Acupuncture robot meridian and acupoint positioning system based on digital twinning

Abstract

The application discloses a meridian and acupoint positioning system of an acupuncture robot based on digital twinning, and belongs to the technical field of acupuncture, and comprises: a data acquisition and digital modeling module, which is used for acquiring visual data of a patient's body surface, reconstructing a 3D digital model of the patient, and creating an initial digital twin model of the patient; a treatment planning simulation and rehearsal module, which is used for doctors to diagnose and design an acupuncture treatment scheme, and simulate and rehearse the whole acupuncture process; a virtual-real synchronous precise positioning module, which is used for controlling the acupuncture robot to move above a target acupoint, tracking the slight movement of the patient in real time, and synchronously updating the target position of the digital twin model of the patient and the acupuncture robot, so as to realize dynamic tracking and precise positioning of the meridian and acupoint. The application solves the problem that the existing acupuncture robot cannot realize precise positioning of the meridian and acupoint of a patient based on digital twinning, and leads to low efficiency of acupuncture treatment of the patient. The application can realize precise positioning of the meridian and acupoint of the patient based on digital twinning, and can improve the efficiency of acupuncture treatment of the patient.

Description

Technical Field

[0001] This invention relates to the field of acupuncture technology, specifically to a meridian and acupoint positioning system for acupuncture robots based on digital twins. Background Technology

[0002] Acupuncture is a general term for needling and moxibustion. Needling refers to inserting needles (usually filiform needles) into the patient's body at a certain angle under the guidance of traditional Chinese medicine theory, and using needling techniques such as twisting and lifting to stimulate specific parts of the body to achieve the purpose of treating diseases. The insertion point is called acupoint, or simply acupuncture point. When patients receive acupuncture, doctors usually perform acupuncture based on their experience.

[0003] Existing technologies cannot accurately locate acupoints and meridians in patients based on digital twins, resulting in low efficiency of acupuncture treatment. Summary of the Invention

[0004] The purpose of this invention is to provide a meridian and acupoint positioning system for acupuncture robots based on digital twins, which can achieve precise positioning of the patient's meridians and acupoints based on digital twins, improve the efficiency of acupuncture treatment for patients, and solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A digital twin-based acupuncture robot meridian and acupoint positioning system includes: The data acquisition and digital modeling module is used to acquire visual data of the patient's body surface based on a visual sensor, reconstruct the patient's 3D digital model, and map a standardized meridian and acupoint database onto the patient's 3D digital model to create an initial digital twin model of the patient. The treatment planning simulation module is used by doctors to make diagnoses and design acupuncture treatment plans, set the acupuncture depth, angle and expected techniques, simulate the entire acupuncture process, assess whether the path is safe and predict tissue deformation and stress. The virtual-real synchronization precision positioning module is used to control the acupuncture robot to move above the target acupoint, track the patient's slight movements in real time, and update the patient's digital twin model and the target position of the acupuncture robot in sync, so as to achieve dynamic tracking and precise positioning of meridian acupoints. The safe acupuncture real-time interaction module is used to provide real-time feedback of acupuncture sensation data during acupuncture by the acupuncture robot, and compare it with the predicted values ​​of the patient's digital twin model. Based on the comparison, the acupuncture treatment plan is automatically adjusted to ensure patient safety.

[0006] Preferably, based on visual sensors, visual data of the patient's body surface is acquired, a 3D digital model of the patient is reconstructed, and a standardized meridian and acupoint database is mapped onto the 3D digital model of the patient to create an initial digital twin model of the patient. The following operations are performed: The RGB-D camera and the thermal imaging camera acquire data at the same time through hardware trigger signals or software synchronization commands. The RGB-D camera acquires RGB color images and depth information of the patient's body surface and directly generates 3D point cloud data. The thermal imaging camera acquires temperature distribution data of the patient's body surface and generates thermal imaging data. Among them, a calibration plate with visible light feature points and heat source is used. Images of the calibration plate at different positions and angles are taken by an RGB-D camera and a thermal imaging camera. The transformation matrix that transforms the points in the thermal imaging camera coordinate system to the RGB-D camera coordinate system is calculated by the calibration algorithm. The relative position and attitude relationship between the RGB-D camera and the thermal imaging camera is found, and a unified coordinate system is established. Using the transformation matrix obtained from calibration, the temperature value of each pixel on the 2D thermal map acquired by the thermal imaging camera is mapped to the 3D point cloud data to obtain multimodal 3D point cloud data, in which each point contains not only spatial location and visual color, but also temperature information. Based on multimodal 3D point cloud data, a 3D digital model of the patient, incorporating shape, color, and temperature, is reconstructed in real time. A standardized meridian and acupoint database is then mapped onto the 3D digital model of the patient to create an initial digital twin model of the patient.

[0007] Preferably, the generated 3D point cloud data and thermal imaging data are preprocessed by performing the following operations: The generated 3D point cloud data and thermal imaging data are cleaned to remove noise and reduce noise interference. The generated 3D point cloud data and thermal imaging data are inspected to identify missing and outlier values. The missing and outlier values ​​in the 3D point cloud data and thermal imaging data are evaluated to determine whether they are valuable for the meridian and acupoint localization of the acupuncture robot. When missing values ​​and outliers in 3D point cloud data and thermal imaging data are valuable for the meridian and acupoint localization of acupuncture robots, the missing values ​​in 3D point cloud data and thermal imaging data are filled in, and the outliers in 3D point cloud data and thermal imaging data are corrected. When missing or outlier values ​​in 3D point cloud data and thermal imaging data are of no value for the meridian and acupoint localization of the acupuncture robot, the missing or outlier values ​​in the 3D point cloud data and thermal imaging data are removed.

[0008] Preferably, the generated 3D point cloud data and thermal imaging data are preprocessed, and the following operations are performed: Standardize 3D point cloud data and thermal imaging data to remove the dimensional differences between them and form standardized 3D point cloud data and thermal imaging data. 3D point cloud data and thermal imaging data are integrated into a unified data view to form patient surface visual data. The integrity of the integrated patient surface visual data is verified to determine whether there are any omissions. After the data integrity verification is qualified, the integrated patient surface visual data is distributed and securely stored.

[0009] Preferably, the generated 3D point cloud data and thermal imaging data are preprocessed, and the following operations are performed: Based on the thermal imaging data on the patient's 3D digital model, thermal feature vectors of acupoint areas are extracted. The thermal feature vectors include: mean temperature, temperature gradient, and thermal distribution entropy. The thermal feature vector is compared with the standard vector in the predefined standard acupoint thermal feature library, and the Euclidean distance is calculated. When the Euclidean distance is greater than the preset Euclidean distance threshold, it is determined whether the temperature deviation is greater than the preset temperature deviation threshold. If so, the acupoint coordinates are moved towards the proximal end along the meridian direction to determine the moving distance. When the Euclidean distance is greater than the preset Euclidean distance threshold, determine whether the gradient deviation is greater than the preset gradient deviation threshold. If so, adjust the acupoint depth and determine the adjustment amount. The coordinates of acupoints on the patient's 3D digital model are dynamically adjusted based on the movement distance and adjustment amount. The adjusted acupoint coordinates are synchronized to the digital twin model for subsequent acupuncture path planning.

[0010] Preferably, the doctor conducts diagnosis and designs acupuncture treatment plans, sets the needle insertion depth, angle, and expected techniques, simulates the entire acupuncture process, and performs the following operations: The patient's digital twin model is rendered using a 3D graphics engine, giving it realistic skin tone and texture, and including meridian points. Based on the patient's digital twin model, doctors conduct patient diagnosis and design acupuncture treatment plans. In this process, doctors use a mouse to click on the patient's digital twin model. Through ray projection technology, rays are emitted from the mouse position and collide with the surface of the patient's digital twin model. The acupoints hit by the rays are selected and highlighted. After selecting the acupoint, perform parametric design. Set the depth using a slider or input box, and intuitively adjust the needle insertion direction or directly input the angle value by dragging the 3D arrow icon. Select the twisting tonifying method or the lifting and thrusting reducing method from the drop-down menu, and set the amplitude and frequency parameters. After parametric design, simulation is performed. Virtual needles are generated at selected acupoints along a set angle. Based on finite element analysis, the soft tissue is divided into thousands of tiny mesh units, and each unit is assigned realistic material properties. The virtual needles are driven to move along the path by the physics engine, and the tissue deformation and needle stress under different manipulations are observed. By simulating the up-and-down reciprocating motion of the virtual needles, the dynamic changes in tissue resistance during each insertion and withdrawal are calculated. By simulating the rotation of the virtual needles, the torque required for rotation is calculated based on the friction model inside the soft tissue.

[0011] Preferably, after selecting the acupoint, perform parametric design and execute the following operations: Obtain basic attribute information of patients and information obtained from doctors' medical history inquiries about patients; Based on the knowledge graph of the selected acupoints, the associated graph items are determined according to the basic attribute information and medical history inquiry information; Determine the first free display area of ​​the parametric design interface; Determine whether a second free display area exists within the first free display area; the second free display area is a continuous area that can just accommodate all associated atlas items; If it does not exist, determine a locally contiguous third free display area within the first free display area; Get the target distance between the third free display area and the input box; Traverse the third free display area in order of target distance from nearest to farthest, and use the third free display area being traversed as the fourth free display area; Obtain the information content of local subnetworks of associated graph items; Based on the order of the network hierarchy of the local subnetwork in the knowledge graph from largest to smallest, the area of ​​the fourth free display region and the information content of the local subnetwork are matched to determine the target local subnetwork with the maximum information content that does not exceed the doctor's cognitive load. Map the target local subnet onto the fourth free display area, and display the remaining subnets in the local subnets other than the target local subnet around the fourth free display area with a preset first transparency; Stop traversing once all local subnetworks have been mapped; If the local subnet is not fully mapped after all the third free display areas have been traversed, then the display corner area closest to the last traversed third free display area is determined. The unmapped local subnetworks are displayed in the corner area of ​​the display based on a preset second transparency, wherein the first transparency is less than the second transparency; Once all mappings or displays are complete, they assist doctors in parametric design.

[0012] Preferably, to assess the safety of the path and predict tissue deformation and stress, the following operations are performed: During collision detection, the path and depth of the virtual needle tip are continuously monitored to determine whether the needle tip will collide with the bones or vital organs embedded in the patient's digital twin model. If a collision risk is detected, the path will be highlighted in red and a warning will be issued. During force analysis, the force on the virtual needle is monitored in real time, and the relationship curve between force and depth is plotted. When the force value suddenly increases sharply at a certain depth, a warning is issued, indicating that fascia or bone has been encountered. When the calculated torque exceeds the preset safety threshold, a warning is issued, indicating that the technique is too heavy, resulting in excessive tissue entanglement or needle breakage. The entire simulation process is presented in a visual format. When the simulation results show that the path is unsafe or the force is abnormal, the doctor immediately modifies the parameters and performs the simulation again until a safe and ideal acupuncture treatment plan is obtained, thus forming a closed-loop management of the patient's acupuncture simulation.

[0013] Preferably, the acupuncture robot is controlled to move above the target acupoint, tracking the patient's minute movements in real time, and simultaneously updating the patient's digital twin model and the target position of the acupuncture robot to achieve dynamic tracking and precise positioning of the meridian acupoints, performing the following operations: The patient lies down in the treatment position, the visual sensor scans the patient and generates a high-precision, static reference point cloud, and based on this, an initial digital twin is created, and the digital twin, acupuncture robot and patient are initially aligned in the same world coordinate system. A high-speed, low-latency RGB-D camera continuously acquires real-time point cloud data of the patient's body surface at a rate of 30 frames per second. The real-time point cloud of the current frame is matched with the initial reference point cloud to calculate the optimal rigid body transformation matrix, which describes the rotational and translational movements that the patient needs to perform in order to return to the initial position. The calculated transformation matrix is ​​directly applied to the digital twin, so that no matter how the patient moves, the digital twin always maintains synchronization with the patient in position and posture, reflecting the patient's real-time status. Based on the real-time coordinates of acupoints on the current digital twin, the system continuously receives updated target points and calculates the joint movement commands of the acupuncture robot in real time based on the new target points. The system then drives the needles to continuously and stably track the moving target acupoints, thereby achieving precise positioning of the patient's meridian acupoints.

[0014] Preferably, the acupuncture robot provides real-time feedback of needling sensation data during acupuncture and compares it with the predicted values ​​from the patient's digital twin model. Based on the comparison, the acupuncture treatment plan is automatically adjusted to ensure patient safety, and the following operations are performed: After the patient's meridian acupoints are accurately located, the acupuncture robot begins to insert needles. Force sensors collect the actual torque data of the needles in real time. At the same time, the real-time physics engine of the patient's digital twin model calculates and predicts the force data based on the current acupuncture depth, speed, and tissue mechanical model. The actual torque data is compared with the predicted force data to determine the deviation; If the deviation is within the preset safety threshold, the acupuncture process conforms to the model prediction, the tissue characteristics are normal, and the acupuncture robot continues to operate. If the deviation exceeds the safety threshold, the safety mechanism is immediately triggered, sending an emergency stop command to the acupuncture robot, freezing all movement instantly. It also conducts cause analysis and rapid diagnosis, and makes adaptive adjustments based on the analysis and diagnosis to ensure the safety of patients receiving acupuncture treatment.

[0015] Compared with the prior art, the beneficial effects of the present invention are: This invention acquires visual data of the patient's body surface using a visual sensor, reconstructs a 3D digital model of the patient, and maps a standardized meridian and acupoint database onto the 3D digital model of the patient to create an initial digital twin model of the patient. A 3D graphics engine is used to render the patient's digital twin model. Based on the patient's digital twin model, doctors design diagnostic and acupuncture treatment plans, setting the acupuncture depth, angle, and expected techniques. The entire acupuncture process is simulated and rehearsed to assess the safety of the path and predict tissue deformation and stress, forming a closed-loop management system for patient acupuncture simulation and rehearsal. After the simulation, an acupuncture robot is controlled to move above the target acupoint, tracking the patient's minute movements in real time and synchronously updating the target position of the patient's digital twin model and the acupuncture robot, achieving dynamic tracking and precise positioning of meridians and acupoints. During acupuncture, the acupuncture robot provides real-time feedback of needling sensation data, which is compared with the predicted values ​​of the patient's digital twin model. Based on the comparison, the acupuncture treatment plan is automatically adjusted to ensure the safety of patient acupuncture treatment. This invention can achieve precise positioning of patient meridians and acupoints based on digital twins, improving the efficiency of patient acupuncture treatment. Attached Figure Description

[0016] Figure 1 This is a block diagram of the meridian and acupoint positioning system for acupuncture robots based on digital twins according to the present invention. Figure 2 This is a flowchart of the algorithm for the acupuncture robot of the present invention to provide real-time feedback and automatically adjust the acupuncture treatment plan during acupuncture. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] To address the current limitations of precise acupoint location based on digital twins, which leads to low efficiency in acupuncture treatment, please refer to [link to relevant documentation]. Figures 1-2 This embodiment provides the following technical solution: The acupuncture robot meridian and acupoint positioning system based on digital twins includes: a data acquisition and digital modeling module, a treatment planning simulation and pre-play module, a virtual-real synchronous precise positioning module, and a safe acupuncture real-time interaction module.

[0019] The data acquisition and digital modeling module is used to acquire visual data of the patient's body surface based on a visual sensor, reconstruct the patient's 3D digital model, and map a standardized meridian and acupoint database onto the patient's 3D digital model to create an initial digital twin model of the patient.

[0020] In this embodiment, visual data of the patient's body surface is acquired based on a visual sensor, a 3D digital model of the patient is reconstructed, and a standardized meridian and acupoint database is mapped onto the 3D digital model of the patient to create an initial digital twin model of the patient. The following operations are performed: The RGB-D camera and the thermal imaging camera acquire data at the same time through hardware trigger signals or software synchronization commands. The RGB-D camera acquires RGB color images and depth information of the patient's body surface and directly generates 3D point cloud data. The thermal imaging camera acquires temperature distribution data of the patient's body surface and generates thermal imaging data. Among them, a calibration plate with visible light feature points and heat source is used. Images of the calibration plate at different positions and angles are taken by an RGB-D camera and a thermal imaging camera. The transformation matrix that transforms the points in the thermal imaging camera coordinate system to the RGB-D camera coordinate system is calculated by the calibration algorithm. The relative position and attitude relationship between the RGB-D camera and the thermal imaging camera is found, and a unified coordinate system is established. Using the transformation matrix obtained from calibration, the temperature value of each pixel on the 2D thermal map acquired by the thermal imaging camera is mapped to the 3D point cloud data to obtain multimodal 3D point cloud data, in which each point contains not only spatial location and visual color, but also temperature information. Based on multimodal 3D point cloud data, a 3D digital model of the patient, incorporating shape, color, and temperature, is reconstructed in real time. A standardized meridian and acupoint database is then mapped onto the 3D digital model of the patient to create an initial digital twin model of the patient.

[0021] In this embodiment, the generated 3D point cloud data and thermal imaging data are preprocessed by performing the following operations: The generated 3D point cloud data and thermal imaging data are cleaned to remove noise and reduce noise interference. The generated 3D point cloud data and thermal imaging data are inspected to identify missing and outlier values. The missing and outlier values ​​in the 3D point cloud data and thermal imaging data are evaluated to determine whether they are valuable for the meridian and acupoint localization of the acupuncture robot. When missing values ​​and outliers in 3D point cloud data and thermal imaging data are valuable for the meridian and acupoint localization of acupuncture robots, the missing values ​​in 3D point cloud data and thermal imaging data are filled in, and the outliers in 3D point cloud data and thermal imaging data are corrected. When missing or outlier values ​​in 3D point cloud data and thermal imaging data are of no value for the meridian and acupoint localization of the acupuncture robot, the missing or outlier values ​​in the 3D point cloud data and thermal imaging data are removed.

[0022] In this embodiment, the generated 3D point cloud data and thermal imaging data are preprocessed, and the following operations are performed: Standardize 3D point cloud data and thermal imaging data to remove the dimensional differences between them and form standardized 3D point cloud data and thermal imaging data. 3D point cloud data and thermal imaging data are integrated into a unified data view to form patient surface visual data. The integrity of the integrated patient surface visual data is verified to determine whether there are any omissions. After the data integrity verification is qualified, the integrated patient surface visual data is distributed and securely stored.

[0023] Among them, the treatment planning simulation module is used by doctors to make diagnoses and design acupuncture treatment plans, set the acupuncture depth, angle and expected techniques, simulate the entire acupuncture process, assess whether the path is safe and predict tissue deformation and stress.

[0024] In this embodiment, the generated 3D point cloud data and thermal imaging data are preprocessed, and the following operations are performed: Based on the thermal imaging data on the patient's 3D digital model, thermal feature vectors of acupoint areas are extracted. The thermal feature vectors include: mean temperature, temperature gradient, and thermal distribution entropy value.

[0025] In this embodiment, the thermal feature vector is a three-dimensional vector describing the thermal imaging characteristics of the acupoint region. It is constructed from vector feature values ​​such as the average temperature, temperature gradient, and thermal distribution entropy. The specific construction rules (the correspondence between feature values ​​and vector element positions) are preset manually. The average temperature is the average temperature of all pixels in the acupoint region. The temperature gradient is the rate of temperature change in the image space, expressed in degrees Celsius per pixel. The thermal distribution entropy is an indicator of the uniformity of temperature distribution.

[0026] The thermal feature vectors are compared with standard vectors in a predefined standard acupoint thermal feature database, and the Euclidean distance is calculated. : ; in, The distance is Euclidean. For the acupoint area Temperature value per pixel This represents the total number of pixels in the acupoint area. The average temperature of the corresponding acupoint in the standard acupoint thermal feature database; This represents the spatial rate of temperature change along the horizontal direction of the image. This represents the spatial rate of temperature change in the vertical direction of the image, expressed in degrees Celsius per pixel. Standard temperature gradient; The first one in the temperature histogram indivual The probability, For temperature histogram total, This is the standard heat distribution entropy value.

[0027] In this embodiment, the predefined standard acupoint thermal feature library is a pre-established database that stores typical thermal feature data (average temperature, temperature gradient, and thermal distribution entropy value) of each standard acupoint in a healthy state, serving as a comparison benchmark.

[0028] When the Euclidean distance is greater than a preset Euclidean distance threshold, it is determined whether the temperature deviation is greater than a preset temperature deviation threshold. If so, the acupoint coordinates are moved proximally along the meridian direction to determine the moving distance. : ; in, This refers to the distance traveled along the meridian towards the proximal end. This is a preset scaling factor, in pixels per degree Celsius; When the Euclidean distance is greater than a preset Euclidean distance threshold, it is determined whether the gradient deviation is greater than a preset gradient deviation threshold. If so, the acupoint depth is adjusted, and the adjustment amount is determined. : ; in, This is the adjustment amount for the depth of the acupoint. This is the preset depth adjustment factor, in millimeters per degree Celsius; The coordinates of acupoints on the patient's 3D digital model are dynamically adjusted based on the movement distance and adjustment amount. The adjusted acupoint coordinates are synchronized to the digital twin model for subsequent acupuncture path planning.

[0029] The working principle and beneficial effects of the above technical solution are as follows: This invention uses thermal imaging data to correct acupoint locations in real time, ensuring accurate positioning of the acupuncture robot. Specifically, thermal feature vectors of the target acupoint region are extracted from thermal imaging data on the patient's 3D digital model. These extracted thermal feature vectors are compared with a standard acupoint thermal feature database to calculate the Euclidean distance. The Euclidean distance is compared with a preset threshold to determine if correction is needed. When the average temperature is below a certain threshold of the standard value, the acupoint position is moved proximally along the meridian direction; when the temperature gradient is above a certain threshold of the standard value, the acupoint depth is adjusted to avoid puncturing abnormal tissue. The corrected acupoint coordinates are synchronized to the digital twin model for subsequent acupuncture planning. This invention can respond to changes in the patient's surface thermal characteristics in real time, resulting in more accurate acupoint positioning. A specific implementation example is given below: Target acupoint: Patients with mild knee inflammation; Target acupoint: Zusanli (ST36); Standard parameters: , , Threshold settings: Euclidean distance threshold is 0.3, temperature deviation threshold is... The gradient deviation threshold is The correction factor is: pixels / , .

[0030] Thermal imaging data of a 2cm × 2cm area around the Zusanli acupoint was obtained by scanning the patient's lower leg, and the average temperature was calculated. Temperature gradient is The heat distribution entropy value is 0.8. Comparing the thermal feature vector with the standard vector yields a Euclidean distance of 0.707, which is greater than 0.3, indicating that correction is needed.

[0031] Next, planar position correction and depth correction are performed. During planar position correction, the temperature deviation is calculated to be 0.7. greater than the temperature deviation threshold This meets the planar correction conditions; further, the calculated movement distance is 1.4 pixels. During depth correction, the gradient deviation is calculated as follows: The depth correction conditions are not met.

[0032] The coordinates of the Zusanli acupoint are moved 1.4 pixels along the Stomach Meridian towards the proximal end (knee direction) to obtain the corrected acupoint coordinates. The corrected acupoint coordinates are then synchronized to the digital twin model.

[0033] In this embodiment, the doctor performs diagnosis and acupuncture treatment plan design, sets the needling depth, angle and expected technique, simulates the entire acupuncture process, and performs the following operations: The patient's digital twin model is rendered using a 3D graphics engine, giving it realistic skin tone and texture, and including meridian points. Based on the patient's digital twin model, doctors conduct patient diagnosis and design acupuncture treatment plans. In this process, doctors use a mouse to click on the patient's digital twin model. Through ray projection technology, rays are emitted from the mouse position and collide with the surface of the patient's digital twin model. The acupoints hit by the rays are selected and highlighted. After selecting the acupoint, perform parametric design. Set the depth using a slider or input box, and intuitively adjust the needle insertion direction or directly input the angle value by dragging the 3D arrow icon. Select the twisting tonifying method or the lifting and thrusting reducing method from the drop-down menu, and set the amplitude and frequency parameters. After parametric design, simulation is performed. Virtual needles are generated at selected acupoints along a set angle. Based on finite element analysis, the soft tissue is divided into thousands of tiny mesh units, and each unit is assigned realistic material properties. The virtual needles are driven to move along the path by the physics engine, and the tissue deformation and needle stress under different manipulations are observed. By simulating the up-and-down reciprocating motion of the virtual needles, the dynamic changes in tissue resistance during each insertion and withdrawal are calculated. By simulating the rotation of the virtual needles, the torque required for rotation is calculated based on the friction model inside the soft tissue.

[0034] In this embodiment, after selecting the acupoint, parameterized design is performed, and the following operations are performed: Obtain basic attribute information of patients and information obtained from doctors' medical history inquiries.

[0035] In this embodiment, the patient's basic attribute information refers to the patient's basic personal data, such as age, gender, weight, and height. The information obtained by the doctor from the patient's medical history inquiry is dynamic information collected by the doctor during the consultation process, including current symptoms, past medical history, family medical history, lifestyle habits, and treatment experience.

[0036] Based on the knowledge graph of the selected acupoints, the associated graph items are determined according to the basic attribute information and medical history inquiry information.

[0037] In this embodiment, the knowledge graph of the selected acupoint is a graph-based storage of relevant knowledge about the selected acupoint (such as the mapping between individual differences and acupoint locations). The associated graph items refer to the relevant nodes and edges in the knowledge graph that match the patient's specific information (basic attributes and medical history).

[0038] Determine the first free display area of ​​the parametric design interface.

[0039] In this embodiment, the parametric design interface is an interactive interface for doctors to adjust parameters (such as acupoint location and stimulation intensity). The first free display area is all the unoccupied blank areas on the parametric design interface.

[0040] Determine whether a second free display area exists within the first free display area; the second free display area is a continuous area that can just accommodate all associated atlas items.

[0041] If it does not exist, determine a locally contiguous third free display area within the first free display area.

[0042] Get the target distance between the third free display area and the input box.

[0043] Traverse the third free display area in order of distance from the target from closest to furthest, and use the third free display area being traversed as the fourth free display area.

[0044] Obtain the information content of local subnetworks of associated graph items.

[0045] In this embodiment, the local subnetwork is the graph network clustering result of the associated graph items. The information content is a complexity index of the local subnetwork, calculated based on the number of nodes, the number of edges, and information entropy, reflecting the richness of the content.

[0046] Based on the order of the network hierarchy of the local subnetworks in the knowledge graph from largest to smallest, the area of ​​the fourth free display region and the information content of the local subnetworks are matched to determine the target local subnetwork with the maximum information content that does not exceed the doctor's cognitive load.

[0047] In this embodiment, the doctor's cognitive load refers to the psychological burden threshold when a doctor processes information. The greater the information density within the fourth idle display area, the heavier the doctor's cognitive load. When matching the area and the information content of the local subnetwork, the target local subnetwork with the maximum information content that does not exceed the doctor's cognitive load is determined as the core display content.

[0048] The target local subnet is mapped onto the fourth free display area, and the remaining subnets in the local subnet, excluding the target local subnet, are displayed around the fourth free display area with a preset first transparency.

[0049] In this embodiment, mapping refers to the process of converting sub-networks (nodes and edges) in the knowledge graph into visual elements and rendering them onto the interface area. The preset first transparency is manually set and is used to display secondary content other than the core display content, making it visually more "fade in".

[0050] Stop traversing once all local subnetworks have been mapped.

[0051] If the local subnet is not fully mapped after all the third free display areas have been traversed, then the display corner area closest to the last traversed third free display area is determined.

[0052] In this embodiment, the display corner area refers to the corner area of ​​the display interface of the display device (such as the upper left corner, lower left corner, upper right corner, and lower right corner).

[0053] The unmapped local subnetworks are displayed in the corner area of ​​the display based on a preset second transparency, wherein the first transparency is less than the second transparency.

[0054] In this embodiment, the preset second transparency is also manually pre-set. The first transparency being less than the second transparency means that the first transparency represents a smaller degree of transparency. A display corner area near the last traversed third free display area is determined to ensure that the doctor can notice it but it does not dominate the interface.

[0055] Once all mappings or displays are complete, they assist doctors in parametric design.

[0056] The working principle and beneficial effects of the above technical solution are as follows: This invention addresses the parameterized design scenario of acupuncture robots by doctors, and designs a dynamic visualization and cognitive load optimization mechanism based on knowledge graphs, specifically including: First, knowledge reasoning is performed based on the patient's basic attribute information and the doctor's medical history. This stage narrows the knowledge graph scope using patient data, avoiding full-graph retrieval and improving efficiency. Next, an adaptive acupuncture knowledge layout is implemented. The system detects the first free display area (currently unoccupied continuous blank space) of the parameterized design interface. If this area contains a sub-area that can accommodate all associated graph items (the second free display area), the content is directly mapped; otherwise, a third free display area (a locally continuous sub-area) is determined. The distance between each third area and the input box (the interaction point for the doctor's input parameters) is calculated, and these sub-areas are traversed in order from closest to furthest. For each fourth free display area, the system matches the information content and area according to the network hierarchy of the local subnetwork of the associated graph item from largest to smallest. It determines that the information content of the target local subnetwork does not exceed the doctor's cognitive load and maps the target subnetwork to the fourth free display area for high-priority display. The remaining subnetworks are displayed with a first level of transparency, fading in from the periphery. This avoids poor knowledge absorption caused by excessive cognitive load when information density is too high, while also allowing doctors to seamlessly access related knowledge when they want to view it. For overflowing knowledge, it is displayed with a second level of transparency in the corner area near the last traversed third free display area, similarly achieving seamless access to overflowing knowledge. After all content is mapped, a visual interface guides the doctor to adjust parameters.

[0057] This invention integrates knowledge graphs, interface layout, and cognitive science to solve the problems of information overload, rigid layout, and low decision-making efficiency in traditional Chinese medicine auxiliary systems. It significantly reduces the cognitive load on doctors and further improves the efficiency and accuracy of treatment plan formulation.

[0058] In this embodiment, the safety of the path and the prediction of tissue deformation and stress are assessed by performing the following operations: During collision detection, the path and depth of the virtual needle tip are continuously monitored to determine whether the needle tip will collide with the bones or vital organs embedded in the patient's digital twin model. If a collision risk is detected, the path will be highlighted in red and a warning will be issued. During force analysis, the force on the virtual needle is monitored in real time, and the relationship curve between force and depth is plotted. When the force value suddenly increases sharply at a certain depth, a warning is issued, indicating that fascia or bone has been encountered. When the calculated torque exceeds the preset safety threshold, a warning is issued, indicating that the technique is too heavy, resulting in excessive tissue entanglement or needle breakage. The entire simulation process is presented in a visual format. When the simulation results show that the path is unsafe or the force is abnormal, the doctor immediately modifies the parameters and performs the simulation again until a safe and ideal acupuncture treatment plan is obtained, thus forming a closed-loop management of the patient's acupuncture simulation.

[0059] Among them, the virtual-real synchronous precision positioning module is used to control the acupuncture robot to move above the target acupoint, track the patient's slight movements in real time, and update the patient's digital twin model and the target position of the acupuncture robot in sync, so as to achieve dynamic tracking and precise positioning of meridian acupoints.

[0060] In this embodiment, the acupuncture robot is controlled to move above the target acupoint, tracking the patient's minute movements in real time, and synchronously updating the patient's digital twin model and the target position of the acupuncture robot to achieve dynamic tracking and precise positioning of the meridian acupoints, and performing the following operations: The patient lies down in the treatment position, the visual sensor scans the patient and generates a high-precision, static reference point cloud, and based on this, an initial digital twin is created, and the digital twin, acupuncture robot and patient are initially aligned in the same world coordinate system. A high-speed, low-latency RGB-D camera continuously acquires real-time point cloud data of the patient's body surface at a rate of 30 frames per second. The real-time point cloud of the current frame is matched with the initial reference point cloud to calculate the optimal rigid body transformation matrix, which describes the rotational and translational movements that the patient needs to perform in order to return to the initial position. The calculated transformation matrix is ​​directly applied to the digital twin, so that no matter how the patient moves, the digital twin always maintains synchronization with the patient in position and posture, reflecting the patient's real-time status. Based on the real-time coordinates of acupoints on the current digital twin, the system continuously receives updated target points and calculates the joint movement commands of the acupuncture robot in real time based on the new target points. The system then drives the needles to continuously and stably track the moving target acupoints, thereby achieving precise positioning of the patient's meridian acupoints.

[0061] Among them, the real-time interactive module for safe acupuncture is used to provide real-time feedback of acupuncture sensation data during acupuncture by the acupuncture robot, and compare it with the predicted value of the patient's digital twin model. Based on the comparison, the acupuncture treatment plan is automatically adjusted to ensure patient safety.

[0062] In this embodiment, the acupuncture robot provides real-time feedback of needle sensation data during acupuncture and compares it with the predicted values ​​from the patient's digital twin model. Based on the comparison, the acupuncture treatment plan is automatically adjusted to ensure patient safety, and the following operations are performed: After the patient's meridian acupoints are accurately located, the acupuncture robot begins to insert needles. Force sensors collect the actual torque data of the needles in real time. At the same time, the real-time physics engine of the patient's digital twin model calculates and predicts the force data based on the current acupuncture depth, speed, and tissue mechanical model. The actual torque data is compared with the predicted force data to determine the deviation; If the deviation is within the preset safety threshold, the acupuncture process conforms to the model prediction, the tissue characteristics are normal, and the acupuncture robot continues to operate. If the deviation exceeds the safety threshold, the safety mechanism is immediately triggered, sending an emergency stop command to the acupuncture robot, freezing all movement instantly. It also conducts cause analysis and rapid diagnosis, and makes adaptive adjustments based on the analysis and diagnosis to ensure the safety of patients receiving acupuncture treatment.

[0063] In summary, this method acquires visual data of the patient's body surface using visual sensors, reconstructs a 3D digital model of the patient, and maps a standardized meridian and acupoint database onto the 3D digital model to create an initial digital twin model of the patient. A 3D graphics engine is then used to render this digital twin model. Based on this model, doctors design diagnostic and acupuncture treatment plans, setting the needling depth, angle, and expected techniques. The entire acupuncture process is simulated and rehearsed to assess the safety of the path and predict tissue deformation and stress, forming a closed-loop management system for patient acupuncture simulation. After the simulation, an acupuncture robot is controlled to move above the target acupoint, tracking the patient's minute movements in real time and simultaneously updating the target position of both the patient's digital twin model and the acupuncture robot. This achieves dynamic tracking and precise positioning of meridians and acupoints. During acupuncture, the acupuncture robot provides real-time feedback of needling sensation data, which is compared with the predicted values ​​from the patient's digital twin model. Based on the comparison, the acupuncture treatment plan is automatically adjusted to ensure patient safety during acupuncture treatment. This method, based on digital twins, enables precise positioning of patient meridians and acupoints, improving the efficiency of acupuncture treatment.

[0064] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0065] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A meridian and acupoint positioning system for acupuncture robots based on digital twins, characterized in that, include: The data acquisition and digital modeling module is used to acquire visual data of the patient's body surface based on a visual sensor, reconstruct the patient's 3D digital model, and map a standardized meridian and acupoint database onto the patient's 3D digital model to create an initial digital twin model of the patient. The treatment planning simulation module is used by doctors to make diagnoses and design acupuncture treatment plans, set the acupuncture depth, angle and expected techniques, simulate the entire acupuncture process, assess whether the path is safe and predict tissue deformation and stress. The virtual-real synchronization precision positioning module is used to control the acupuncture robot to move above the target acupoint, track the patient's slight movements in real time, and update the patient's digital twin model and the target position of the acupuncture robot in sync, so as to achieve dynamic tracking and precise positioning of meridian acupoints. The safe acupuncture real-time interaction module is used to provide real-time feedback of acupuncture sensation data during acupuncture by the acupuncture robot, and compare it with the predicted value of the patient's digital twin model. Based on the comparison, the acupuncture treatment plan is automatically adjusted. Based on the thermal imaging data on the patient's 3D digital model, thermal feature vectors of acupoint areas are extracted. The thermal feature vectors include: mean temperature, temperature gradient, and thermal distribution entropy. The thermal feature vector is a three-dimensional vector describing the thermal imaging characteristics of the acupoint area. It is constructed from vector feature values ​​such as temperature mean, temperature gradient, and thermal distribution entropy. The temperature mean is the average temperature of all pixels in the acupoint area. The temperature gradient is the rate of change of temperature in the image space, in degrees Celsius per pixel. The thermal distribution entropy is an indicator of the uniformity of temperature distribution. The thermal feature vectors are compared with standard vectors in a predefined standard acupoint thermal feature database, and the Euclidean distance is calculated. : ； in, The distance is Euclidean. For the acupoint area Temperature value per pixel This represents the total number of pixels in the acupoint area. The average temperature of the corresponding acupoint in the standard acupoint thermal feature database; This represents the spatial rate of temperature change along the horizontal direction of the image. This represents the spatial rate of temperature change in the vertical direction of the image, expressed in degrees Celsius per pixel. Standard temperature gradient; The first one in the temperature histogram indivual The probability, For temperature histogram total, The standard heat distribution entropy value; The predefined standard acupoint thermal feature library is a pre-established database that stores typical thermal feature data of each standard acupoint under healthy conditions, serving as a comparison benchmark. When the Euclidean distance is greater than a preset Euclidean distance threshold, it is determined whether the temperature deviation is greater than a preset temperature deviation threshold. If so, the acupoint coordinates are moved proximally along the meridian direction to determine the moving distance. : ； in, This refers to the distance traveled along the meridian towards the proximal end. This is a preset scaling factor, in pixels per degree Celsius; When the Euclidean distance is greater than a preset Euclidean distance threshold, it is determined whether the gradient deviation is greater than a preset gradient deviation threshold. If so, the acupoint depth is adjusted, and the adjustment amount is determined. : ； in, This is the adjustment amount for the depth of the acupoint. This is the preset depth adjustment factor, in millimeters per degree Celsius; The coordinates of acupoints on the patient's 3D digital model are dynamically adjusted based on the movement distance and adjustment amount; the adjusted acupoint coordinates are then synchronized to the digital twin model for subsequent acupuncture path planning.

2. The acupuncture robot meridian and acupoint positioning system based on digital twin as described in claim 1, characterized in that, Based on visual data of the patient's body surface acquired by a visual sensor, a 3D digital model of the patient is reconstructed. A standardized meridian and acupoint database is then mapped onto the 3D digital model of the patient to create an initial digital twin model of the patient. The following operations are then performed: The RGB-D camera and the thermal imaging camera acquire data at the same time through hardware trigger signals or software synchronization commands. The RGB-D camera acquires RGB color images and depth information of the patient's body surface and directly generates 3D point cloud data. The thermal imaging camera acquires temperature distribution data of the patient's body surface and generates thermal imaging data. Among them, a calibration plate with visible light feature points and heat source is used. Images of the calibration plate at different positions and angles are taken by an RGB-D camera and a thermal imaging camera. The transformation matrix that transforms the points in the thermal imaging camera coordinate system to the RGB-D camera coordinate system is calculated by the calibration algorithm. The relative position and attitude relationship between the RGB-D camera and the thermal imaging camera is found, and a unified coordinate system is established. Using the transformation matrix obtained from calibration, the temperature value of each pixel on the 2D thermal map acquired by the thermal imaging camera is mapped to the 3D point cloud data to obtain multimodal 3D point cloud data, in which each point contains not only spatial location and visual color, but also temperature information. Based on multimodal 3D point cloud data, a 3D digital model of the patient, incorporating shape, color, and temperature, is reconstructed in real time. A standardized meridian and acupoint database is then mapped onto the 3D digital model of the patient to create an initial digital twin model of the patient.

3. The acupuncture robot meridian and acupoint positioning system based on digital twins according to claim 2, characterized in that, The generated 3D point cloud data and thermal imaging data are preprocessed by performing the following operations: The generated 3D point cloud data and thermal imaging data are cleaned to remove noise and reduce noise interference. The generated 3D point cloud data and thermal imaging data are inspected to identify missing and outlier values. The missing and outlier values ​​in the 3D point cloud data and thermal imaging data are evaluated to determine whether they are valuable for the meridian and acupoint localization of the acupuncture robot. When missing values ​​and outliers in 3D point cloud data and thermal imaging data are valuable for the meridian and acupoint localization of acupuncture robots, the missing values ​​in 3D point cloud data and thermal imaging data are filled in, and the outliers in 3D point cloud data and thermal imaging data are corrected. When missing or outlier values ​​in 3D point cloud data and thermal imaging data are of no value for the meridian and acupoint localization of the acupuncture robot, the missing or outlier values ​​in the 3D point cloud data and thermal imaging data are removed.

4. The acupuncture robot meridian and acupoint positioning system based on digital twins according to claim 3, characterized in that, The generated 3D point cloud data and thermal imaging data are preprocessed, and the following operations are performed: Standardize 3D point cloud data and thermal imaging data to remove the dimensional differences between them and form standardized 3D point cloud data and thermal imaging data. 3D point cloud data and thermal imaging data are integrated into a unified data view to form patient surface visual data. The integrity of the integrated patient surface visual data is verified to determine whether there are any omissions. After the data integrity verification is qualified, the integrated patient surface visual data is distributed and securely stored.

5. The acupuncture robot meridian and acupoint positioning system based on digital twins according to claim 4, characterized in that, The doctor conducts diagnosis and designs acupuncture treatment plans, sets the needle insertion depth, angle, and expected techniques, and simulates the entire acupuncture process, performing the following operations: The patient's digital twin model is rendered using a 3D graphics engine, giving it realistic skin tone and texture, and including meridian points. Based on the patient's digital twin model, doctors conduct patient diagnosis and design acupuncture treatment plans. In this process, doctors use a mouse to click on the patient's digital twin model. Through ray projection technology, rays are emitted from the mouse position and collide with the surface of the patient's digital twin model. The acupoints hit by the rays are selected and highlighted. After selecting the acupoint, perform parametric design. Set the depth using a slider or input box, and intuitively adjust the needle insertion direction or directly input the angle value by dragging the 3D arrow icon. Select the twisting tonifying method or the lifting and thrusting reducing method from the drop-down menu, and set the amplitude and frequency parameters. After parametric design, simulation is performed. Virtual needles are generated at selected acupoints along a set angle. Based on finite element analysis, the soft tissue is divided into thousands of tiny mesh units, and each unit is assigned realistic material properties. The virtual needles are driven to move along the path by the physics engine, and the tissue deformation and needle stress under different manipulations are observed. By simulating the up-and-down reciprocating motion of the virtual needles, the dynamic changes in tissue resistance during each insertion and withdrawal are calculated. By simulating the rotation of the virtual needles, the torque required for rotation is calculated based on the friction model inside the soft tissue.

6. The acupuncture robot meridian and acupoint positioning system based on digital twins according to claim 5, characterized in that, After selecting the acupoint, perform parametric design by doing the following: Obtain basic attribute information of patients and information obtained from doctors' medical history inquiries about patients; Based on the knowledge graph of the selected acupoints, the associated graph items are determined according to the basic attribute information and medical history inquiry information; Determine the first free display area of ​​the parametric design interface; Determine whether a second free display area exists within the first free display area; the second free display area is a continuous area that can just accommodate all associated atlas items; If it does not exist, determine a locally contiguous third free display area within the first free display area; Get the target distance between the third free display area and the input box; Traverse the third free display area in order of target distance from nearest to farthest, and use the third free display area being traversed as the fourth free display area; Obtain the information content of local subnetworks of associated graph items; Based on the order of the network hierarchy of the local subnetwork in the knowledge graph from largest to smallest, the area of ​​the fourth free display region and the information content of the local subnetwork are matched to determine the target local subnetwork with the maximum information content that does not exceed the doctor's cognitive load. Map the target local subnet onto the fourth free display area, and display the remaining subnets in the local subnets other than the target local subnet around the fourth free display area with a preset first transparency; Stop traversing once all local subnetworks have been mapped; If the local subnet is not fully mapped after all the third free display areas have been traversed, then the display corner area closest to the last traversed third free display area is determined. The unmapped local subnetworks are displayed in the corner area of ​​the display based on a preset second transparency, wherein the first transparency is less than the second transparency; Once all mappings or displays are complete, they assist doctors in parametric design.

7. The acupuncture robot meridian and acupoint positioning system based on digital twins according to claim 6, characterized in that, To assess the safety of the pathway and predict tissue deformation and stress, perform the following actions: During collision detection, the path and depth of the virtual needle tip are continuously monitored to determine whether the needle tip will collide with the bones or vital organs embedded in the patient's digital twin model. If a collision risk is detected, the path will be highlighted in red and a warning will be issued. During force analysis, the force on the virtual needle is monitored in real time, and the relationship curve between force and depth is plotted. When the force value suddenly increases sharply at a certain depth, a warning is issued, indicating that fascia or bone has been encountered. When the calculated torque exceeds the preset safety threshold, a warning is issued, indicating that the technique is too heavy, resulting in excessive tissue entanglement or needle breakage.

8. The acupuncture robot meridian and acupoint positioning system based on digital twins according to claim 7, characterized in that, The acupuncture robot is controlled to move above the target acupoint, tracking the patient's minute movements in real time and simultaneously updating the patient's digital twin model and the target position of the acupuncture robot to achieve dynamic tracking and precise positioning of meridian acupoints. The following operations are performed: The patient lies down in the treatment position, the visual sensor scans the patient and generates a high-precision, static reference point cloud, and based on this, an initial digital twin is created, and the digital twin, acupuncture robot and patient are initially aligned in the same world coordinate system. A high-speed, low-latency RGB-D camera continuously acquires real-time point cloud data of the patient's body surface at a rate of 30 frames per second. The real-time point cloud of the current frame is matched with the initial reference point cloud to calculate the optimal rigid body transformation matrix, which describes the rotational and translational movements that the patient needs to perform in order to return to the initial position. The calculated transformation matrix is ​​directly applied to the digital twin, so that no matter how the patient moves, the digital twin always maintains synchronization with the patient in position and posture, reflecting the patient's real-time status. Based on the real-time coordinates of acupoints on the current digital twin, the system continuously receives updated target points and calculates the joint movement commands of the acupuncture robot in real time based on the new target points. The system then drives the needles to continuously and stably track the moving target acupoints, thereby achieving precise positioning of the patient's meridian acupoints.

9. The acupuncture robot meridian and acupoint positioning system based on digital twins according to claim 8, characterized in that, The acupuncture robot provides real-time feedback on needle sensation data during acupuncture and compares it with the predicted values ​​from the patient's digital twin model. Based on the comparison, it automatically adjusts the acupuncture treatment plan to ensure patient safety and performs the following operations: After the patient's meridian acupoints are accurately located, the acupuncture robot begins to insert needles. Force sensors collect the actual torque data of the needles in real time. At the same time, the real-time physics engine of the patient's digital twin model calculates and predicts the force data based on the current acupuncture depth, speed, and tissue mechanical model. The actual torque data is compared with the predicted force data to determine the deviation; If the deviation is within the preset safety threshold, the acupuncture process conforms to the model prediction, the tissue characteristics are normal, and the acupuncture robot continues to operate. If the deviation exceeds the safety threshold, the safety mechanism is immediately triggered, sending an emergency stop command to the acupuncture robot, freezing all movement instantly.

Images