Physiotherapy system and method for AI-assisted acupoint positioning

Abstract

The invention discloses an AI-assisted acupoint positioning physiotherapy system and method, and relates to the technical field of traditional Chinese medicine intelligent physiotherapy, and the system comprises a holographic interaction module which reconstructs a limb model through AR glasses and an SLAM technology, and generates a holographic dialectical interface; the intelligent deformation ring sleeve is attached to the limbs through an SMP material, and electronic skin is arranged in the intelligent deformation ring sleeve to monitor acupuncture points; the quantum correlation dialectical module is used for constructing a knowledge graph, calculating quantum state similarity and correcting matched acupoints; the multi-field cooperative physiotherapy module integrates various energy fields to form a six-mode matrix; the biosafety barrier module monitors impedance and HRV and has a child mode. According to the system, multiple technologies are fused to improve the accuracy and safety of intelligent physiotherapy, holographic interaction assists in acupoint positioning, the intelligent deformation loop collects physiological signals, quantum correlation differentiation accelerates the process, multi-field cooperative physiotherapy synergism is achieved, a biological safety barrier guarantees safety, space-time rhythm adjustment improves the curative effect, and the physiotherapy effective rate and the patient satisfaction degree are greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent traditional Chinese medicine physiotherapy technology, and in particular to a physiotherapy system and method for AI-assisted acupoint positioning. Background Art

[0002] With the development of intelligent TCM physiotherapy, traditional acupoint positioning and treatment methods face numerous technical bottlenecks. Existing systems often rely on manual palpation or two-dimensional imaging to locate acupoints, which is greatly influenced by the physician's experience. The positioning accuracy is generally less than 1mm, making it difficult to adapt to precise operation under complex anatomical structures. Although AR technology has been applied to some physiotherapy scenarios, the 3D reconstruction delay of existing equipment often exceeds 30ms, and the SLAM algorithm often has an error greater than 1mm in dynamic limb tracking, which cannot meet the needs of real-time interaction. Furthermore, it lacks the ability to fusion multimodal data, making it difficult to synchronously integrate dialectical information such as tongue image and odor.

[0003] Traditional systems for syndrome differentiation and treatment generally rely on rule-based acupuncture point selection. This means complex syndrome differentiation takes over five minutes, and the learning cycle for new symptoms can stretch for months, making them unsuitable for clinical needs. Drug delivery relies on topical or oral administration, resulting in poor skin penetration and a time to effect exceeding 40 minutes. Energy field therapy often utilizes a single electrical pulse mode, with a penetration depth of less than 1 cm, making it difficult to reach deep meridians. Furthermore, it lacks adaptive mechanisms to address individual body constitutions, leading to significant individual variability in efficacy.

[0004] In terms of safety, existing physiotherapy devices often lack real-time physiological monitoring mechanisms. The response time to impedance anomalies exceeds one second, making it impossible to promptly block hazardous energy output. Child-mode protection is simple, posing operational safety risks. Furthermore, systems for regulating spatiotemporal rhythms and evaluating efficacy are lacking, making it impossible to optimize treatment periods based on the theory of meridian flow. Long-term efficacy tracking relies on manual record-keeping. These technical shortcomings make it difficult for traditional physiotherapy systems to meet the demands of modern healthcare in terms of accuracy, safety, and intelligence. Summary of the Invention

[0005] The present invention proposes an AI-assisted acupoint positioning physical therapy system and method to solve the problems mentioned in the above-mentioned prior art.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: An AI-assisted acupoint positioning physical therapy system includes the following modules: Holographic interaction module: Equipped with AR glasses, it uses SLAM technology to reconstruct a 3D model of the user's limbs in real time, superimposing dynamic meridian projections. Users can slide gestures or select symptoms through eye tracking, and the system will simultaneously generate a holographic diagnosis interface that includes anatomical structures. Smart deformable ring: Made of shape memory polymer with a glass transition temperature of 37±2°C, it automatically conforms to the curve of the hand or wrist when exposed to body temperature. It has a built-in 48-channel flexible electronic skin, which can draw a real-time pressure-temperature matrix diagram of the acupoint area and dynamically adjust the ring tension. Quantum correlation syndrome differentiation module: Constructs a knowledge graph of TCM meridians and uses a quantum neural network algorithm to calculate the quantum state similarity between symptoms and acupoints. The formula is: ,in, is the inner product of the quantum state, For symptomatic state, For acupoint status, pulsed infrared thermal imaging is used to detect acupoint temperature differences and automatically correct acupoint matching plans; Multi-field synergistic therapy module: Integrates microsecond pulsed electric field, terahertz wave thermal transmission and bionic sound wave technology, combined with traditional electric pulse, moxibustion and vibration modes to form a six-modal energy matrix. Each treatment contact independently outputs at least three energy fields, achieving a combined therapeutic effect of "acupuncture-moxibustion-medication-massage-sound-light"; Biosafety Barrier Module: Deploys an impedance-heart rate dual monitoring mechanism. When an impedance abnormality is detected, energy output is cut off within 0.1 seconds. Heart rate variability monitoring is combined with the Kullback-Leibler divergence algorithm to evaluate physiological stress responses in real time. When the calculated Dkl value is less than 0.3, it is determined to be a safe state. Facial recognition is enabled in child mode, and the device is automatically locked when operated by non-guardians.

[0007] Furthermore, it also includes a spatiotemporal rhythm regulation module: a built-in atomic clock automatically calculates the best treatment period according to the meridian flow theory in the Yellow Emperor's Classic of Internal Medicine, and plays the Wave Music regulates melatonin secretion through optogenetic technology and enhances the duration of therapeutic effect.

[0008] Furthermore, it also includes a nano-targeted drug delivery module: the ring-shaped contact has a built-in microfluidic drug reservoir, which uses electroporation technology to open the skin channel and deliver Chinese medicine nanoparticles in a targeted manner. The drug reservoir supports the replacement of 20 types of drug solutions, and the drug permeability is monitored in real time through near-infrared spectroscopy.

[0009] Furthermore, in the holographic interaction module, the AR meridian projection has the function of augmented reality acupuncture simulation. In terms of puncture path planning, advanced algorithms are used for intelligent calculation. At the same time, in conjunction with force feedback gloves, the soreness, swelling and numbness produced when qi is obtained are simulated, so that the similarity between the simulated feeling and the real experience reaches more than 90%.

[0010] Furthermore, in the quantum correlation diagnosis module, the quantum neural network contains 5 quantum bit layers, uses a variational quantum algorithm to optimize parameters, and the update cycle of the disease-acupoint association rules is ≤24 hours, and new rules are automatically synchronized to all devices.

[0011] Furthermore, in the multi-field collaborative therapy module, terahertz wave thermal penetration is combined with adaptive phase modulation technology to adjust the wavefront phase according to different body constitutions, so that energy is selectively deposited in the deep layers of the meridians.

[0012] Furthermore, in the biosafety barrier module, HRV analysis uses the approximate entropy algorithm ApEn. When ApEn>1.5, it is determined to be a stress state, and the white noise soothing program is automatically triggered to reduce ApEn to a safe range within 5 minutes, that is, ApEn<1.0.

[0013] Furthermore, the method of the AI-assisted acupoint positioning physical therapy system includes the following steps: Five-sense fusion syndrome differentiation steps: The user wears AR glasses and looks at the holographic meridian model, while simultaneously inhaling nine herbal scents and taking a tongue image. The system uses a multimodal data fusion algorithm with a visual weighting coefficient of 0.4, an olfactory weighting coefficient of 0.3, and a tongue image weighting coefficient of 0.3 to ultimately generate a personalized syndrome differentiation vector V=[v1,v2,v3]; Dynamic acupoint calibration steps: The ring-shaped electronic skin detects the temperature change rate of the acupoint area, triggers AI dynamic bone point tracking, and combines the quantum state similarity algorithm to adjust the contact position in real time to ensure that energy is accurately applied to the active meridian target; Six-dimensional energy tuning steps: Based on the syndrome differentiation vector V, the six-modal energy parameters are optimized through genetic algorithms to generate a personalized treatment sequence. During the treatment, skin impedance and HRV physiological indicators are collected every 2 minutes to dynamically adjust the energy output.

[0014] Furthermore, it also includes cross-temporal and spatial efficacy evaluation steps: establishing a blockchain medical data chain, uploading patient treatment data to the cloud after encryption through federated learning, predicting long-term efficacy through survival analysis algorithm, and automatically pushing 3-month health management recommendations after the treatment is completed.

[0015] Furthermore, it also includes the meridian energy guidance step: at the end of the treatment, the gradient energy attenuation program is started, and at the same time, the patient's thoughts are guided through the meridians through voice guidance, combined with breathing frequency adjustment, so that the energy can dissipate naturally along the meridians, reducing discomfort after treatment.

[0016] Compared with the existing technology, the beneficial effects of the present invention are: This application significantly improves the accuracy and safety of intelligent physical therapy through the integration of multiple technologies. The holographic interaction module uses SLAM and high-precision AR projection to control acupoint positioning errors to a very small range, shortening the three-dimensional reconstruction delay to an extremely short time. The force feedback gloves simulate the feeling of air with a very high degree of similarity, greatly improving the positioning efficiency of novice doctors. The SMP material of the intelligent deformable ring and the flexible electronic skin achieve high pressure sensing accuracy and fast dynamic fitting response time, providing stable support for the acquisition of physiological signals in the acupoint area.

[0017] The quantum correlation syndrome differentiation module significantly compresses the traditionally lengthy syndrome differentiation process, significantly improving the accuracy of acupoint selection. Newly added symptoms can be quickly updated with knowledge. Combined with pulsed infrared thermal imaging, the efficiency of acupoint selection plan revisions is significantly improved. Multi-field synergistic therapy utilizes a six-modal energy matrix to significantly increase drug skin penetration, significantly shortening the onset of action. Terahertz wave penetration reaches a certain depth, and phase modulation tailored to individual body constitutions minimizes energy deposition errors, significantly improving clinical efficacy compared to single therapies.

[0018] The biosafety barrier enables rapid response to impedance anomalies, while HRV analysis improves the accuracy of identifying physiological stress states, with an extremely low misidentification rate for children's patterns, ensuring treatment safety. Spatiotemporal rhythm regulation, based on atomic clock synchronization and optogenetics, significantly increases melatonin secretion and prolongs the duration of therapeutic effects. The overall system significantly improves treatment effectiveness and patient satisfaction, providing a comprehensive technical solution for intelligent Traditional Chinese Medicine (TCM) physiotherapy. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a schematic block diagram of an AI-assisted acupoint positioning physiotherapy system proposed in the present invention; Figure 2 This is a schematic block diagram of an AI-assisted acupoint positioning physiotherapy method proposed in the present invention; Figure 3 This is a bar chart comparing the delay time of traditional AR physiotherapy equipment and the interaction technology of this system in different interaction scenarios; Figure 4 The following is a line chart comparing the pressure resolution of traditional pressure sensors and the intelligent deformation ring sensor of this system in different scenarios. DETAILED DESCRIPTION

[0020] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0021] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise", "counterclockwise" and the like to indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0022] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the said features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be a connection between the two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below with reference to the accompanying drawings.

[0023] Reference Figures 1 to 4 : An AI-assisted acupoint positioning physical therapy system, including the following modules: The holographic interaction module utilizes HoloLens 2 AR glasses and features a 2K×2K resolution see-through display, ensuring high-definition image presentation while strictly controlling latency to within 16ms, guaranteeing smooth real-time interaction. The device integrates laser SLAM and visual inertial odometry (VIO), rapidly scanning and modeling limbs at a 15Hz frequency. The 3D point cloud has minimal error and accurately reproduces human anatomy.

[0024] In terms of interactive functions, gesture recognition relies on the MobileNetV3+SSD deep learning model, running on the NVIDIA Jetson AGX Orin platform, capable of real-time detection of up to 20 types of gestures and accurately identifying user operation intentions. The eye tracking system uses a four-channel infrared light source to achieve high-precision eye positioning, monitoring subtle changes in pupil diameter, and assisting in precise interactive control.

[0025] The software is developed based on Unreal Engine 4.27 and features a built-in VisibleHumanProject voxel model. Its fine voxel resolution supports dynamic rendering of muscle fiber orientation and nerve bundle distribution, visualizing the internal structure of the human body. Meridian projection utilizes GPU particle effects technology, mapping blood flow perfusion status in real time through particle color changes. The force feedback gloves utilize advanced pneumatic tendon drive technology and are equipped with 24 independent air chambers to simulate a variety of pressure levels. Verified by electromyographic (EMG) feedback, their acupressure simulation is highly reproducible to realistic tactile sensations such as soreness, swelling, and numbness, achieving a highly similar experience at the neuroelectrophysiological level.

[0026] Smart Deformable Ring: The main material is a composite shape memory polymer (SMP) composed of polycaprolactone (PCL) and graphene nanosheets (GNPs). PCL imparts biocompatibility and degradability to the material, while GNPs enhance its electrical conductivity, thermal conductivity, and mechanical strength, giving the composite SMP both flexible deformability and intelligent thermal response. Produced through selective laser sintering (SLS) 3D printing, with a precisely controlled layer thickness of 50μm, the ring creates a hollow structure that conforms to the curved contours of the human acupuncture point area. The material has a glass transition temperature (Tg) of 37±2°C (close to human body temperature) and a thermal response time of <8s. The shape memory effect can be triggered by body temperature or an external heat source, enabling dynamic conformation to the acupuncture points and adaptive pressure adjustment.

[0027] The built-in 96-channel flexible electronic skin integrates a capacitive MEMS pressure sensor array and a platinum-rhodium alloy thin-film temperature sensor. The capacitive MEMS array detects pressure through capacitance changes within its micro-nanostructures, with a resolution of 0.03kPa, accurately capturing subtle pressure fluctuations associated with acupoint massage. The platinum-rhodium alloy thin film utilizes the thermoelectric effect to detect acupoint temperature with an accuracy of ±0.05°C. A 500Hz sampling rate ensures real-time monitoring of the physiological thermal field. Data is transmitted via 2.4GHz Bluetooth low energy (latency <10ms), supporting simultaneous multi-sensor data acquisition and edge computing. This provides high-frame-rate, low-latency physiological signal input for therapeutic feedback control, enabling dual-modal "pressure-temperature" physiological sensing.

[0028] The control core is based on an STM32H7A3 microcontroller and features a fuzzy adaptive PID algorithm. This algorithm uses fuzzy logic to automatically tune PID parameters in real time, adapting to the nonlinear dynamic changes in acupoint pressure and temperature during treatment. When a temperature change rate greater than 0.7°C / min (indicating an abnormal physiological state) and a pressure distribution coefficient of variation greater than 0.2 (indicating decreased fit stability) are detected in the acupoint area, a 0.2mm diameter Ni-Ti shape memory alloy (SMA) spring is activated, achieving stepless tension adjustment from 0 to 20N in 0.6 seconds. Leveraging its thermally induced phase change properties, the SMA spring rapidly deforms when heated by an electric current, precisely compensating for the dynamic pressure demands of the acupoint area and ensuring stable therapeutic force output, forming a closed-loop "perception-computation-adjustment" control system.

[0029] The surface of the ring sleeve is coated with a polydopamine-silver nanoparticle (PDA-AgNP) composite coating. PDA strongly adheres to the substrate through catechol groups, constructing a three-dimensional network structure to load AgNPs. AgNPs achieve broad-spectrum antibacterial properties by destroying bacterial cell membranes and DNA. At the same time, the biocompatibility of PDA reduces the risk of skin irritation. The coating adapts to the 0-90° range of motion of the finger joints. Through bionic texture and elastic modulus optimization, it ensures uniform stress distribution during dynamic fitting. The skin sensitization rate after long-term wear is less than 0.1%, meeting medical-grade biosafety requirements. In addition, the SLS-printed hollow structure has built-in micro-heat dissipation channels. Combined with the thermal response characteristics of SMP, it enhances air convection through shape memory recovery when physical therapy energy is input to avoid overheating of the skin. The temperature sensor is linked to feedback to adjust the physical therapy parameters to improve thermal safety and energy utilization efficiency.

[0030] Quantum correlation syndrome differentiation module: A system that deeply integrates traditional Chinese medicine meridian theory with modern intelligent computing has been constructed. First, a traditional Chinese medicine meridian knowledge graph was built based on the Neo4j5.1 graph database. This graph is based on the attribute graph model and stores in detail the characteristic information of 361 acupoint entities, the topological connection relationship of 12 main meridians, and more than 3,000 association rules between symptoms and acupoints, forming a knowledge base for traditional Chinese medicine physiotherapy syndrome differentiation. In order to achieve efficient use of knowledge, the entities and relationships in the graph are embedded through the Transformer architecture and converted into a 512-dimensional vector representation, so that the cosine similarity calculation error between entity vectors is controlled within an extremely small range, thereby providing high-precision knowledge mapping capabilities for subsequent syndrome differentiation calculations.

[0031] On this basis, the module deploys a quantum neural network with 7 quantum bit layers (a total of 128 quantum bits), uses the time-varying quantum eigenvalue solver (TV-VQE) to optimize the network parameters, and uses the IBM Qiskit quantum computing framework to realize the quantum state inner product calculation between the disease and the acupoint. Its core formula is ,in, is the inner product of the quantum state, used to measure the symptom state Acupoint state This quantum computing method breaks through the bottleneck of traditional algorithms, transforming the TCM syndrome differentiation process into a quantum state matching problem. It greatly improves the computational accuracy and efficiency of acupoint matching schemes, making the analysis of the association between symptoms and acupoints more scientific and quantitative.

[0032] To enhance the accuracy of syndrome differentiation, the module integrates pulsed infrared thermal imaging technology. Using the FLIRA8580sc thermal imaging device (640×512 pixel resolution, 60Hz frame rate, and equipped with an 8-14μm band filter), it can precisely detect temperature differences across acupoint areas with an accuracy of 0.03°C. By combining real-time acupoint temperature data with the syndrome differentiation results of a quantum neural network, acupoint selection schemes are dynamically optimized using the Monte Carlo Tree Search (MCTS) algorithm. This correction mechanism, based on multimodal data (quantum state matching results and infrared thermal imaging data), significantly improves efficiency compared to traditional rule-based matching methods. It enables rapid adjustment of treatment plans based on the patient's real-time physiological state, ensuring accurate and personalized acupoint selection.

[0033] In terms of knowledge updating and system evolution, the knowledge graph is automatically updated daily using remote supervised learning. Newly added symptom-acupoint association rules undergo three rounds of federated learning verification to ensure their effectiveness and generalization in multi-center data scenarios, and are then synchronized to edge devices with a latency of less than 0.5 seconds. This mechanism enables the system to quickly adapt to new clinical symptoms and treatment needs, continuously iterating dialectical algorithms to maintain the timeliness and accuracy of treatment plans. Even in an edge computing environment, devices can obtain the latest knowledge updates in real time, enabling continuous self-optimization of the therapy system, providing strong knowledge support and dynamic adaptability for intelligent TCM therapy.

[0034] Multi-field synergistic therapy module: The energy field integration module adopts a modular design. Its microsecond pulse electric field is generated by a solid-state Tesla coil. The field strength can be flexibly adjusted between 0 and 50 kV / m, the pulse width range is 1 microsecond to 200 microseconds, and the rising edge is extremely fast (less than 100 nanoseconds), which can quickly provide electrical energy stimulation; terahertz wave thermal transmission relies on a quantum cascade laser (QCL) array, with a frequency continuously adjustable from 0.1 to 10 THz and a power density of 0.1 to 10 mW / mm 2, through the metamaterial phase plate to achieve phase modulation from 0 to 2π (resolution of π / 16), the terahertz wave phase can be precisely controlled to adjust the energy deposition depth and distribution; the bionic sound wave is generated by the piezoelectric ceramic transducer, with a frequency range of 20 to 2000Hz and a sound pressure level of 40 to 90dB, supporting five-tone twelve-tone modulation, using the sound wave frequency to simulate traditional Chinese medicine sound wave therapy and bionic stimulation of the meridians. The six-modal treatment contact is manufactured using 3D micro-electromechanical system (MEMS) technology, with a diameter of 1 mm and an integrated 0.5mm 2 The terahertz antenna, 0.1mm-pitch electrode pair, and micro-speaker simultaneously output at least four energy fields, evenly distributed across space, ensuring stable and uniform delivery of therapeutic energy to acupuncture points. For patients with Yang deficiency, the terahertz wave phase is adjusted to π / 4, coupled with a 25kV / m pulsed electric field (50µs pulse width), resulting in energy deposition depth of approximately 1.8cm. For patients with Yin deficiency, the phase is adjusted to 7π / 8, resulting in a deposition depth of approximately 0.6cm. Ultrasound tomography has verified minimal error (less than 0.1cm). This module integrates multimodal energy fields such as pulsed electric fields, terahertz waves, and bionic sound waves, and uses advanced generation and control technologies, combined with constitution-adapted energy deposition optimization, to enable the treatment contacts to achieve multi-energy field integration and uniform output, providing diversified and precise energy support for intelligent Chinese medicine physiotherapy, improving the accuracy and effectiveness of physiotherapy, and assisting in the intelligent upgrade of Chinese medicine physiotherapy in energy field applications. Through the synergy of different energy fields and constitution-adapted adjustment, it meets personalized physiotherapy needs, enhances the physiotherapy effect, and ensures precise control of energy deposition, providing core technical support for multi-field collaborative physiotherapy of intelligent physiotherapy systems.

[0035] Biosafety Barrier Module: Impedance monitoring utilizes four-electrode frequency sweep technology, sweeping across a frequency range of 10 to 100 kHz. A 200-microampere excitation current is applied to detect changes in human impedance in real time. When the system detects an impedance value below 500 ohms or above 10 kiloohms, with an impedance phase angle less than -30 degrees, indicating a possible skin contact abnormality or physiological risk, the field-programmable gate array (FPGA) rapidly cuts off energy output within 0.05 seconds, providing immediate safety protection. HRV (heart rate variability) analysis utilizes an improved approximate entropy algorithm (ApEn), setting an embedding dimension of 3 and a threshold of 0.15 standard deviations. Heart rate signal characteristics are continuously monitored at a high sampling rate of 1000 Hz. When an ApEn value greater than 1.7 is detected, indicating excessive physiological stress, the system automatically triggers 4000 Hz white noise (55 dB SPL, 100 Hz chirp rate) for emotional soothing and simultaneously provides 0.5 Hz breathing guidance. Through this dual adjustment method, the ApEn value is reduced by more than 50% within 5 minutes, effectively alleviating the patient's physiological stress. Child mode utilizes 3D structured light and infrared ToF (time-of-flight) dual-modal recognition technology with a wide 90° x 70° field of view. It is trained on the EfficientNet-B4 model using 200,000 facial images of children aged 0 to 12, ensuring the recognition algorithm accurately distinguishes children of different age groups.

[0036] A dual protection mechanism is immediately activated when a non-guardian attempts to operate the device: First, the device is physically locked, preventing any therapeutic operation; second, a real-time alert is sent to the parent via the 5G network, simultaneously transmitting a 1080p HD video stream. This ensures that guardians receive notification of any abnormal operation within one second, comprehensively safeguarding the child's safety. This comprehensive biosafety barrier technology utilizes multi-dimensional physiological signal monitoring, intelligent algorithm analysis, and a hierarchical protection strategy to establish a comprehensive safety system encompassing risk detection, emergency response, status adjustment, and protection of special populations, providing reliable security for the clinical application of intelligent therapy equipment.

[0037] The present invention also includes a spatiotemporal rhythm regulation module: The time synchronization system utilizes a rubidium atomic clock with dual-frequency timing from both GPS and Beidou as its core component, achieving a time synchronization accuracy of ±0.3 nanoseconds. The system automatically calibrates every 10 minutes to ensure high consistency of time references across all modules. In line with the Traditional Chinese Medicine (TCM) concept of "harmony between heaven and man," the system incorporates the theory of qi and blood circulation in the Huangdi Neijing (Huangdi Neijing) to construct a meridian peak-time calculation model, enabling precise determination of the peak periods of qi and blood flow in each meridian.

[0038] The wave music generation module is based on real-time EEG feedback and effectively reduces environmental noise interference through adaptive Wiener filtering technology to ensure that the collected EEG signals are clear and reliable. The system can dynamically adjust the music frequency according to the user's current EEG characteristics, covering 8 to 12 Hz. The frequency range is wide, with frequency adjustment steps as fine as 0.1Hz, accurately matching the EEG rhythms of different individuals. Music played through bone conduction headphones maintains a sound pressure level of 45 to 50dB, ensuring clarity while avoiding overstimulation of the ear canal. Furthermore, the headphones minimize harmonic distortion, resulting in pure and natural sound quality, providing users with a comfortable listening experience.

[0039] The optogenetic regulation module uses a 470nm blue light and 590nm yellow light LED matrix as the light source. The peak wavelength error of the blue light is controlled within ±5nm, and the luminous flux ranges from 100 to 500lux. The peak wavelength of the yellow light also maintains an accuracy of ±5nm, and the luminous flux ranges from 50 to 300lux. The duty cycle of the LED matrix can be flexibly adjusted between 0 and 100%, and the light intensity and duration can be accurately controlled according to the needs of physical therapy. The entire system is synchronized with high-precision time, personalized Wave music regulation and precise optogenetic intervention have built a closed-loop system of "time-neuro-endocrine" coordinated regulation, deeply integrating the theory of traditional Chinese medicine time medicine with modern optogenetics and EEG feedback technology, providing scientific and effective technical support for the spatiotemporal rhythm regulation of intelligent physical therapy, and helping to achieve personalized physical therapy plans that are more in line with the physiological rhythm of the human body.

[0040] The present invention also includes a nano-targeted drug delivery module: the microfluidic drug reservoir is precision-manufactured using PDMS / SU-8 multilayer soft lithography technology. It integrates 25 independent liquid storage chambers, each with a capacity of 50 microliters. The flow channel network is equipped with a fishbone-shaped mixer measuring 500 microns by 500 microns. The unique flow channel design ensures extremely uniform drug mixing. The electroporation technology module can output biphasic rectangular pulses with a voltage continuously adjustable between 100 and 300 volts and a pulse width of 10 to 100 milliseconds. The rising and falling edge times of the pulses are both less than 1 microsecond. Combined with a pulse train mode with a 0.5 millisecond interval, it can open nanochannels with a pore size of approximately 100 nanometers on the skin surface. These channels can persist for 5 minutes, creating an efficient transmission path for drug penetration.

[0041] Traditional Chinese medicine nanoparticles are prepared via an emulsification-solvent evaporation-lyophilization method, surface-modified with hyaluronic acid (HA) targeting ligands, and controlled to a particle size of 120±10 nanometers. Experimental verification shows that the skin penetration rate of danshensu nanoparticles is seven times higher than that of traditional decoctions. Near-infrared spectroscopy monitoring technology (wavelength 780 nanometers, sampling rate 1Hz) allows real-time tracking of the penetration process with minimal monitoring error. The drug reservoir replacement system utilizes a magnetically driven microvalves with a response time of less than 50 milliseconds, supporting automatic switching between 25 different drug solutions, with a single replacement time of less than 10 seconds. Raman spectroscopy also monitors drug solution composition in real time to ensure extremely high drug delivery accuracy. The entire microfluidic drug reservoir system, through sophisticated structural design, efficient electroporation technology, targeted nanoparticle preparation, and intelligent drug solution management, achieves precise mixing, efficient penetration, and safe switching of drug solutions. This provides stable and reliable technical support for drug delivery in intelligent physical therapy, significantly improving drug utilization efficiency and treatment safety.

[0042] In the present invention, a physical therapy method for AI-assisted acupoint positioning includes the following steps: Five-Sense Integration Diagnosis Steps: Tongue image acquisition utilizes a dual-spectral (RGB + near-infrared) industrial camera with a resolution of 4000×3000 and a frame rate of 30fps, clearly capturing tongue details. The camera is equipped with an improved U-Net model with a high Dice coefficient (≥0.95), accurately identifying 20 tongue colors and 25 tongue coating types, while maintaining a tongue segmentation error within 0.5mm, providing highly accurate visual data for tongue diagnosis.

[0043] Odor recognition utilizes a 48-sensor metal oxide semiconductor (MOS) array, combined with a dynamic time warping (DTW) algorithm, to discern 15 herbal odors with high accuracy. The sensor array self-cleans through periodic heating (400°C / 10s), resulting in a drift rate of less than 5% per month, ensuring long-term stable detection performance and effectively capturing human odor information for diagnostic analysis.

[0044] Multimodal fusion uses a dynamic weight algorithm, the formula is ( are the variances of vision, smell, and tongue data respectively), generating a four-dimensional syndrome differentiation vector (including the virtual and real dimensions), such as the wind-cold syndrome differentiation vector V = [−0.3, 0.2, −0.8, 0.6]. This dynamically integrates visual, olfactory, and tongue image data to improve the comprehensiveness and accuracy of syndrome differentiation. AI skeletal point tracking uses OpenPose and an improved YOLOv8, with an update frequency of 150Hz and an acupoint location prediction error of ≤ 0.1mm. It can track human skeletal points in real time, providing high-precision spatial data for acupoint positioning, ensuring accurate acupoint positioning during physical therapy.

[0045] Dynamic acupoint calibration steps: Multiple technologies collaborate to achieve precise, real-time adjustments. The flexible electronic skin within the ring continuously monitors temperature changes in the acupoint area. If a temperature change rate exceeding 0.5°C per minute is detected (indicating a potential state of active meridian qi or a significant physiological change in the area), the system immediately triggers the AI ​​skeletal point dynamic tracking module. This module utilizes high-precision tracking algorithms (such as a composite model combining OpenPose and an improved YOLOv8) to collect the coordinates of key skeletal points in real time at a high-frequency update rate of 100Hz. This allows for precise capture of the spatial dynamics of acupoints, even with slight movements or changes in body position. Simultaneously, the system utilizes a quantum state similarity algorithm to calculate the quantum state match between the current contact tip position and the target acupoint, assessing the positional deviation between the contact tip and the active meridian qi target in real time. Based on this assessment, the system drives a fine-tuning mechanism to precisely adjust the contact tip position, with an accuracy of less than 0.2mm. This closed-loop control process of "temperature detection trigger-dynamic tracking of bone points-quantum algorithm adjustment" ensures that the physical therapy contacts can quickly and accurately follow the dynamic changes of acupoints, so that energy can always act accurately on the target area where the meridian qi is active, effectively improving the pertinence and effectiveness of physical therapy, adapting to the body position changes or physiological state fluctuations that may occur during physical therapy, and providing key technical support for the dynamic adaptability of intelligent physical therapy.

[0046] Six-dimensional energy tuning steps: Through the collaboration of multiple algorithms, the precise optimization and dynamic adjustment of physical therapy energy are achieved. First, the six modal parameters are optimized based on the NSGA-III algorithm, and the population size is set to 100 and the number of iterations is set to 200. The fitness function ( The efficacy weight is 0.4, The safety weight is 0.3, The initial treatment sequence was generated using a Kalman filter (process noise covariance Q = 0.01I, measurement noise covariance R = 0.1I) with a comfort weight of 0.3; E = quantum state similarity, P = physiological stress index, and Q = patient feedback score. During treatment, parameters were adjusted in real time using a Kalman filter (process noise covariance Q = 0.01I, measurement noise covariance R = 0.1I). When the RMSSD of HRV decreased by more than 20%, the electric field intensity was automatically reduced by 10%. Energy gradient decay was performed using a double exponential function. ( is the initial field strength, =2s, =10s), combined with voice guidance (80 words / minute) and respiratory rate regulation (6-8 breaths / minute), the incidence of post-treatment discomfort is reduced. Through algorithm optimization, real-time feedback, and gradient attenuation, this process balances efficacy, safety, and comfort, improving energy regulation precision and patient experience, and providing a scientific framework for intelligent physical therapy energy regulation.

[0047] This invention also includes a cross-temporal and spatial efficacy evaluation step: the underlying blockchain uses Hyperledger Fabric 2.5, running with the PBFT consensus mechanism, with 7 nodes, 2 fault tolerances, and a block generation time of less than 2 seconds. Data encryption uses the SM4 algorithm (128-bit key length) to ensure data security and efficiency. The survival analysis model processes more than 150 dimensional features, and after XGBoost screening (importance threshold > 0.01), retains 28 key variables. The long-term efficacy prediction C-index reaches 0.89, accurately evaluating the therapeutic effect. Health management recommendations are based on the Transformer architecture (6 layers, 8 heads), combining patient genetic data (23andMe format) with lifestyle data to provide personalized guidance. Efficacy traceability is achieved through digital twins, with the mapping error between the virtual model and the physical entity's physiological indicators less than 5%, allowing for full-process traceability of therapeutic effects. The entire system builds a complete system from data security, efficacy prediction, health advice to effect tracing, to improve the scientificity and sustainability of intelligent Chinese medicine physiotherapy, provide data support for physiotherapy effect evaluation and health management, promote the intelligent development of physiotherapy, ensure data security and efficiency, accurate personalized advice, reliable effect tracing, and multi-link collaboration to enhance the overall effectiveness of the physiotherapy system.

[0048] The present invention also includes a meridian energy guidance step: at the end of the treatment, the meridian energy guidance step is started: the electric field intensity gradually decreases at a rate of 0.1kV / m / s to achieve gradient attenuation of energy. At the same time, the system provides voice guidance at a speaking speed of 80 words / minute, allowing the patient's mind to follow the direction of the meridians for meridian guidance, and adjusts the breathing frequency to 6-8 times / minute to allow the energy to dissipate naturally along the meridians. This combination of physical energy attenuation and mind and breathing guidance promotes the orderly flow of energy and effectively reduces discomfort after treatment. This step integrates the traditional Chinese medicine concept of "guiding and moving qi" with modern technology. Through gradient energy control, voice guidance and breathing regulation, it achieves the orderly conclusion of physical therapy energy, enhances the physical therapy effect, and provides guarantees for the comfort and safety of intelligent physical therapy. It embodies the synergistic effect of Chinese and Western medicine technologies in the final stage of physical therapy, making energy dissipation more in line with the operation rules of the human meridians, reducing postoperative discomfort and improving the overall physical therapy experience.

[0049] The above are only preferred specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solutions and inventive concepts of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. An AI-assisted acupoint positioning physical therapy system, characterized in that: Includes the following modules: Holographic interaction module: Equipped with AR glasses, it uses SLAM technology to reconstruct a 3D model of the user's limbs in real time, superimposing dynamic meridian projections. Users can slide gestures or select symptoms through eye tracking, and the system will simultaneously generate a holographic diagnosis interface that includes anatomical structures. Smart deformable ring: Made of shape memory polymer with a glass transition temperature of 37±2°C, it automatically conforms to the curve of the hand or wrist when exposed to body temperature. It has a built-in 48-channel flexible electronic skin, which can draw a real-time pressure-temperature matrix diagram of the acupoint area and dynamically adjust the ring tension. Quantum correlation syndrome differentiation module: Constructs a knowledge graph of TCM meridians and uses a quantum neural network algorithm to calculate the quantum state similarity between symptoms and acupoints. The formula is: ,in, is the inner product of the quantum state, For symptomatic state, For acupoint status, pulsed infrared thermal imaging is used to detect acupoint temperature differences and automatically correct acupoint matching plans; Multi-field synergistic therapy module: Integrates microsecond pulsed electric fields, terahertz wave thermal diathermy, and bionic sound wave technologies, combined with traditional electric pulses, moxibustion, and vibration modes to form a six-modal energy matrix. Each treatment contact independently outputs at least three energy fields, achieving a combined therapeutic effect of "acupuncture-moxibustion-medication-massage-sound-light"; Biosafety Barrier Module: Deploys an impedance-heart rate dual monitoring mechanism. When an impedance abnormality is detected, energy output is cut off within 0.1 seconds. Heart rate variability monitoring is combined with the Kullback-Leibler divergence algorithm to evaluate physiological stress responses in real time. When the calculated Dkl value is less than 0.3, it is determined to be a safe state. Facial recognition is enabled in child mode, and the device is automatically locked when operated by non-guardians.

2. The AI-assisted acupoint positioning physical therapy system according to claim 1, characterized in that: It also includes a spatiotemporal rhythm regulation module: a built-in atomic clock automatically calculates the best treatment period based on the meridian flow theory in the Yellow Emperor's Classic of Internal Medicine, and plays the rhythm during treatment. Wave Music regulates melatonin secretion through optogenetic technology and enhances the duration of therapeutic effect.

3. The AI-assisted acupoint positioning physical therapy system according to claim 1, characterized in that: It also includes a nano-targeted drug delivery module: the ring-shaped contact has a built-in microfluidic drug reservoir, which uses electroporation technology to open the skin channel and deliver Chinese medicine nanoparticles in a targeted manner. The drug reservoir supports the replacement of 20 types of drug solutions, and the drug solution permeability is monitored in real time through near-infrared spectroscopy.

4. The AI-assisted acupoint positioning physical therapy system according to claim 1, characterized in that: In the holographic interactive module, the AR meridian projection has the function of augmented reality acupuncture simulation. In terms of puncture path planning, advanced algorithms are used for intelligent calculation. At the same time, combined with force feedback gloves, the soreness, swelling and numbness caused by Qi are simulated, so that the similarity between the simulated feeling and the real experience reaches more than 90%.

5. The AI-assisted acupoint positioning physical therapy system according to claim 1, characterized in that: In the quantum correlation diagnosis module, the quantum neural network contains 5 quantum bit layers, uses a variational quantum algorithm to optimize parameters, and the update cycle of the disease-acupoint association rules is ≤24 hours. New rules are automatically synchronized to all devices.

6. The AI-assisted acupoint positioning physical therapy system according to claim 1, characterized in that: In the multi-field collaborative therapy module, terahertz wave thermal penetration is combined with adaptive phase modulation technology to adjust the wavefront phase according to different body constitutions, so that energy is selectively deposited in the deep layers of the meridians.

7. The AI-assisted acupoint positioning physical therapy system according to claim 1, characterized in that: In the biosafety barrier module, HRV analysis uses the approximate entropy algorithm ApEn. When ApEn>1.5, it is determined to be a stress state, and the white noise soothing program is automatically triggered to reduce ApEn to a safe range within 5 minutes, that is, ApEn<1.

0.

8. A method for applying the AI-assisted acupoint positioning physical therapy system according to any one of claims 1 to 7, characterized in that: The following steps are involved: Five-sense fusion syndrome differentiation steps: The user wears AR glasses and looks at the holographic meridian model, while simultaneously inhaling nine herbal scents and taking a tongue image. The system uses a multimodal data fusion algorithm with a visual weighting coefficient of 0.4, an olfactory weighting coefficient of 0.3, and a tongue image weighting coefficient of 0.3 to ultimately generate a personalized syndrome differentiation vector V=[v1,v2,v3]; Dynamic acupoint calibration steps: The ring-shaped electronic skin detects the temperature change rate of the acupoint area, triggers AI dynamic tracking of bone points, and combines the quantum state similarity algorithm to adjust the contact position in real time to ensure that energy is accurately applied to the active meridian target; Six-dimensional energy tuning steps: Based on the syndrome differentiation vector V, the six-modal energy parameters are optimized through genetic algorithms to generate a personalized treatment sequence. During the treatment, skin impedance and HRV physiological indicators are collected every 2 minutes to dynamically adjust the energy output.

9. The method of the AI-assisted acupoint positioning physical therapy system according to claim 8, characterized in that: It also includes steps for cross-temporal and spatial efficacy evaluation: establishing a blockchain medical data chain, uploading patient treatment data to the cloud after encryption through federated learning, predicting long-term efficacy through survival analysis algorithms, and automatically pushing 3-month health management recommendations after the treatment is completed.

10. The method of the AI-assisted acupoint positioning physical therapy system according to claim 8, characterized in that: It also includes the meridian energy guidance step: at the end of the treatment, the gradient energy attenuation program is started, and at the same time, the patient's thoughts are guided through the meridians through voice guidance, combined with breathing frequency adjustment, so that the energy can dissipate naturally along the meridians, reducing discomfort after treatment.

Images