Intelligent fascia physiotherapy instrument based on multi-mode bioelectric pulse
By integrating the collection of tongue, pulse, facial, and voice information, and using AI to generate collaborative physiotherapy plans, the problem of the single function of existing fascia therapy equipment and the separation of traditional Chinese medicine detection systems has been solved, achieving intelligent and personalized multimodal physiotherapy effects.
Patent Information
- Application Number
- CN202511328362.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-09-17
- Publication Date
- 2026-01-27
AI Technical Summary
Existing fascia therapy equipment has limited functionality, lacks multimodal synergistic optimization, relies on operator experience, and cannot achieve personalized and precise therapy. Traditional Chinese medicine digital detection systems separate information collection from therapy and lack the fusion analysis of multiple biological information.
The design incorporates a smart fascia therapy device based on multimodal bioelectric pulses, integrating the acquisition of tongue, pulse, facial, and voice information. Through AI-based central identification of syndromes, it generates collaborative therapy plans, achieving an integrated solution from information acquisition to therapy application.
It improves the comprehensiveness and convenience of information acquisition, reduces the impact of human operational differences, realizes the synergistic effect of multiple physiotherapy methods, improves the physiotherapy effect, and achieves standardization and objectification.
Smart Images
Figure CN121401611A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent physiotherapy technology, and more specifically, to an intelligent fascia physiotherapy device based on multimodal bioelectric pulses. Background Technology
[0002] Existing myofascial therapy equipment mainly employs a single physical therapy modality, such as intermediate-frequency electrical stimulation, ultrasound, or infrared radiation. Reference 1 (Segura-Pérez M, Hernández-Criado MT, Calvo-Lobo C, et al. A multimodal approach for myofascial pain syndrome: A prospective study[J]. Journal of manipulative and physiological therapeutics, 2017, 40(6):397-403.) reports a multimodal rehabilitation program that combines myofascial trigger point dry needling, stretching training, kinesiology tape, eccentric exercises, and patient education for myofascial pain rehabilitation. After four weeks of application, the pain intensity was significantly reduced. However, existing multimodal therapy equipment still suffers from relatively limited functionality and a lack of synergistic optimization between different therapy modalities. Furthermore, these devices typically rely on the operator's subjective experience to develop therapy plans, failing to achieve personalized and precise therapy.
[0003] The four diagnostic methods of traditional Chinese medicine (TCM) – observation, auscultation, inquiry, and palpation – have made some progress in digitalization in recent years. Reference 2 (Yang L, Dong Q, Lin D, et al. TongueNet: a multi-modal fusion and multi-label classification model for traditional Chinese Medicine tongue diagnosis[J]. Frontiers in Physiology, 2025, 16: 1527751.) proposes a tongue image analysis system based on multimodal deep learning. This system can integrate tongue image analysis and textual features, demonstrating good performance in disease nature and lesion location classification tasks. However, existing TCM digital detection systems typically separate information collection from physiotherapy applications, lacking a complete process from information gathering to physiotherapy plan generation. Furthermore, most systems focus only on a single detection method (such as only tongue image or only pulse image), failing to achieve the fusion analysis of multiple biological information. In addition, traditional physiotherapy equipment cannot automatically match corresponding physiotherapy parameters according to TCM theory, resulting in TCM physiotherapy still heavily relying on the operator's clinical experience, making standardization and objectification difficult.
[0004] Therefore, existing technologies lack devices that can effectively integrate multiple bio-information acquisition, intelligent analysis and processing, and multimodal collaborative physiotherapy functions to achieve an integrated solution from information acquisition to physiotherapy application, thereby improving the intelligence level and application effect of physiotherapy equipment. Summary of the Invention
[0005] To overcome the aforementioned shortcomings of existing technologies, this invention provides an intelligent fascia therapy device based on multimodal bioelectric pulses. This device integrates multi-source information from tongue, pulse, facial, and speech data, with an AI central processing unit identifying symptoms and generating a coordinated therapy plan. This integrated design establishes a closed loop from objective data to therapy instructions, addressing the technical deficiencies of existing technologies, such as the separation of diagnosis and therapy, lack of multi-information fusion, and reliance on subjective experience in treatment plans.
[0006] To achieve the above objectives, the present invention provides the following technical solution:
[0007] This intelligent fascia therapy device based on multimodal bioelectric pulses includes a four-diagnostic data acquisition module, an artificial intelligence control center, and a multimodal therapy module. The four-diagnostic data acquisition module includes a tongue image acquisition unit, a pulse image acquisition unit, and a facial and voice information acquisition unit. The tongue image acquisition unit includes an image sensor and a standard light source for acquiring standardized user tongue images. The pulse image acquisition unit includes a sensor array and a pressure application device for acquiring pulse waveform data at different pressure depths at the cun, guan, and chi positions. The facial and voice information acquisition unit includes a camera and a microphone for acquiring user facial images and voice information.
[0008] The artificial intelligence control center is connected to the four diagnostic methods data acquisition module to receive pulse waveform data, tongue image, facial image and voice information. It converts the received multi-source information into standardized TCM feature vectors, identifies the TCM feature vectors as a specific TCM syndrome through a pre-trained TCM syndrome classification model, and generates a set of multi-parameter physiotherapy control instructions that uniquely correspond to the TCM syndrome according to a preset syndrome-physiotherapy plan mapping rule library.
[0009] The multimodal physiotherapy module is connected to the artificial intelligence control center and is used to receive the multi-parameter physiotherapy control commands and synchronously drive the pulse electromagnetic field unit, the intermediate frequency bioelectric unit, the far-infrared radiation unit and the physiotherapy ultrasound unit inside it to work together according to the commands.
[0010] As a further aspect of the present invention, the process by which the artificial intelligence control center identifies the TCM syndrome through the TCM syndrome classification model and generates the multi-parameter physiotherapy control command specifically involves: inputting the standardized TCM feature vector into the pre-trained TCM syndrome classification model to obtain the probability score of the TCM syndrome in the preset syndrome database, and selecting the syndrome with the highest probability score as the identified syndrome; then, retrieving and extracting the multi-parameter physiotherapy control command that uniquely corresponds to the identified syndrome from the syndrome-physiotherapy scheme mapping rule database, which includes parameters such as pulse electromagnetic field frequency and intensity, mid-frequency bioelectric beat frequency and intensity, far-infrared radiation temperature, and physiotherapy ultrasound frequency, intensity, and duty cycle.
[0011] As a further embodiment of the present invention, the sensor array in the pulse acquisition unit consists of three piezoelectric thin-film sensors corresponding to the cun, guan, and chi positions respectively. The pressure application device is a linear voice coil motor actuator positioned directly above the sensor array and used to apply normal pressure. The artificial intelligence control center is configured to control the linear voice coil motor actuator to apply a sequential static pressure from 5 kPa to 30 kPa to the sensor array within a preset time, so that the sensor array continuously records pulse waveform data at different static pressure levels. The pulse waveform data is then processed to extract the time-domain and frequency-domain features from the traditional Chinese medicine feature vector. The time-domain features include the peak value and width of the main wave, and the frequency-domain features include the energy and ratio of the first five harmonic components obtained through fast Fourier transform.
[0012] As a further embodiment of the present invention, the image sensor in the tongue image acquisition unit is a CMOS image sensor, and the standard light source is a D65 standard light source; the artificial intelligence control center is configured to perform color correction on the acquired tongue image, and then automatically segment the tongue image into four regions of interest—the tip area, the middle area, the root area, and the side area—through a predefined geometric mask, and independently calculate the CIELab color space feature value of each region of interest.
[0013] As a further embodiment of the present invention, the camera in the facial and voice information acquisition unit is a high-definition camera, and the microphone is a MEMS microphone array; the artificial intelligence control center is configured to process the facial image acquired by the high-definition camera under a D65 standard light source to extract the color feature value of the cheek area, drive the text-to-speech engine to broadcast standardized prompts to guide the user to generate speech, and perform acoustic analysis on the user speech recorded by the MEMS microphone array to extract its average fundamental frequency and amplitude mean features.
[0014] As a further embodiment of the present invention, the intermediate frequency bioelectric unit includes two independent output channels, wherein the first channel outputs a sinusoidal carrier current with a fixed frequency, and the second channel outputs a carrier current with a variable frequency. The frequency difference between the two carrier currents constitutes a low-frequency beat frequency. The artificial intelligence control center sets the frequency value of the low-frequency beat frequency by adjusting the output frequency of the second channel according to the multi-parameter physiotherapy control command.
[0015] As a further embodiment of the present invention, the far-infrared radiation unit uses a flexible electrothermal radiation film as a radiation source and has an NTC thermistor built in. Through a PID controller, the temperature of the physiotherapy surface is maintained at the target temperature value set by the multi-parameter physiotherapy control command.
[0016] As a further aspect of the present invention, the pulsed electromagnetic field unit is used to generate a pulsed electromagnetic field according to the frequency and intensity set by the multi-parameter physiotherapy control command.
[0017] As a further aspect of the present invention, the therapeutic ultrasound unit is used to generate ultrasound waves according to the frequency, intensity, and duty cycle set by the multi-parameter therapeutic control command.
[0018] As a further aspect of the present invention, a skin impedance sensor is integrated on the electrodes of the intermediate frequency bioelectric unit; the artificial intelligence control center is also configured to continuously monitor the skin impedance value obtained by the skin impedance sensor during the physiotherapy process, compare the skin impedance value with a preset safety threshold, and automatically reduce the output current intensity of the intermediate frequency bioelectric unit when the skin impedance value exceeds the safety threshold.
[0019] Compared with existing technologies, the beneficial effects of the intelligent myofascial therapy device based on multimodal bioelectric pulses of this invention are as follows:
[0020] This invention overcomes the shortcomings of existing technologies, which involve separate functions of related analysis devices and require multiple switching for information acquisition, by integrating a tongue image acquisition unit, a pulse image acquisition unit, and a facial and voice information acquisition unit into a single device. Specifically, this integrated design effectively improves the comprehensiveness and objectivity of data acquisition related to the four diagnostic methods of traditional Chinese medicine: the tongue image acquisition unit achieves objectification and standardization of tongue image information; the pulse image acquisition unit achieves objectification and standardization of multi-dimensional pulse image information; and the facial and voice information acquisition unit standardizes the traditional observation and auscultation processes by objectively quantifying facial color and standardizing the acquisition of voice features. This integrated design not only greatly improves the comprehensiveness and convenience of information acquisition but also minimizes the impact of human operational differences on the accuracy of subsequent AI analysis through a standardized acquisition process.
[0021] This invention's multimodal physiotherapy module integrates four different physical therapy methods: pulsed electromagnetic field, mid-frequency bioelectricity, far-infrared radiation, and therapeutic ultrasound. Through an AI control center, it automatically generates multi-parameter control commands based on identified traditional Chinese medicine syndromes to achieve the synergistic output of these four therapy methods. Compared to existing physiotherapy devices that are single-function, offering only a single therapy mode, or requiring manual parameter adjustment by the operator, this invention produces better therapeutic effects through the synergistic effect of multiple therapy methods. Furthermore, automated parameter matching avoids the problem of poor therapeutic results due to insufficient operator experience. Attached Figure Description
[0022] Figure 1 This is a sample example of tongue feature extraction and syndrome classification from Reference 2.
[0023] Figure 2 This is a schematic diagram of the structure of the intelligent fascia therapy device based on multimodal bioelectric pulses of the present invention.
[0024] Figure 3 This is a time-frequency domain analysis diagram of the pulse signal of the intelligent myofascial therapy device based on multimodal bioelectric pulses according to the present invention.
[0025] Figure 4 This is a performance verification diagram of the random forest classifier for the intelligent fascia therapy device based on multimodal bioelectric pulses of this invention.
[0026] Figure 5 This is a thermogram of the therapeutic parameters of the intelligent fascia therapy device based on multimodal bioelectric pulses according to the present invention.
[0027] Figure 6 This is a comparison chart showing the therapeutic effects of the intelligent fascia therapy device based on multimodal bioelectric pulses of the present invention. Detailed Implementation
[0028] The technical solutions of this embodiment will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.
[0029] Example 1
[0030] A smart fascia therapy device based on multimodal bioelectric pulses. It includes a four-diagnostic data acquisition module, an AI control center, and a multimodal therapy module.
[0031] During operation, the four diagnostic methods data acquisition module collects multi-source physiological data from the user and sends digital signals to the AI control center. The AI control center processes and analyzes the received data, identifies the user's TCM syndrome, and generates a set of multi-parameter physiotherapy control instructions based on preset mapping logic. Finally, these instructions are sent to the multimodal physiotherapy module, which drives multiple physical therapy units to perform coordinated physiotherapy on the target areas of the user's body.
[0032] The four diagnostic methods data acquisition module is used to transform the subjective perception process of traditional Chinese medicine's four diagnostic methods into objective, repeatable, and measurable engineering data. In this embodiment of the invention, the four diagnostic methods data acquisition module includes a tongue image acquisition unit, a pulse image acquisition unit, and a facial and voice information acquisition unit. This module achieves a comprehensive assessment of the user's physiological state through the objective acquisition of relevant data from the four diagnostic methods (inspection, auscultation, inquiry, and palpation).
[0033] The tongue image acquisition unit includes a D65 standard light source with a color temperature of 6504K and a CMOS image sensor. The standard light source ensures that tongue image images acquired under different environments have a consistent color reference, avoiding interference from ambient light.
[0034] The workflow is as follows: First, a digital image of the user's tongue is acquired using a CMOS image sensor; then, the AI control center performs color correction on the acquired digital image using a pre-calibrated ICC color profile to accurately reproduce the tongue's color; the corrected image is converted to the CIELAB color space for subsequent quantitative analysis; finally, based on a preset geometric mask, the AI control center automatically segments the tongue image into four regions of interest: the tip, middle, root, and sides of the tongue, and independently calculates and records the L* (luminance), a* (red-green saturation), and b* (brightness) values for each region.
[0035] Value (yellow-blue). The geometric mask is a digital image template whose internal region division strictly follows the correspondence between the tongue and internal organs in Traditional Chinese Medicine theory, namely, "the tip of the tongue corresponds to the heart and lungs, the middle of the tongue to the spleen and stomach, the root of the tongue to the kidneys, and the sides of the tongue to the liver and gallbladder." After color correction and spatial transformation of the acquired tongue image, this mask template is called and aligned with the tongue image using an image registration algorithm and then superimposed. After superimposition, the complete tongue image can be automatically and repeatedly segmented into four regions of interest, and the visual features such as color and texture of each region can be independently quantified and extracted.
[0036] Traditional Chinese medicine pulse diagnosis is a multi-dimensional, qualitative assessment method that relies heavily on the physician's experience. The examination site is the "cun kou" (radial artery at the wrist). Traditional pulse diagnosis revolves around four core principles: "location, rate, shape, and momentum." "Location" refers to the depth at which the pulse is most clearly perceived, categorized as superficial or deep; "rate" refers to the frequency and rhythm of the pulse, such as a slow pulse (<60 bpm) or a rapid pulse (>90 bpm).
[0037] "Form" describes the specific shape of the pulse fluctuation, such as a wiry pulse (straight and long, like pressing a harp string) or a slippery pulse (flowing smoothly, like beads rolling on a plate); "strength" describes the strength of the pulse, divided into full and weak. Transforming these rich, subjective, yet clinically significant descriptions into objective, quantifiable engineering signals is the core challenge facing the objectification and modernization of Traditional Chinese Medicine. The pulse acquisition unit in this invention is designed to solve this technical problem.
[0038] The pulse acquisition unit comprises a sensor array consisting of three independent polyvinylidene fluoride (PVDF) piezoelectric film sensors, and a linear voice coil motor actuator. The three PVDF sensors are spaced 15 mm apart, and their physical layout corresponds to the cun, guan, and chi positions in traditional Chinese medicine pulse diagnosis.
[0039] Traditional Chinese medicine pulse diagnosis is not a static perception, but rather an observation of the layers and changes in strength of the pulse through three different depths of pressure: "lifting," "pressing," and "searching." To engineer this process, the linear voice coil motor actuator of the pulse acquisition unit applies a dynamically changing pressure to the piezoelectric thin-film sensor array. In specific implementation, the AI control center controls the linear voice coil motor actuator to linearly increase the applied pressure from 5 kPa to 30 kPa within 10 seconds. Traditional Chinese medicine pulse diagnosis assesses different levels of pathological information through three different pressure levels: "superficial," "medium," and "deep." A pressure of 5 kPa is similar to the "superficial" pressure of a practitioner lightly touching the patient's skin with their fingertips; the pulse perceived at this pressure mainly reflects superficial disease information. A pressure of 30 kPa, on the other hand, is similar to the "deep" pressure, pressing firmly to the bones and muscles; the pulse perceived at this pressure can reveal deeper organ dysfunctions. By systematically scanning the entire pressure range from 5 to 30 kPa, the system can comprehensively capture and quantify pulse waveform information at all depths required for traditional Chinese medicine pulse diagnosis, thereby enabling precise syndrome identification for the subsequent AI control center. During this pressure change process, the sensor array continuously records and outputs the corresponding pulse waveform data at each pressure gradient, thus obtaining richer pulse information than under a single pressure.
[0040] The facial and voice information acquisition unit in this embodiment of the invention is used to acquire objective physiological information about a user's face and voice. Structurally, this unit integrates a high-definition camera and a MEMS microphone array. The high-definition camera acquires static images of the user's face under the same D65 standard light source as the tongue image acquisition unit, allowing the AI control center to analyze the average chromaticity values of key areas such as the cheeks and forehead, thereby quantifying facial color. The MEMS microphone array uses a text-to-speech (TTS) engine to broadcast standardized questions to guide the user to generate voice responses, which are then recorded. The AI control center subsequently performs acoustic analysis on the voice signal, extracting its average fundamental frequency and average amplitude characteristics to quantify the user's voice features.
[0041] The internal processing flow of the AI control center in this embodiment of the invention is mainly divided into three stages: feature extraction, syndrome classification, and scheme generation.
[0042] First, the feature extraction stage transforms the raw data collected by each sensor into a standardized, quantified TCM feature vector for subsequent model input. For example, for the pulse waveform data collected by the pulse acquisition unit, the AI control center performs time-domain and frequency-domain analysis. Time-domain analysis mainly extracts features that characterize the waveform morphology, including the peak value of the main wave, the peak value of the dicrotic wavefront, and the width of the main wave. Frequency-domain analysis obtains the harmonic composition of the signal through Fast Fourier Transform (FFT), mainly extracting the energy and ratio of the first five harmonic components. Figure 3 This demonstrates the feature extraction process. For example... Figure 3 (a) shows the original pulse signals collected from the cun, guan, and chi positions. Figure 3 (d) shows how to calibrate and extract time-domain features such as h1 (peak value of the main wave), h4 (trough value of the diabetic wave), and W1 (width of the main wave) within a typical pulse cycle. Figure 3 (e) shows the power spectral density map obtained after performing an FFT on the signal, and Figure 3 (f) further demonstrates the quantitative analysis results of the energy of the five key harmonic components C1 to C5. For the static facial images and voice data of the user acquired by the facial and voice information acquisition units, the AI control center extracts facial and acoustic features. Facial processing focuses on extracting the average chromaticity values of the user's cheeks and forehead as facial color features. Acoustic processing analyzes the user's voice, extracting its average fundamental frequency and average amplitude as speech acoustic features.
[0043] The core of the syndrome classification stage in this embodiment of the invention is a pre-trained machine learning classifier. In this embodiment of the invention, a random forest multi-label classifier is used.
[0044] Random forest is a well-known ensemble learning algorithm in machine learning, consisting of multiple decision trees, and is widely used due to its high accuracy and strong robustness against overfitting. This invention uses this classic algorithm as the basis for its classification model. The syndrome differentiation process in traditional Chinese medicine is essentially a complex cognitive process highly dependent on the physician's personal knowledge, experience, and senses. Different physicians observing the same patient may arrive at different judgments due to differences in experience. This subjectivity and inconsistency in differentiation is a key bottleneck in the modernization and standardization of traditional Chinese medicine. The development of artificial intelligence and machine learning technologies provides powerful tools to solve this problem. By training a classification model on a large-scale clinical dataset labeled with consensus by senior experts, the implicit, unstructured differentiation logic and knowledge of experts can be transformed into an objective, stable, and repeatable mathematical model. This model can automatically learn and extract feature patterns strongly correlated with specific TCM syndromes from high-dimensional quantitative physiological data, thereby achieving automated and standardized syndrome differentiation. The training process of this model is based on a database containing 1,850 clinical samples. This database associates the quantified TCM feature vectors obtained through the aforementioned feature extraction method with authoritative syndrome labels marked by senior TCM doctors, thereby training a classifier that can automatically identify TCM syndromes.
[0045] In this embodiment of the invention, the scheme generation stage is completed through a pre-defined syndrome-physiotherapy scheme mapping rule base. This rule base stores the mapping relationship between TCM syndromes and multi-parameter physiotherapy control commands. Its construction includes: First, associating specific TCM syndrome pathologies with modern medical physiological states. For example, the TCM syndrome of "qi stagnation and blood stasis" is highly correlated with microcirculatory disorders and poor blood flow in modern physiology. Second, based on physiotherapy modalities and parameters proven to improve corresponding physiological states, preliminary scheme matching is performed. For instance, pulsed electromagnetic fields and mid-frequency interference currents have been proven to promote blood circulation and can therefore be used to achieve the TCM physiotherapy goal of promoting qi and blood circulation. Finally, through a clinical sample database with authoritative syndrome annotations by senior TCM physicians, these preliminarily matched physiotherapy parameters are trained, optimized, and calibrated, ultimately forming a deterministic rule base capable of mapping identified syndromes to a set of synergistic physiotherapy parameters. Once a specific TCM syndrome is identified in the previous stage, the system will query the syndrome-treatment plan mapping rule library and match a uniquely corresponding multi-parameter physiotherapy control command. This command is a set of data used to simultaneously control the collaborative output of at least two different physical units in the multimodal physiotherapy module.
[0046] For example, when the syndrome identified in the previous stage is 'qi stagnation and blood stasis', the system will query the syndrome-physiotherapy plan mapping rule library and match the corresponding multi-parameter physiotherapy control instructions. The specific parameters are shown in Table 1.
[0047] Table 1 - Multi-parameter physiotherapy control instructions
[0048]
[0049] The multimodal physiotherapy module in this embodiment of the invention is the physical execution unit of the physiotherapy plan, and all its parameters are precisely controlled by the AI control center. In this embodiment of the invention, the multimodal physiotherapy module includes a pulsed electromagnetic field unit, a mid-frequency bioelectric unit, a far-infrared radiation unit, and a therapeutic ultrasound unit.
[0050] In this embodiment of the invention, the pulsed electromagnetic field unit is an array of six planar helical coils. These coils can be directly integrated onto a flexible or rigid printed circuit board. Through precise control and layout optimization of the current in each coil of the array, a sufficiently uniform pulsed electromagnetic field with a frequency between 0.5Hz and 50.0Hz and a magnetic induction intensity between 10μT and 800μT can be generated in the target therapeutic area. Pulsed electromagnetic field (PEMF) is a non-invasive therapeutic method that uses a pulsed magnetic field of a specific frequency to interact with the body's own electromagnetic field, thereby influencing physiological processes such as cell membrane potential and ion transmembrane transport.
[0051] The mid-frequency bioelectric unit employs the principle of interferometric electrotherapy (IFC) and comprises two independent output channels. The first channel outputs a fixed 4000Hz sinusoidal carrier current, while the second channel outputs a carrier current with a frequency varying between 4001Hz and 4250Hz. The frequency difference between the two currents constitutes a beat frequency varying within the range of 1Hz to 250Hz. The AI control center precisely sets the beat frequency based on the identified different symptoms to achieve different therapeutic objectives. For example, for symptoms requiring analgesia, a high beat frequency of 80Hz to 100Hz is set; while for symptoms requiring improved local circulation, a low beat frequency of 1Hz to 10Hz is set to effectively stimulate the muscle pump.
[0052] The far-infrared radiation unit uses a flexible electrothermal radiation film as its radiation source, emitting electromagnetic waves with energy concentrated in the 8-14 micrometer wavelength band. To achieve precise and safe temperature control, the unit integrates an NTC thermistor for real-time surface temperature monitoring and uses a PID (proportional-integral-derivative) controller for closed-loop feedback control. This allows the therapeutic surface temperature to be precisely maintained between 36.0℃ and 55.0℃, with a control accuracy of ±0.5℃, thus ensuring the therapeutic effect while eliminating the risk of low-temperature burns. Far-infrared rays are invisible electromagnetic waves found in sunlight. Their 8-14 micrometer wavelength band is similar to the far-infrared wavelength emitted by the human body, allowing them to be effectively absorbed by water molecules within the body. This absorption converts energy into heat through a resonance effect, a process known as resonant absorption.
[0053] The therapeutic ultrasound unit employs a piezoelectric transducer, whose operating frequency can switch between 1.0 MHz for deep tissues and 3.0 MHz for superficial tissues. Its output sound intensity can be adjusted within the range of 0.1 W to 2.0 W per square centimeter. The AI control center can also control its duty cycle to emphasize different therapeutic effects: for example, setting it to a 100% continuous mode to primarily generate a deep thermal effect, or setting it to a 10%-20% pulse mode to primarily generate non-thermal effects, such as micro-mechanical massage, to loosen tissue adhesions.
[0054] In addition, this physiotherapy device includes a skin impedance sensor integrated with the mid-frequency bioelectric unit to provide proactive safety. During treatment, the AI control center makes real-time judgments based on the skin impedance value monitored by this sensor. In this embodiment, the safety threshold is set as follows: if the real-time skin impedance value increases by more than 30% compared to the initial reference value within 5 seconds, the system determines that the contact is poor or the skin is excessively dry. At this time, the AI control center automatically reduces the output current intensity of the mid-frequency bioelectric unit by 20% and issues a prompt to the user to ensure safe use.
[0055] This invention, through the aforementioned structure and method, firstly utilizes standardized sensors to objectively collect physiological information such as the user's tongue, pulse, facial features, and voice; then, an AI control center processes the data to automatically identify TCM syndromes; finally, based on the identification results, a precise multi-parameter physiotherapy plan is matched from a preset database, and multiple physical therapy units are driven to perform synergistic therapy. The entire process achieves a complete closed loop from objective data collection to intelligent, personalized synergistic therapy, overcoming the shortcomings of existing technologies where information collection and therapy are separated, and the plans cannot be customized.
[0056] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
[0057] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.
Claims
1. An intelligent fascia therapy device based on multimodal bioelectric pulses, comprising a four-diagnosis data acquisition module, an artificial intelligence control center, and a multimodal therapy module; characterized in that, The four diagnostic methods data acquisition module includes a tongue image acquisition unit, a pulse image acquisition unit, and a facial and voice information acquisition unit. The tongue image acquisition unit includes an image sensor and a standard light source for acquiring standardized user tongue images. The pulse image acquisition unit includes a sensor array and a pressure application device for acquiring pulse waveform data at different pressure depths at the cun, guan, and chi positions. The facial and voice information acquisition unit includes a camera and a microphone for acquiring user facial images and voice information. The artificial intelligence control center is connected to the four diagnostic methods data acquisition module to receive pulse waveform data, tongue image, facial image and voice information. It converts the received multi-source information into standardized TCM feature vectors, identifies the TCM feature vectors as a specific TCM syndrome through a pre-trained TCM syndrome classification model, and generates a set of multi-parameter physiotherapy control instructions that uniquely correspond to the TCM syndrome according to a preset syndrome-physiotherapy plan mapping rule library. The multimodal physiotherapy module is connected to the artificial intelligence control center and is used to receive the multi-parameter physiotherapy control commands and synchronously drive the pulse electromagnetic field unit, the intermediate frequency bioelectric unit, the far-infrared radiation unit and the physiotherapy ultrasound unit inside it to work together according to the commands.
2. The intelligent myofascial therapy device based on multimodal bioelectric pulses according to claim 1, characterized in that, The process by which the artificial intelligence control center identifies the TCM syndrome through the TCM syndrome classification model and generates the multi-parameter physiotherapy control command specifically involves: inputting the standardized TCM feature vector into the pre-trained TCM syndrome classification model to obtain the probability score of the TCM syndrome in the preset syndrome database, and selecting the syndrome with the highest probability score as the identified syndrome; then, retrieving and extracting the multi-parameter physiotherapy control command that uniquely corresponds to the identified syndrome from the syndrome-physiotherapy scheme mapping rule database, which includes parameters such as pulse electromagnetic field frequency and intensity, mid-frequency bioelectric beat frequency and intensity, far-infrared radiation temperature, and physiotherapy ultrasound frequency, intensity, and duty cycle.
3. The intelligent myofascial therapy device based on multimodal bioelectric pulses according to claim 1, characterized in that, The sensor array in the pulse acquisition unit consists of three piezoelectric thin-film sensors corresponding to the cun, guan, and chi positions, respectively. The pressure application device is a linear voice coil motor actuator positioned directly above the sensor array to apply normal pressure. The artificial intelligence control center is configured to control the linear voice coil motor actuator to apply a sequential static pressure ranging from 5 kPa to 30 kPa to the sensor array within a preset time period. This allows the sensor array to continuously record pulse waveform data at different static pressure levels. The pulse waveform data is then processed to extract the time-domain and frequency-domain features from the traditional Chinese medicine feature vector. The time-domain features include the peak value and width of the main wave, while the frequency-domain features include the energy and ratio of the first five harmonic components obtained through a fast Fourier transform.
4. The intelligent myofascial therapy device based on multimodal bioelectric pulses according to claim 1, characterized in that, The image sensor in the tongue image acquisition unit is a CMOS image sensor, and the standard light source is a D65 standard light source. The artificial intelligence control center is configured to perform color correction on the acquired tongue image, and then automatically segment the tongue image into four regions of interest (ROIs): the tip of the tongue, the middle of the tongue, the root of the tongue, and the side of the tongue, using a predefined geometric mask, and independently calculate the CIELab color space feature value of each ROI.
5. The intelligent myofascial therapy device based on multimodal bioelectric pulses according to claim 1, characterized in that, The camera in the facial and voice information acquisition unit is a high-definition camera, and the microphone is a MEMS microphone array. The artificial intelligence control center is configured to process the facial image acquired by the high-definition camera under D65 standard light source to extract the color feature value of the cheek area, drive the text-to-speech engine to broadcast standardized prompts to guide the user to generate speech, and perform acoustic analysis on the user speech recorded by the MEMS microphone array to extract its average fundamental frequency and amplitude mean features.
6. The intelligent myofascial therapy device based on multimodal bioelectric pulses according to claim 1, characterized in that, The intermediate frequency bioelectric unit includes two independent output channels, wherein the first channel outputs a sinusoidal carrier current with a fixed frequency, and the second channel outputs a carrier current with a variable frequency. The frequency difference between the two carrier currents constitutes a low-frequency beat frequency. The artificial intelligence control center sets the frequency value of the low-frequency beat frequency by adjusting the output frequency of the second channel according to the multi-parameter physiotherapy control command.
7. The intelligent myofascial therapy device based on multimodal bioelectric pulses according to claim 1, characterized in that, The far-infrared radiation unit uses a flexible electrothermal radiation film as the radiation source and has an NTC thermistor built in. Through a PID controller, the temperature of the physiotherapy surface is maintained at the target temperature value set by the multi-parameter physiotherapy control command.
8. The intelligent myofascial therapy device based on multimodal bioelectric pulses according to claim 1, characterized in that, The pulsed electromagnetic field unit is used to generate a pulsed electromagnetic field according to the frequency and intensity set by the multi-parameter physiotherapy control command.
9. The intelligent myofascial therapy device based on multimodal bioelectric pulses according to claim 1, characterized in that, The therapeutic ultrasound unit is used to generate ultrasound waves according to the frequency, intensity, and duty cycle set by the multi-parameter therapeutic control command.
10. The intelligent myofascial therapy device based on multimodal bioelectric pulses according to claim 1, characterized in that, The electrodes of the intermediate frequency bioelectric unit are integrated with skin impedance sensors; the artificial intelligence control center is also configured to continuously monitor the skin impedance value obtained by the skin impedance sensor during the physiotherapy process, compare the skin impedance value with a preset safety threshold, and automatically reduce the output current intensity of the intermediate frequency bioelectric unit when the skin impedance value exceeds the safety threshold.