a vibration attenuation coefficient required for the vibration amplitude to decay to 30% of the initial value; the extraction process of the characteristic parameter comprises signal peak value detection, exponential decay curve fitting or time window integral calculation, so as to quantify the hardness, elasticity and viscosity characteristics of the muscle tissue.

[0132] It should be noted that, for the convenience and brevity of description, the specific working processes of the muscle hardness detection and fascia gun hitting parameter intelligent matching system and each module described above can be clearly understood by those skilled in the art, and can refer to the corresponding content in each embodiment of the muscle hardness detection and fascia gun hitting parameter intelligent matching method described above, which will not be described here.

[0133] The muscle hardness detection and fascia gun hitting parameter intelligent matching method described above can be implemented in the form of a computer program, which can run on the device as shown in Figure 3 .

[0134] Please refer to Figure 4 , Figure 4 is a structural schematic block diagram of the fascia gun provided by the embodiment of the present application. The fascia gun includes a processor, a memory and a network interface connected through a device bus, wherein the memory can include a storage medium and an internal memory.

[0135] The storage medium can store an operating device and a computer program. The computer program includes program instructions which, when executed, can cause the processor to execute any muscle hardness detection and fascia gun hitting parameter intelligent matching method.

[0136] The processor is used to provide computing and control capabilities to support the operation of the entire fascia gun.

[0137] The internal memory provides an environment for the running of the computer program in the non-volatile storage medium, which, when executed by the processor, can cause the processor to execute any muscle hardness detection and fascia gun hitting parameter intelligent matching method.

[0138] The network interface is used for network communication, such as sending assigned tasks. Those skilled in the art can understand that Figure 4 the structure shown in the figure is only a block diagram of part of the structure related to the scheme of the present application, and does not constitute a limitation on the terminal to which the scheme of the present application is applied. The specific fascia gun can include more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0139] It should be appreciated that the processor can be a central processing unit (CPU), and can also be other general-purpose processors, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0140] In one embodiment, the processor is configured to run a computer program stored in the memory to implement the following steps:

[0141] The original hardness detection signal of the target muscle area in the static pressing state is obtained by a pressure sensor, a vibration sensor or a displacement sensor arranged on the fascia gun head or a separate detection device; the original hardness detection signal includes one or more of pressure displacement data, vibration response data and elastic deformation data;

[0142] The original hardness detection signal is subjected to noise reduction filtering processing, and a characteristic parameter reflecting muscle hardness is extracted, the characteristic parameter including one or more of pressure peak value, deformation recovery time and vibration attenuation coefficient;

[0143] The characteristic parameter is input into a preset intelligent matching algorithm model, the intelligent matching algorithm model being trained in advance by a machine learning method to output a hitting parameter combination of the fascia gun according to muscle hardness characteristic parameters, user preset hitting preferences and user body basic data; the user body basic data including one or more of age, weight, height, muscle mass and exercise frequency, and the hitting parameter combination including one or more of hitting frequency, hitting strength, hitting time and amplitude depth;

[0144] The fascia gun is controlled to perform a hitting operation according to the hitting parameter combination, and a feedback signal of the target muscle area is collected in real time during the hitting process, the feedback signal including muscle vibration amplitude change, surface electromyogram signal change or pressing hardness real-time data; the feedback signal is input into the intelligent matching algorithm model, and the hitting parameter combination is dynamically adjusted so that the hitting parameter combination is matched with the real-time hardness state of the target muscle area.

[0145] In some embodiments, the original hardness detection signal of the target muscle area in the static pressing state is obtained by a pressure sensor, a vibration sensor or a displacement sensor arranged on the head of the fascia gun or a separate detection device, including: the head of the fascia gun is contacted with the target muscle area at a preset initial pressure, and during the static pressing state, the pressure change data during pressing is collected by the pressure sensor, the vibration response data of the muscle tissue to the static pressure is collected by the vibration sensor, and the displacement change data of the head relative to the muscle surface is collected by the displacement sensor, forming a multi-dimensional original signal containing pressure, vibration and displacement information; wherein the preset initial pressure is a fixed value or a phased increasing pressure value within the safe pressure threshold value that the human muscle can withstand.

[0146] In some embodiments, the noise reduction filtering processing of the original hardness detection signal includes: adopting a digital filtering algorithm to eliminate noise of the original hardness detection signal, the digital filtering algorithm including one or more combinations of Gaussian filtering, median filtering or adaptive filtering; different filtering parameters are set for the pressure signal, the vibration signal and the displacement signal according to different noise characteristics, the low-frequency trend component and the high-frequency characteristic component reflecting the muscle elasticity in the signal are retained, and the environmental vibration noise, the sensor zero drift noise or the human respiratory motion interference signal is removed.

[0147] In some embodiments, the intelligent matching algorithm model is trained in advance by a machine learning method, including: constructing a training data set by using the historical muscle hardness detection data, the corresponding human body basic data and the actual hitting effect feedback data, the training data set containing the mapping relationship between the muscle hardness characteristics of different age, weight and muscle mass populations and the adaptive hitting parameters; an initial model is trained by using a supervised learning algorithm, the supervised learning algorithm including one or more of a neural network algorithm, a random forest algorithm or a support vector machine algorithm, and the error between the hitting parameter combination output by the model and the actual effective hitting parameter is less than a preset precision threshold by iteratively optimizing the model parameters.

[0148] In some embodiments, the hitting parameter combination of the adaptive fascia gun is output according to the muscle hardness characteristic parameters, the user's preset hitting preference and the user's body basic data, including: the intelligent matching algorithm model determines a basic range of hitting parameters based on the muscle hardness characteristic parameters, and then adjusts the basic parameter range by a weight allocation algorithm to generate an individualized hitting parameter combination in combination with the user's preset hitting preference and the user's body basic data; wherein the weight allocation algorithm dynamically optimizes the influence weight of each input factor according to the user's historical use data.

[0149] In some embodiments, the controlling the fascia gun to perform the hitting operation according to the hitting parameter combination comprises: driving the motor and the hitting component to move according to the hitting frequency, the hitting force, and the amplitude depth in the hitting parameter combination; before the hitting operation starts, pre-hitting is performed at a preset low frequency and low force to obtain an initial feedback signal of the muscle to verify the safety of the parameters; during the hitting operation, the motor speed, the battery power, and the mechanical state of the hitting component are monitored in real time, and when an abnormal vibration or overload signal is detected, a parameter protection mechanism is automatically triggered to reduce the hitting force or pause the hitting.

[0150] In some embodiments, the real-time collection of the feedback signal of the target muscle region during the hitting operation comprises: during the intermittent period of the fascia gun or the static stage of a single hitting cycle, the real-time pressure change, vibration attenuation characteristics, or muscle electrical signal amplitude change of the muscle surface are collected by the pressure sensor, the vibration sensor, or an additional surface electromyography sensor; the corresponding collection process is triggered synchronously with the hitting action to avoid interference of the hitting vibration on the collection signal and ensure the accuracy of the feedback signal.

[0151] In some embodiments, the inputting of the feedback signal into the intelligent matching algorithm model to dynamically adjust the hitting parameter combination comprises: calculating the feature difference value between the feedback signal and the initial detection signal before the hitting, and when the difference value exceeds a preset dynamic adjustment threshold, triggering a parameter adjustment mechanism; the intelligent matching algorithm model adjusts the hitting frequency, force, or amplitude in a one-way or two-way fine-tuning manner according to the real-time feedback of the muscle hardness change trend on the basis of the current hitting parameter combination, and the adjusted parameter combination needs to meet the human muscle safe hitting parameter range, and the continuity of the hitting operation is maintained during the adjustment process.

[0152] In some embodiments, the extracting of the feature parameter reflecting the muscle hardness comprises: extracting the pressure peak value when the pressure reaches a stable value and the deformation recovery time required for the muscle deformation to recover to 80% of the initial state after the pressure is removed from the denoised pressure-displacement signal; and extracting the vibration attenuation coefficient required for the vibration amplitude to decay to 30% of the initial value from the vibration response signal; the extraction process of the feature parameter comprises signal peak value detection, exponential decay curve fitting, or time window integral calculation to quantify the hardness, elasticity, and viscosity characteristics of the muscle tissue.

[0153] The application also provides a computer readable storage medium storing a computer program, wherein the computer program is executed by a processor to make the processor implement the steps of the muscle hardness detection and fascia gun hitting parameter intelligent matching method provided by any one of the embodiments of the application.

[0154] The computer readable storage medium can be an internal storage unit of the fascia gun, such as a hard disk or a memory of the fascia gun. The computer readable storage medium can also be an external storage device of the fascia gun, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc.

[0155] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of various equivalent modifications or replacements within the technical range disclosed by the present application, and these modifications or replacements shall be covered in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method for intelligent matching of muscle stiffness detection and fascia gun firing parameters, characterized in that, Applied to a fascia gun; the method includes: The original hardness detection signal of the target muscle area under static pressure is obtained by using a pressure sensor, vibration sensor, or displacement sensor installed on the fascia gun head or independent detection device; the original hardness detection signal includes pressure displacement data, vibration response data, and elastic deformation data. The original hardness detection signal is subjected to noise reduction and filtering to extract characteristic parameters reflecting muscle hardness, including pressure peak value, deformation recovery time, and vibration attenuation coefficient. The feature parameters are input into a preset intelligent matching algorithm model, which is pre-trained using machine learning methods to output a combination of striking parameters adapted to the fascia gun based on muscle hardness feature parameters, the user's preset striking preferences, and the user's basic body data. This includes: the intelligent matching algorithm model determining a basic range of striking parameters based on the muscle hardness feature parameters, then adjusting the basic parameter range using a weight allocation algorithm in conjunction with the user's preset striking preferences and basic body data to generate a personalized combination of striking parameters; wherein the weight allocation algorithm dynamically optimizes the influence weight of each input factor based on the user's historical usage data; wherein the user's basic body data includes age, weight, height, muscle mass, and exercise frequency; and the striking parameter combination includes striking frequency, striking force, striking time, and amplitude depth. The fascia gun is controlled to perform striking operations according to the striking parameter combination, and feedback signals from the target muscle area are collected in real time during the striking process. The feedback signals include changes in muscle vibration amplitude, changes in surface electromyography signals, or real-time data on pressure hardness. The feedback signals are input into the intelligent matching algorithm model to dynamically adjust the striking parameter combination, including: calculating the feature difference value between the feedback signal and the initial detection signal before striking; when the difference value exceeds a preset dynamic adjustment threshold, a parameter adjustment mechanism is triggered; the intelligent matching algorithm model, based on the real-time feedback trend of muscle hardness change, makes unidirectional or bidirectional fine adjustments to the striking frequency, force, or amplitude according to a preset step size on the basis of the current striking parameter combination. The adjusted parameter combination must meet the safe striking parameter range of human muscles, and the striking operation must be maintained continuously during the adjustment process so that the striking parameter combination matches the real-time hardness state of the target muscle area.

2. The method according to claim 1, characterized in that, The process of acquiring the original hardness detection signal of the target muscle area under static pressure by means of a pressure sensor, vibration sensor, or displacement sensor installed on the fascia gun head or an independent detection device includes: The fascia gun tip is brought into contact with the target muscle area with a preset initial pressure. While maintaining a static pressing state, the pressure sensor collects pressure change data during the pressing process, the vibration sensor collects vibration response data of muscle tissue to static pressure, and the displacement sensor collects displacement change data of the gun tip relative to the muscle surface, forming a multi-dimensional raw signal containing pressure, vibration, and displacement information. The preset initial pressure is a fixed value or a gradually increasing pressure value within the safe pressure threshold range that human muscles can withstand.

3. The method according to claim 1, characterized in that, The noise reduction and filtering process for the original hardness detection signal includes: The original hardness detection signal is noise-reduced using a digital filtering algorithm, which includes one or more combinations of Gaussian filtering, median filtering, or adaptive filtering. Filtering parameters are set according to the different noise characteristics of pressure signals, vibration signals, and displacement signals. Low-frequency trend components and high-frequency characteristic components reflecting muscle elasticity are retained in the signals, while environmental vibration noise, sensor zero-point drift noise, or human respiratory movement interference signals are removed.

4. The method according to claim 1, characterized in that, The intelligent matching algorithm model is pre-trained using machine learning methods, including: A training dataset is constructed using historically collected muscle hardness detection data, corresponding human body basic data, and actual impact effect feedback data. The training dataset contains the mapping relationship between muscle hardness characteristics and appropriate impact parameters for people of different ages, weights, and muscle mass. The initial model is trained using a supervised learning algorithm, which includes one or more of neural network algorithms, random forest algorithms, or support vector machine algorithms. The model parameters are iteratively optimized so that the error between the combination of impact parameters output by the model and the actual effective impact parameters is less than a preset accuracy threshold.

5. The method according to claim 1, characterized in that, The step of controlling the fascia gun to perform the striking operation according to the striking parameter combination includes: The fascia gun motor and striking components are driven to move according to the striking frequency, striking force, and amplitude depth parameters in the striking parameter combination. Before the striking operation begins, a pre-striking is performed with preset low frequency and low force parameters to obtain initial muscle feedback signals and verify the safety of the parameters. During the striking process, the motor speed of the fascia gun, the battery level, and the mechanical status of the striking components are monitored in real time. When abnormal vibration or overload signals are detected, the parameter protection mechanism is automatically triggered to reduce the striking force or pause the striking.

6. The method according to claim 1, characterized in that, The real-time acquisition of feedback signals from the target muscle area during the striking process includes: During the intervals between fascia gun strikes or the static phase of a single strike cycle, the pressure sensor, vibration sensor, or additional surface electromyography (EMG) sensor is used to collect real-time pressure changes, vibration attenuation characteristics, or EMG signal amplitude changes on the muscle surface. The corresponding acquisition process is triggered synchronously with the striking action to avoid interference from striking vibrations on the acquired signals and ensure the accuracy of the feedback signals.

7. The method according to claim 1, characterized in that, The extraction of feature parameters reflecting muscle stiffness includes: The pressure peak value when the pressure reaches a stable value and the deformation recovery time required for the muscle deformation to recover to 80% of the initial state after the pressure is removed are extracted from the denoised pressure-displacement signal; the vibration attenuation coefficient required for the vibration amplitude to decay to 30% of the initial value is extracted from the vibration response signal; the extraction process of the feature parameters includes signal peak detection, exponential decay curve fitting or time window integral calculation to quantify the hardness, elasticity and viscosity characteristics of muscle tissue.

8. A system for intelligently matching muscle stiffness detection with fascia gun firing parameters, characterized in that, Applied to a fascia gun; the system includes: The signal acquisition unit is used to acquire the original hardness detection signal of the target muscle area under static pressure state through a pressure sensor, vibration sensor or displacement sensor set on the fascia gun head or independent detection device; the original hardness detection signal includes pressure displacement data, vibration response data and elastic deformation data. The filtering unit is used to perform noise reduction filtering on the original hardness detection signal and extract feature parameters reflecting muscle hardness, including pressure peak value, deformation recovery time, and vibration attenuation coefficient. The parameter input unit is used to input the feature parameters into a preset intelligent matching algorithm model. The intelligent matching algorithm model is pre-trained using machine learning methods to output a combination of striking parameters adapted to the fascia gun based on muscle hardness feature parameters, user-preset striking preferences, and user's basic body data. This includes: the intelligent matching algorithm model determining a basic range of striking parameters based on the muscle hardness feature parameters, then adjusting the basic parameter range using a weight allocation algorithm in conjunction with the user's preset striking preferences and user's basic body data to generate a personalized combination of striking parameters; wherein the weight allocation algorithm dynamically optimizes the influence weight of each input factor based on the user's historical usage data; wherein the user's basic body data includes age, weight, height, muscle mass, and exercise frequency; and the striking parameter combination includes striking frequency, striking force, striking time, and amplitude depth. The parameter adjustment unit is used to control the fascia gun to perform the striking operation according to the striking parameter combination, and to collect feedback signals from the target muscle area in real time during the striking process. The feedback signals include changes in muscle vibration amplitude, changes in surface electromyography signals, or real-time data on pressure hardness. The feedback signals are input into the intelligent matching algorithm model to dynamically adjust the striking parameter combination, including: calculating the feature difference value between the feedback signal and the initial detection signal before striking; when the difference value exceeds a preset dynamic adjustment threshold, the parameter adjustment mechanism is triggered; the intelligent matching algorithm model, based on the real-time feedback trend of muscle hardness change, makes unidirectional or bidirectional fine adjustments to the striking frequency, force, or amplitude according to a preset step size on the basis of the current striking parameter combination. The adjusted parameter combination must meet the safe striking parameter range of human muscles, and the striking operation must be maintained continuously during the adjustment process so that the striking parameter combination matches the real-time hardness state of the target muscle area.

Images