Dragon tank robot system based on intelligent treatment and recuperation and control method

Abstract

The invention discloses a pitaya robot system based on intelligent treatment and recuperation and a control method, and the system comprises a high-precision image collection and human body image collection and three-dimensional reconstruction module which is used for precisely recognizing and positioning a treatment region and acupoints; the core physician manipulation digitization module is used for finely capturing professional treatment manipulations and standardized recuperating manipulations of deep physicians for operating pitayas through a multi-dimensional sensor array and forming a manipulation model library; the intelligent strategy generation module based on artificial intelligence can call, fuse or optimize and generate an optimal execution scheme from a corresponding model library according to a mode selected by a user; and the high-precision and compliant force control robot execution module reproduces the digitized technique of the pitaya in a high-fidelity manner under the real-time closed-loop control of visual guidance and force sense feedback. Through the double-track parallel strategy, quantification, inheritance and popularization of the high-level pitaya therapy are achieved, and application of the high-level pitaya therapy is pushed to the wide field.

Description

Technical Field

[0001] This invention relates to the field of big health industry technology, specifically to a fire dragon cupping robot system and control method based on intelligent treatment and recuperation. Background Technology

[0002] Fire Dragon Cupping Therapy, a highly distinctive physical therapy in traditional Chinese medicine, integrates moxibustion, scraping, massage, and cupping. It is highly regarded for its significant effects in dispelling dampness and cold, promoting blood circulation, and warming the meridians. However, its core efficacy largely depends on the practitioner's skill, experience, and ineffable "feel." This highly experience-dependent model, in the context of modern medicine and the development of the health industry, exposes several inherent challenges: 1. Lack of Standardization and Difficulty in Inheritance: The techniques of top physicians in operating the Fire Dragon Cupping Ceremony involve extremely precise control of pressure, movement speed, rotation angle, and temperature at the rim of the cup. These techniques are difficult to describe accurately through language or writing, resulting in low efficiency in teaching and inheritance, and putting high-level skills at risk of being lost.

[0003] 2. Inconsistent service quality: Due to the lack of objective evaluation standards, the skill levels of practitioners in the market vary, resulting in huge differences in service effects. There is even a safety risk of skin burns or injuries caused by improper control of force or temperature.

[0004] 3. High workload for physicians and scarcity of resources: Fire Dragon Cupping therapy is a high-intensity physical labor, and the training period for senior physicians is long, resulting in limited service capacity and difficulty in meeting the growing market demand, which has led to a serious shortage of high-quality medical resources.

[0005] At the same time, with the rapid development of the social economy and the awakening of public health awareness, the concept of "big health" has taken root in people's hearts. Public demand has expanded from simply "treating existing illnesses" to "preventing illnesses." In these broad scenarios, the market urgently needs safe, comfortable, standardized, convenient, and highly accessible professional health care services.

[0006] To address these challenges, existing technologies have explored various approaches, but most focus on simple massage or cupping devices, failing to provide effective solutions for the complex and comprehensive characteristics of fire cupping therapy. These devices cannot simulate the dynamic process of force, heat, and motion coupling during fire cupping, nor can they reproduce the exquisite skills of experts.

[0007] In conclusion, existing technologies fail to provide a comprehensive solution that can systematically and clearly distinguish between professional cupping therapy "treatment techniques" and standardized "therapeutic techniques," enabling in-depth digital data acquisition, intelligent modeling, and high-fidelity reproduction. Therefore, the industry urgently needs a new technological paradigm to drive the modernization and popularization of cupping therapy. Summary of the Invention

[0008] The technical problem to be solved by this invention is to provide a fire cupping robot system and control method based on intelligent treatment and recuperation, which can learn, quantify, store and reproduce the exquisite techniques of fire cupping therapy experts with high fidelity, and innovatively combines professional "treatment" and public "recuperation" functions to improve the user experience.

[0009] To address the aforementioned technical problems, this invention provides a fire dragon cupping robot system based on intelligent treatment and recuperation, comprising: The image acquisition and 3D reconstruction module is used to acquire high-definition image information of the user's body parts to be processed, and generate a 3D point cloud model containing surface normal vector information based on the image information. The machine vision algorithm is used to identify and locate the target acupoints or physiotherapy areas on the 3D point cloud model. The digital module for physician's fire dragon cupping technique is used to collect and process multi-dimensional dynamic technique data generated when physicians manually operate fire dragon cupping, and to transform the multi-dimensional dynamic technique data into a structured digital technique model. The manipulation model library is used to store digital manipulation models, which include a therapeutic manipulation model library and a recuperative manipulation model library. The intelligent strategy generation module is connected to the technique model library and is used to select or fuse one or more digital technique models from the corresponding technique model library according to the "treatment mode" or "recuperation mode" selected by the user, so as to generate a comprehensive execution plan containing robot motion trajectory, posture sequence and force control parameters. The robot execution module includes a multi-joint robotic arm, a six-axis force sensor, and an end effector for gripping or integrating the fire dragon cupping device, connected in sequence, for performing dynamic actions on the target acupoint or physiotherapy area according to the comprehensive execution scheme; The main control and feedback system, serving as the central neural network of the system, connects the image acquisition and 3D reconstruction module, the physician's fire dragon cupping technique digitization module, the technique model library, the intelligent strategy generation module, and the robot execution module. Based on real-time feedback data from the six-axis force sensor and the image acquisition and 3D reconstruction module, it performs microsecond-level closed-loop adjustments to the motion of the robot execution module.

[0010] Furthermore, the digital module for the physician's fire dragon cupping technique includes: The technique acquisition terminal is used to simulate the fire dragon pot, and it integrates a six-axis force sensor and a nine-axis inertial measurement unit (IMU). A high frame rate spatial positioning unit is used to track the real-time position and attitude of the method acquisition terminal in three-dimensional space with sub-millimeter accuracy; The data synchronization processing unit is used to synchronously acquire and record multidimensional dynamic data streams from a six-axis force sensor and a nine-axis inertial measurement unit (IMU), forming a multidimensional dataset of "force-position-attitude-time" with a unified time reference.

[0011] Furthermore, the data synchronization processing unit performs in-depth processing on the multidimensional dataset, including data filtering, noise removal, and action primitive segmentation based on events or time series. It uses Hidden Markov Model (HMM) or Recurrent Neural Network (RNN) algorithms to decompose the continuous Fire Dragon Can operation into several standardized action primitives.

[0012] Furthermore, the structured digital manipulation model includes a sequence of target acupoints, the type of action primitive corresponding to each target acupoint, the duration of each action primitive, the three-dimensional spatial trajectory of each action primitive, the end effector posture change curve of each action primitive, and the six-axis force change curve profile of each action primitive.

[0013] Furthermore, the intelligent strategy generation module employs reinforcement learning or generative adversarial network (GAN) machine learning algorithms to continuously optimize the models in the method model library or interpolate and fuse multiple basic models based on historical data and effect feedback, in order to generate new execution schemes that are more suitable for specific scenarios.

[0014] Furthermore, the main control and feedback system adopts a model-based variable parameter admittance controller or impedance controller. By comparing the actual contact force measured by the six-axis force sensor with the target contact force defined in the digital manual model in real time, the position and posture of the robotic arm are dynamically adjusted to achieve compliant and high-precision force tracking control.

[0015] A method for controlling a fire dragon cupping robot based on intelligent treatment and recuperation, employing any one of the systems described above, includes the following steps: S1: Technique Digitization and Model Library Construction Phase (Offline): a. A physician demonstrates the operation using a manual data acquisition terminal that integrates a six-axis force sensor and a nine-axis inertial measurement unit (IMU); b. Real-time and synchronous acquisition of multi-dimensional dynamic data streams of force, position, and attitude during the exemplary operation process; c. Process, segment, and structure the acquired multidimensional dynamic data stream, and transform it into one or more digital manipulation models; d. Based on the purpose of the operation, classify and store the generated digital manipulation models into the treatment manipulation model library or the recuperation manipulation model library; S2: Automated Execution and Real-Time Feedback Phase (Online): e. The user selects either "treatment mode" or "recuperation mode"; f. Use the image acquisition function in the image acquisition and 3D reconstruction module to scan the body area of ​​the user to be processed, and perform 3D reconstruction and automatic identification and positioning of target acupoints; g. Based on the user's selection and specific needs, automatically or manually select one or more suitable digital manipulation models from the corresponding manipulation model library to generate a comprehensive execution plan that the robot can execute; h. Control the robot execution module to move to the identified target acupoint according to the comprehensive execution scheme, and reproduce the action sequence defined by the selected digital manipulation model; i. During execution, the robot's motion and force application are adjusted in a closed loop through real-time feedback provided by a six-axis force sensor and an image acquisition and 3D reconstruction module.

[0016] Furthermore, in step c), the process of transforming the multidimensional dynamic data stream into a digital manipulation model includes: using time series analysis, Hidden Markov Model (HMM) or Long Short-Term Memory (LSTM) network algorithms to perform pattern recognition on the data stream and automatically segment and label it into different action primitives.

[0017] Furthermore, in step i), the closed-loop adjustment also includes: when the user's movement is detected, updating the coordinates of all target acupoints in real time, and dynamically replanning the robot's motion trajectory to adapt to changes in the user's body position.

[0018] The beneficial effects of this invention are: 1. Dual-track approach with broad market coverage: By distinguishing between treatment and recuperation, this invention can not only serve patients who need professional fire dragon cupping therapy, but also meet the daily health care needs of a large number of sub-healthy people.

[0019] 2. Balancing professionalism and universality: The treatment module solves the problem of passing on the skills of Fire Dragon Cupping experts. The recuperation module ensures the standardization, safety, and ease of use of the service.

[0020] 3. It has realized the scientification and inheritance of the core techniques of Fire Dragon Cupping: the techniques of traditional Chinese medicine practitioners in operating Fire Dragon Cupping have been transformed into precise data models, opening up a new path for the modernization of this traditional therapy.

[0021] 4. Ensures standardized and consistent operation: The robot strictly follows the digital model, eliminating the arbitrariness of human operation.

[0022] 5. Improved process accuracy and safety: Combining 3D vision and force feedback enables precise and safe physical operation, avoiding the risk of uncontrolled force or temperature that may occur during human operation.

[0023] 6. It liberated the productivity of doctors and technicians: freeing professionals from repetitive and heavy physical labor. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the overall architecture of the present invention; Figure 2 This is a block diagram of the digital module for the physician's fire dragon cupping technique of the present invention; Figure 3 This is a flowchart of the control method of the present invention; Figure 4 This is a three-dimensional spatial trajectory change curve of the action primitive of the present invention in a time series; Figure 5 This is a graph showing the force variation of the action element of the present invention over a time series; Figure 6 This is a graph showing the change in torque of the action element of the present invention over a time series. Detailed Implementation

[0025] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, so that those skilled in the art can better understand and implement the present invention. However, the embodiments described are not intended to limit the present invention.

[0026] Reference Figure 1 As shown, an embodiment of the intelligent treatment and recuperation-based fire dragon cupping robot system and control method of the present invention is capable of in-depth digital acquisition, intelligent modeling, classification and storage, and high-fidelity reproduction of professional fire dragon cupping techniques. To achieve this goal, the present invention has made fundamental innovations in system design, clearly dividing the system functions into two parallel and complementary core modules: a treatment module and a recuperation module.

[0027] Treatment Module (Expert Replication): This module aims to precisely replicate, through cutting-edge technology, the complex, sophisticated, and effective treatment techniques employed by experienced physicians using cupping therapy for specific ailments (such as cervical spondylosis, lumbar muscle strain, and rheumatic pain). Its core concept is "replicating the expert," pursuing depth, precision, and repeatability in treatment.

[0028] Standardized Wellness Module: This module aims to provide a series of standardized rehabilitation and wellness services centered around cupping therapy (such as relieving muscle fatigue, regulating uterine coldness, and removing dampness from the whole body). Its core concept lies in "standardized service," pursuing the ultimate comfort of the user experience, the accessibility of services, and the wide applicability.

[0029] To achieve the above functional division, the present invention adopts the following interconnected technical solution, which mainly includes: an image acquisition and three-dimensional reconstruction module, a digital module for physician's fire dragon cupping techniques, a technique model library (containing two sub-libraries: treatment and recuperation), an intelligent strategy generation module, a high-precision compliant force control robot execution module, and a main control and feedback system.

[0030] The dual-track fire dragon pot technique model library: This is the cornerstone of functional division. It is a strictly partitioned central database containing: Treatment Technique Model Library: This library stores "prescription-level" cupping techniques captured from the clinical practice of multiple senior physicians. Each model corresponds to a specific ailment and embodies the culmination of expert experience.

[0031] Therapeutic Technique Model Library: Stores a set of carefully designed and optimized "health product-grade" standardized fire dragon cupping techniques, with a greater emphasis on universality, safety, and a comfortable experience.

[0032] Pattern-driven intelligent strategy generation module: This module acts as the "decision brain" of the system. Based on the "treatment mode" or "recuperation mode" selected by the user, it retrieves, matches and calls models from the corresponding method model library to ensure the correctness of the execution strategy.

[0033] Modular, multi-functional robotic actuator module: The end effector of this module is designed with a quick-change structure. In "Therapy Mode," a professional fire cupping therapy head can be installed. In "Recuperation Mode," a fire cupping massage head with additional comfort features can be installed.

[0034] Specifically, refer to Figure 1 As shown, in one embodiment of the present invention, The hardware configuration of the Intelligent Fire Dragon Cupping Robot System 100 is basically the same as described above. Its revolutionary nature lies in the deep integration and innovation of software, data and workflow, especially in the precise acupoint positioning of the Fire Dragon Cupping therapy, the compliant control of the robotic arm posture, and the synergy of multiple physical fields (force, heat).

[0035] I. High-precision acupoint positioning and three-dimensional perception, corresponding to the deepening of the image acquisition and three-dimensional reconstruction module 110.

[0036] Precise acupoint location is a prerequisite for effective treatment. This invention employs a strategy based on "standard model registration + multimodal feature fusion" to achieve individualized, millimeter-level acupoint location.

[0037] Personalized 3D model generation: Once the user lies on the treatment bed, the depth camera in the image acquisition and 3D reconstruction module 110 is activated to quickly scan the user's body area to be treated (such as the back and neck). Within seconds, the system generates a high-density 3D point cloud model containing millions of points, fully reproducing the precise surface morphology of the user's current body part.

[0038] Standard Human Anatomy Model and Acupoint Atlas: The system includes a standard, high-precision digital human anatomical model. This model not only includes the body surface morphology but also pre-annotates the three-dimensional coordinates of all meridians and acupoints defined in the national standard. This serves as a universal "template."

[0039] Coarse and fine registration: By identifying obvious anatomical features in the user's 3D model (such as scapular contours, spinal grooves, and cervical curves) and matching them with corresponding features on the standard model, a rough alignment is achieved. Subsequently, the Iterative Closest Point (ICP) algorithm and its variants are used to precisely fit the surface of the standard model with the user's real-time generated point cloud model, maximizing the overlap between the two models. After this process, the pre-annotated acupoints on the standard model are mapped onto the user's personalized 3D body model, completing the initial acupoint localization.

[0040] Local feature correction: Taking into account individual differences (such as muscle development and fat thickness), the system also utilizes higher-resolution 2D images or thermal images (optional) as supplementary information. For example, by using a deep learning network (such as U-Net) to identify the edges and direction of muscles, the mapped acupoint locations are fine-tuned to better match the user's actual physiological state.

[0041] Normal vector calculation: After localization, the system calculates the surface normal vector at the location of each target acupoint. This vector is perpendicular to the skin surface at that point, providing a crucial attitude reference for the subsequent robotic actuator to perform the cupping therapy at the correct angle.

[0042] II. Improvement of the robotic arm’s compliant posture and force-thermal coordinated control, corresponding to the robot execution module 150 and the main control and feedback system 160.

[0043] To faithfully reproduce the expert's exquisite technique in operating the fire dragon pot, the robot must not only be "accurate in its arrival," but also "move gently," "control precisely," and "heat accurately."

[0044] Kinematics and Posture Planning: Inverse kinematics solution: Once the three-dimensional coordinates of the target acupoint and the desired contact posture (determined by the normal vector) are determined, the main control and feedback system 160 solves the inverse kinematics for the seven-DOF robotic arm, calculating the angles of each joint. The redundancy of the seven-DOF allows the robotic arm to not only meet the end-effector pose requirements but also actively "avoid obstacles" or optimize its own configuration to achieve the best operating posture.

[0045] Dynamic posture control: For complex techniques such as "cupping," the digital technique model records the posture change curve over time (such as periodic rotation around the normal vector). The motion planner translates these instructions into smooth joint angular velocity commands, driving the robotic arm's "wrist" joint to perform delicate movements.

[0046] Hybrid position / force control strategy: Approach phase (position control): During the process of the robotic arm's end effector moving from the air to approach the skin surface, a high-speed, high-precision position control mode is adopted.

[0047] Contact and execution phase (impedance / admittance control): The control system seamlessly switches to impedance control mode the instant the end effector is about to contact the skin (either by a force sensor sensing an extremely small force or by a vision system determining that the distance is extremely close).

[0048] Core: Applications of impedance control Simulating "Hand Feel": In impedance control mode, the robot is no longer a rigid positioning device, but rather a programmable, virtual "mass-spring-damping" system. This means that when it comes into contact with the human body, it exhibits a compliantness similar to that of a human hand. If the user moves slightly, the robot will not resist abruptly, but will "follow" the movement to produce a small displacement while maintaining the set contact force, greatly improving safety and comfort.

[0049] Precise force tracking: The digital model defines the ideal magnitude and direction of the contact force (Fx, Fy, Fz) at every moment. The controller reads the feedback values ​​from the six-axis force sensor in real time and compares them with the target values ​​in the model. The error between the two drives the impedance controller to adjust the tiny displacements of the robotic arm in real time, thereby accurately tracking the target force curve. This is the key to reproducing the "force" of the expert's Fire Dragon Can.

[0050] Temperature closed-loop control: The robot's end effector can also integrate temperature sensors (such as infrared thermocouples) and a controllable heating module. During execution, the system controls not only the force but also the temperature at the can opening. The digital model also includes temperature change curves. The main control system adjusts the power of the heating module using PID or more advanced control algorithms, ensuring the can opening temperature precisely follows the curve set in the model, achieving gentle, continuous, and safe heat penetration.

[0051] Through the above-mentioned in-depth design of acupoint positioning, posture control, and temperature control, this invention constructs a complete "perception-decision-execution" technical closed loop, ensuring that the robot can not only find the correct position, but also, in a highly human-like, safe, and precise manner, coordinate the control of the two physical fields of force and heat to reproduce the expert's fire cupping therapy and recuperation techniques, thereby truly realizing the core value of this invention.

[0052] III. Implementation of Technique Model Library 140 At the database level, two core data tables with different structures but interrelated relationships are created: one is treatment_models, and the other is wellness_models.

[0053] The `treatment_models` table stores digital manipulation models collected from multiple rigorously certified TCM practitioners, orthopedic surgeons, or rehabilitation therapists for specific conditions such as "cervical spondylosis segmental repositioning" and "deep muscle release for lumbar muscle strain." Fields in this table will include the condition code (ICD code), treatment stage, contraindications, and extremely detailed force-position-posture-time-temperature data.

[0054] The `wellness_models` table stores standardized manipulative models designed by rehabilitation therapists, sports scientists, and ergonomics experts for general purposes such as "deep neck and shoulder relaxation," "promoting whole-body circulation," and "post-exercise recovery." The fields in this table focus more on comfort ratings, target audience labels (such as "office workers," "fitness enthusiasts"), and safety threshold parameters.

[0055] IV. Interaction process of intelligent strategy generation module 130.

[0056] It can use a graphical user interface (GUI) to provide two visually distinct and logically independent entry buttons on the main interface: "Professional Treatment Mode" and "Health and Wellness Mode".

[0057] When "Professional Treatment Mode" is clicked, the system will enter a rigorous process similar to a Medical Information System (HIS). The interface will guide the operator (who must be an authorized physician or therapist) to enter the patient ID, retrieve medical records, or enter detailed diagnostic information. Subsequently, based on the diagnosis, the system will intelligently recommend the most suitable manipulation model from the treatment_models table and display it to the operator in a 3D visualization for final confirmation or fine-tuning.

[0058] When the "Health and Wellness Mode" is selected, the interface switches to a user-friendly and intuitive consumer-grade application style. Users can directly select body parts on the touchscreen and then browse a series of illustrated wellness programs (such as "60-Minute Deep Relaxation Journey" and "30-Minute Vitality Recovery"). The system reads and executes the corresponding standardized procedures from the wellness_models table.

[0059] V. The control method flow of the present invention is as follows: Figure 3 The above includes the following steps: Phase 1, S100 - Digitalization of techniques for offline learning and modeling.

[0060] S110-S130: Demonstration and Data Acquisition: This stage is the same as the aforementioned embodiments, but emphasizes the professionalism of the acquisition environment and the fidelity of the data. Using the manual acquisition terminal 121, digital motion primitive models are accurately acquired and generated on human models or volunteers equipped with markers.

[0061] Specifically, the process of transforming data from the coordinate system of the technique acquisition terminal to the world coordinate system through the spatial positioning unit 122, and then processing it through the data synchronization processing unit 123 before finally mapping it to the robot base coordinate system and the target human body coordinate system is the spatial transformation basis for achieving accurate reproduction of techniques.

[0062] The coordinate system T of the aforementioned data acquisition terminal is a local coordinate system with the sensor center as the origin, and it is the direct source of all physical interaction data. It acquires raw six-axis force / torque and IMU attitude data, used to capture the most direct and raw physical information during the physician's operation.

[0063] The world coordinate system W is a fixed global reference coordinate system that serves as a "world anchor point" connecting all other dynamic coordinate systems. It can be a fixed point in physical space (such as a device base). The transformation from T to W is obtained through a spatial positioning unit (such as an optical camera).

[0064] The robot's base coordinate system R is a coordinate system with the robot's base as its origin. It serves as the robot's own reference system for executing all motion planning. It is used to define the robot's workspace and calculate all joint movements. A fixed transformation from W to R is obtained through a one-time "hand-eye calibration".

[0065] The target human coordinate system B is a dynamic coordinate system established on the user's body, used for precise acupoint location to adapt to even the slightest movements of the user. The treatment target is defined within this user-bound coordinate system to achieve adaptive tracking. It calculates the dynamic transformation from W to B in real time using visual positioning (3D camera).

[0066] S140: Deep Modeling and Classification Storage: When storing models in the database, perform key classification and annotation steps. Add a core field "Model Type". If the model collected is Dr. X's "Du Meridian Cupping Therapy", this field is marked as "Treatment" and associated with tags such as "Du Meridian" and "Dampness Removal", and stored in the treatment_models table. If the model collected is a standardized "Shoulder and Neck Relief" procedure, it is marked as "Wellness" and associated with tags such as "Relaxation" and "Sleep Aid", and stored in the wellness_models table.

[0067] Phase 2, S200 - Automated execution for online reproduction and interaction.

[0068] S205: Mode Selection: Users or operators first select "Therapy Mode" or "Convalescence Mode" on the system GUI.

[0069] S210-S220: Personalized 3D Perception: The system activates the image acquisition and 3D reconstruction module 110 to perform high-precision 3D reconstruction and acupoint positioning.

[0070] S230: Intelligent Solution Generation: The intelligent strategy generation module 130 retrieves and calls the technique model from the corresponding database table according to the selection in S205.

[0071] S240-S290: Safe and Compliant Execution and Monitoring: Subsequent robot program generation, compliant force control-based execution, and multi-sensor fusion-based safety monitoring are consistent with the aforementioned embodiments, but the execution employs sophisticated techniques with distinctly different characteristics (professionalism or comfort) selected according to different modes. Throughout the process, the user always holds an emergency stop button to ensure absolute safety.

[0072] Specifically, refer to Figures 4 to 6 The diagram shown is a digital model of a typical "pushing the can" motion element in a fire dragon can. The diagram uses multi-dimensional curves to illustrate the changes in the three-dimensional spatial trajectory (x, y, z) of this motion over time, as well as the dynamic profiles of the forces and torques (Fx, Fy, Fz, Tx, Ty, Tz) in the six degrees of freedom, intuitively presenting the essence of digital techniques.

[0073] The following is a detailed explanation of a typical "pushing" action of a fire dragon cupping device, demonstrating how the system of this invention can realize the entire process from offline data collection to online reproduction, and provides specific data as an illustration.

[0074] Phase 1: Offline digitization and model building.

[0075] a) Demonstration Operation and Data Collection A senior TCM doctor used the technique collection terminal of this invention to perform a standard fire dragon cupping "pushing cupping" operation on the Du meridian (e.g., from Dazhui acupoint to Mingmen acupoint) on the back of a volunteer.

[0076] Equipment: Manual acquisition terminal (integrating a six-axis force sensor and IMU), high frame rate optical spatial positioning unit.

[0077] Data Acquisition: The system synchronously acquires the following multidimensional data at a frequency of 1000Hz: Position and attitude: The real-time position P(t)=[x,y,z] and attitude Q(t)=[qx,qy,qz,qw] of the terminal in the world coordinate system W are collected by the optical spatial positioning unit tracking method.

[0078] Force and torque: The contact force F(t) = [Fx, Fy, Fz] and torque T(t) = [Tx,Ty,Tz] are obtained through a six-axis force sensor.

[0079] Angular velocity and angular acceleration: The terminal's own angular velocity and angular acceleration are obtained through the IMU unit.

[0080] b) Data processing and action primitive segmentation The system processes the collected data stream and uses time series analysis algorithms to automatically divide the continuous operation of "pushing the can" into a series of discrete action primitives. For example, a typical "pushing the can" process is divided into: Element 1: Contact and Pressure Application (0s-2s) Element 2: Uniform movement (2s-15s) Element 3: Deceleration and Exit (15s-17s) c) Deep modeling and classification for database entry The system performs in-depth modeling of the action primitive "uniform speed movement," forming a structured digital technique model. This model includes: Basic type: Push can.

[0081] Duration: 13 seconds.

[0082] Three-dimensional spatial trajectory: from the three-dimensional coordinates P_start=[0.1, 0.2, 0.5] (unit: meters) near the Dazhui acupoint to P_end=[0.1, -0.1, 0.45] near the Mingmen acupoint. Its trajectory is modeled as a smooth cubic spline curve.

[0083] End effector attitude change curve: The can opening is always perpendicular to the skin surface, and its attitude quaternion Q(t) is planned as a smooth interpolation sequence to ensure that it is always aligned with the skin normal vector.

[0084] Six-axis force variation curve profile: Normal force Fz: linearly increases from 20N to 35N, holds for 10 seconds, and then linearly decreases to 0N.

[0085] Tangential force Fx / Fy: During the displacement process, there is a tangential force pointing in the direction of P_end, with a constant magnitude of 5N.

[0086] Classification: This model is labeled as "Treatment" and stored in the manipulation model library.

[0087] Phase Two: Online Automated Execution and Real-Time Feedback.

[0088] A) Personalized 3D perception and acupoint localization When a user needs to receive this "cupping" therapy, the system first activates the image acquisition and 3D reconstruction module.

[0089] 3D Reconstruction: A depth camera scans the user's back to generate a high-density point cloud model.

[0090] Standard registration: The system performs ICP registration between the built-in standard human body model and the user point cloud, maps the coordinates of the pre-labeled "Dazhui" and "Mingmen" acupoints onto the user's body, and calculates the surface normal vector of each acupoint, which is used as a reference for the posture adjustment of the multi-joint robotic arm.

[0091] Location results: Dazhui acupoint coordinates [0.105, 0.21, 0.52], Mingmen acupoint coordinates [0.11, -0.09, 0.46].

[0092] B) Intelligent solution generation and robot execution: Solution Generation: The intelligent strategy generation module retrieves the "cupping" model from the technique model library and dynamically generates a comprehensive solution that the robot can execute based on the actual acupoint coordinates on the user's body. This solution includes a new motion trajectory and corrected force control parameters.

[0093] The robot's execution includes the following: Position control: The robotic arm first moves to a position about 5cm above the Dazhui acupoint using a high-speed position control mode.

[0094] Switch to impedance control: When the end force sensor senses a small contact force (e.g., > 0.5N), the system smoothly switches to impedance control mode.

[0095] Force and trajectory tracking: The robot begins to precisely execute the "pushing the can" action. It reads feedback from the six-axis force sensor in real time and compares the actual normal force (Fz) with the target force curve. If the actual force is less than the target force, the controller will slightly drive the robotic arm to move downwards; conversely, it will slightly move upwards, ensuring that the normal force accurately tracks the change curve from 20N to 35N. At the same time, it moves along the planned trajectory from the Dazhui acupoint to the Mingmen acupoint, ensuring that the tangential force remains at around 5N.

[0096] Real-time feedback and adjustment: If the user's body position changes due to breathing or slight movement during the process, the image acquisition module will update the acupoint coordinates in real time, and the main control system will dynamically replan the trajectory to ensure that the "cupping" action is always performed in the correct acupoint area and the force remains constant.

[0097] C) Result: Through the above process, the robot not only accurately moved the cupping jars to the designated area, but also reproduced the expert-level "pushing" force and trajectory in a programmable, compliant and high-fidelity manner, thereby ensuring the standardization and safety of the treatment effect.

[0098] The embodiments described above are merely preferred embodiments for fully illustrating the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention.

Claims

1. A robotic system for intelligent treatment and recuperation using a traditional Chinese medicine cupping device, characterized in that: include: The image acquisition and 3D reconstruction module is used to acquire high-definition image information of the user's body parts to be processed, and generate a 3D point cloud model containing surface normal vector information based on the image information. The machine vision algorithm is used to identify and locate the target acupoints or physiotherapy areas on the 3D point cloud model. The digital module for physician's fire dragon cupping technique is used to collect and process multi-dimensional dynamic technique data generated when physicians manually operate fire dragon cupping, and to transform the multi-dimensional dynamic technique data into a structured digital technique model. The manipulation model library is used to store digital manipulation models, which include a therapeutic manipulation model library and a recuperative manipulation model library. The intelligent strategy generation module is connected to the technique model library and is used to select or fuse one or more digital technique models from the corresponding technique model library according to the "treatment mode" or "recuperation mode" selected by the user, so as to generate a comprehensive execution plan containing robot motion trajectory, posture sequence and force control parameters. The robot execution module includes a multi-joint robotic arm, a six-axis force sensor, and an end effector for gripping or integrating the fire dragon cupping device, connected in sequence, for performing dynamic actions on the target acupoint or physiotherapy area according to the comprehensive execution scheme; The main control and feedback system, serving as the central neural network of the system, connects the image acquisition and 3D reconstruction module, the physician's fire dragon cupping technique digitization module, the technique model library, the intelligent strategy generation module, and the robot execution module. Based on real-time feedback data from the six-axis force sensor and the image acquisition and 3D reconstruction module, it performs microsecond-level closed-loop adjustments to the motion of the robot execution module.

2. The Fire Dragon Cupping Robot System based on intelligent treatment and recuperation as described in claim 1, characterized in that, The digital module for the physician's cupping technique includes: The technique acquisition terminal is used to simulate the fire dragon pot, and it integrates a six-axis force sensor and a nine-axis inertial measurement unit (IMU). A high frame rate spatial positioning unit is used to track the real-time position and attitude of the method acquisition terminal in three-dimensional space with sub-millimeter accuracy; The data synchronization processing unit is used to synchronously acquire and record multidimensional dynamic data streams from a six-axis force sensor and a nine-axis inertial measurement unit (IMU), forming a multidimensional dataset of "force-position-attitude-time" with a unified time reference.

3. The Fire Dragon Cupping Robot System based on intelligent treatment and recuperation as described in claim 2, characterized in that, The data synchronization processing unit performs in-depth processing on the multidimensional dataset, including data filtering, noise removal, and action primitive segmentation based on events or time series. It uses Hidden Markov Model (HMM) or Recurrent Neural Network (RNN) algorithms to decompose continuous Fire Dragon Can operations into several standardized action primitives.

4. The Fire Dragon Cupping Robot System based on intelligent treatment and recuperation as described in claim 3, characterized in that, The structured digital manipulation model includes a sequence of target acupoints, the type of action primitive corresponding to each target acupoint, the duration of each action primitive, the three-dimensional spatial trajectory of each action primitive, the end effector posture change curve of each action primitive, and the six-axis force change curve profile of each action primitive.

5. The Fire Dragon Cupping Robot System based on intelligent treatment and recuperation as described in claim 1, characterized in that, The intelligent strategy generation module employs reinforcement learning or generative adversarial network (GAN) machine learning algorithms. Based on historical data and performance feedback, it continuously optimizes the models in the method model library or interpolates and fuses multiple basic models to generate new execution schemes that are more suitable for specific scenarios.

6. The Fire Dragon Cupping Robot System based on intelligent treatment and recuperation as described in claim 1, characterized in that, The main control and feedback system adopts a model-based variable parameter admittance controller or impedance controller. By comparing the actual contact force measured by the six-axis force sensor with the target contact force defined in the digital manual model in real time, the position and posture of the robotic arm are dynamically adjusted to achieve compliant and high-precision force tracking control.

7. A method for controlling a fire dragon cupping robot based on intelligent treatment and recuperation, characterized in that, The system according to any one of claims 1-6 includes the following steps: S1: Technique digitization and model library construction phase: a. A physician demonstrates the operation using a manual data acquisition terminal that integrates a six-axis force sensor and a nine-axis inertial measurement unit (IMU); b. Real-time and synchronous acquisition of multi-dimensional dynamic data streams of force, position, and attitude during the exemplary operation process; c. Process, segment, and structure the acquired multidimensional dynamic data stream, and transform it into one or more digital manipulation models; d. Based on the purpose of the operation, classify and store the generated digital manipulation models into the treatment manipulation model library or the recuperation manipulation model library; S2: Automated Execution and Real-Time Feedback Phase (Online): e. The user selects either "treatment mode" or "recuperation mode"; f. Use the image acquisition function in the image acquisition and 3D reconstruction module to scan the body area of ​​the user to be processed, and perform 3D reconstruction and automatic identification and positioning of target acupoints; g. Based on the user's selection and specific needs, automatically or manually select one or more suitable digital manipulation models from the corresponding manipulation model library to generate a comprehensive execution plan that the robot can execute; h. Control the robot execution module to move to the identified target acupoint according to the comprehensive execution scheme, and reproduce the action sequence defined by the selected digital manipulation model; i. During execution, the robot's motion and force application are adjusted in a closed loop through real-time feedback provided by a six-axis force sensor and an image acquisition and 3D reconstruction module.

8. The control method for the Fire Dragon Jar robot based on intelligent treatment and recuperation as described in claim 1, characterized in that, In step c), the process of transforming the multidimensional dynamic data stream into a digital manipulation model includes: using time series analysis, Hidden Markov Model (HMM) or Long Short-Term Memory (LSTM) network algorithms to perform pattern recognition on the data stream and automatically segment and label it into different action primitives.

9. The control method for the Fire Dragon Jar robot based on intelligent treatment and recuperation as described in claim 1, characterized in that, In step i), the closed-loop adjustment also includes: when the user moves, updating the coordinates of all target acupoints in real time and dynamically replanning the robot's motion trajectory to adapt to changes in the user's body position.

Images