Humanoid robot hand system and method for manipulative therapy integrated training gloves

Abstract

The invention discloses a humanoid robot arm system and method for a manual treatment integrated training glove, and belongs to the technical field of medical robots. The system comprises a humanoid robot hand, a training glove and intelligent terminal application software. The human-shaped robot hand is made by covering a hand skeleton plate with hydrogel to have the elastic touch feeling and appearance highly imitating the human hand, and touch force monitoring, heating and three-dimensional vibration function modules are integrated in the hand; the training glove is integrated with a finger joint bending degree and pose sensor, can be connected with or separated from the robot arm, and supports combined training and remote control training of the robot arm; and the intelligent terminal application software manages data and is connected with the cloud server. The method comprises the steps of collecting manipulation data of a trainer, driving a robot arm to reproduce manipulation, and generating expert-level treatment data by fusing a cloud expert database or large model optimization. According to the invention, digitalization of expert manipulation treatment, high-precision robot reproduction and intelligent improvement are realized, and the robot is suitable for scenes such as tumor edema, stroke rehabilitation, massage physiotherapy relaxation and the like.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of rehabilitation robots, in particular to a humanoid robot hand system and method for integrated training gloves for manual therapy, which is suitable for tumor lymphedema, stroke rehabilitation, diabetes rehabilitation, massage and physiotherapy rehabilitation and relaxation after medium and large surgery, and prevention of lymphedema. BACKGROUND

[0002] In the rehabilitation treatment of lymphedema, stroke, diabetes, prevention and treatment of lymphedema after medium and large surgery, and physiotherapy and massage relaxation, the professional manual therapy of doctors determines the curative effect, especially the treatment path and manual force control, speed and rhythm which require expert level, which highly depends on the experience of the treating doctors, which has the following limitations: 1) The treating doctors are tired due to long time and high intensity of each manual therapy, which leads to unstable manual output and cannot guarantee the quality of continuous treatment; 2) The experience of expert manual therapy is different from person to person, and the treatment consistency is poor, and there is no standardization and digital quantification; 3) The training period of novice doctors is long, and patients cannot obtain expert-level manual therapy.

[0003] The contact massage head of the treatment auxiliary device in the prior art is usually a single fixed structure, and different shapes of massage heads are often replaced by pausing treatment, which cannot simulate the manual operation required by the doctor's manual drainage treatment; the existing dexterous hand is a mechanical hard material, which can complete finger combination action, but it does not have the material elasticity of human hand and cannot meet the clinical needs of manual therapy field; the current tumor edema treatment technology lacks accurate digital collection capability for the process of doctor's expert manual therapy, which makes it difficult to learn and train neural network to drive the robot to reproduce the expert manual therapy level. SUMMARY

[0004] The present application aims to overcome the limitations of the prior art and provide a humanoid robot hand system and method for integrated training gloves for manual therapy, which realizes accurate digitalization of the process of doctor's treatment manual therapy by bionic hand material touch, performance design and finger bending angle, palm touch force digital sensing, installation at the end of the bionic mechanical arm equipped with six-dimensional force sensor, collects high-quality edema doctor and expert manual therapy data, drives the robot hand to reproduce and train the expert manual therapy large model, and realizes digital, intelligent and automatic expert-level manual therapy.

[0005] To achieve the above-mentioned application purposes, the present application provides a humanoid robot hand system and method for integrated training gloves for manual therapy, which comprises a humanoid robot hand unit, a training glove unit and an intelligent terminal application software.

[0006] When the anthropomorphic robot hand unit and the training glove unit are connected and integrated through the magnetic suction snap array, the trainer wearing the glove drives the robot hand to move and operate, and at the same time, the space is integrated and the training is carried out; when the two are separated, the trainer wearing the glove controls the robot hand remotely and synchronously, and supports the networking of the same type of training gloves to remotely control the robot hand synchronously;

[0007] The anthropomorphic robot hand unit comprises a robot hand skeleton plate and a robot hand palm;

[0008] The robot hand skeleton plate comprises a palm skeleton and a finger skeleton, is made of hard materials such as metal and plastic, has a inner side of the palm surface, integrates installation of a finger driving motor, an anthropomorphic finger mechanical component and a function module, can realize adaptive deformation actions such as grabbing, holding and pinching, and the outer side of the robot hand skeleton plate is a robot hand back surface through threaded or suitable glue installation of an array of male snaps, used for connecting an array of female snap seats of the corresponding training glove palm surface and fixedly integrated.

[0009] The inner side of the robot hand skeleton plate integrates installation of the following function modules, comprising:

[0010] 1) A robot hand thin film stress sensor array module, fixed to the palm area of the robot hand skeleton plate and the finger mechanical component finger pulp area, used for real-time detection of touch force feedback in the treatment process;

[0011] 2) A three-dimensional vibration motor module, comprising two same type micro-vibration motors, orthogonally fixed to a steel base and integrated as a whole, synthesizing a three-dimensional vibration force vector in a three-dimensional space, setting the rotation speed and phase difference parameters of the two vibration motors through the intelligent terminal application software, generating a spiral three-dimensional vibration force through the microprocessor program according to the parameters to control the PWM signal duty cycle, adjusting the combined vibration force size by adjusting the rotation speed of the two motors, generating a spiral of the resultant vector through the rotation speed difference of the two motors, changing the spatial trajectory of the resultant vector through the phase difference of the two motors, installing and fixing the module at a suitable position in the palm center area of the robot hand skeleton plate, driving the robot hand to generate a spiral three-dimensional vibration force and its energy wave, which can achieve the effect of exceeding the doctor's hand treatment of edema, the vibration frequency range is adjustable 0-300Hz, and the amplitude is adjustable 0-5mm;

[0012] 3) A heating sheet module, fixed to the surface of the robot hand skeleton plate and the finger skeleton support plate, with a temperature control heating range of 25-45℃, used for low temperature environment and auxiliary treatment;

[0013] The machine palm, comprising a palm surface and five finger pads, is made of hydrogel material and process with high imitation human hand muscle and skin material, and is fixed and packaged on the machine hand skeleton plate and its functional modules by pouring the hydrogel in the mold matched with the real hand guide mold, to make the machine palm with high imitation human hand muscle and skin elasticity and touch;

[0014] The machine palm, comprising a palm surface and five finger pads, is made of hydrogel material and process with high imitation human hand muscle and skin material, and is fixed and packaged on the machine hand skeleton plate and its functional modules by pouring the hydrogel in the mold matched with the real hand guide mold, to make the machine palm with high imitation human hand muscle and skin elasticity and touch;

[0015] The wrist part of the humanoid machine hand is configured with standard mechanical interface and power supply data interface, which is used for docking the standard interface of the end of the humanoid mechanical arm, performing data communication between the master computer of the machine hand and the mechanical arm, and realizing integrated linkage control of the mechanical arm and the machine hand.

[0016] The training glove unit is a glove component that can be worn on the hand of the trainer, comprising:

[0017] 1) Multi-modal sensor module, comprising five thin film inductive bending sensor modules integrated on the side of the five finger pads or the back of the fingers, for real-time detection and collection of the bending angles of the joints of the five fingers; array thin film touch force sensor modules integrated on the abdomen of the five fingers and the palm, for real-time detection of the touch force feedback of the trainer's hand method, which can be preferentially deployed under the array buckle base on the palm surface of the training glove; IMU pose sensor module integrated in the cross-sectional view of the back of the glove, for real-time detection and collection of the spatial pose and acceleration data of the hand;

[0018] 2) Communication control module, integrated on the back of the training glove, comprising a microprocessor, a memory, a wireless communication module, a rechargeable battery and an electrical data interface, which automatically switches to wired communication and power supply through the electrical interface when connected with the humanoid machine hand unit, and automatically switches to wireless communication and battery power supply of the training glove when separated from the humanoid machine hand unit;

[0019] 3) Human-computer interaction module, comprising magic tape straps suitable for different hand shapes and functional keys arranged on the back of the glove, the keys including system switch, start training, delete data and execute treatment function keys;

[0020] The intelligent terminal application software is used for managing and collecting storage data, communication control, networking cloud expert hand method database or expert hand method training large model optimization treatment data, to drive the robot to perform hand method treatment; the intelligent terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer and an industrial computer.

[0021] The working method of the present application comprises the following steps:

[0022] S1. System preparation, connect the integrated training glove of the humanoid robot hand to the humanoid robot arm with a six-dimensional force sensor at the end, and the edema treatment doctor, patient and family members can be used as the trainer user. Wear the training glove installed on the robot hand with the treatment hand, press the training glove switch, and prepare for hand-to-hand training of robot treatment method;

[0023] S2. Training method treatment, press the training button on the training glove to enter the training state, start real-time collection of multi-modal treatment data according to the application software setting parameters, and the trainer drives the robot hand installed on the humanoid robot arm to gently adhere to the planned treatment path on the patient's body surface, and runs in the order, direction, posture, force, speed and rhythm of each treatment path. The rhythm is the combination state of the interval time and force, speed of each treatment path. When completing the training of one treatment path, press the "training" button to stop training data collection. The intelligent terminal application software automatically stores the treatment data according to the trainer user type, which can be stored in the intelligent terminal storage or the main control computer storage of the humanoid robot arm. The data includes the bending degree sensor, IMU posture sensor in the training glove, and the motion trajectory, posture, palm force, vibration frequency and heating temperature of the robot hand at the end of the robot arm. According to the order of all planned treatment paths, complete the treatment training and data collection and storage of each treatment path;

[0024] S3. Reproduction of training method treatment, through the intelligent terminal application software wireless or wired communication with the main control computer of the humanoid robot arm, configure parameters for the collected training method data, including: the number of repeated traversals and the time length of each treatment path and all paths, and use these comprehensive data to drive the robot hand to reproduce the trainer's treatment method. When the trainer is an experienced edema treatment doctor and expert, the trainer's own training method treatment can be directly reproduced;

[0025] S4. Improve the training method treatment, based on the training treatment method data, keep the method treatment path trajectory and hand posture data unchanged, and increase, decrease and set the safety range to adjust the force, speed and rhythm of the treatment. A matching expert treatment data of the current patient's edema degree can be selected from the doctor's expert method database in the local storage or cloud data center, and the expert's force, speed and rhythm are extracted to generate new treatment data;

[0026] The training glove therapy data can also be sent to a cloud data center expert method training large model to directly generate expert level force, speed and rhythm of therapy, and reserve the training data of the method path trajectory coordinates and posture data, and be fused into new therapy data, and the new therapy data is downloaded to drive the machine hand to perform expert level method therapy.

[0027] When some scenes need to separate the training glove from the machine hand to remotely control the training machine hand therapy method, the built-in Bluetooth or WIFI module of the training glove can be used to realize real-time data transmission; and another training glove of the same type can be used to remotely control the training machine hand therapy method, and the built-in Bluetooth or WIFI module of the training glove can be used to realize real-time data transmission through 4G / 5G network, and the trainer wears the training glove, and remotely controls the training glove through short-range human eye or remote real-time video observation feedback, and remotely controls the machine hand to adhere to the treatment path marked on the patient's body surface to perform virtual method operation, and the machine hand performs real-time synchronous spatial follow-up in proportion, and the working method steps are the same as those of the integrated training glove humanoid machine hand.

[0028] The beneficial effects of the present application are: 1) the machine hand based on human-like muscle and skin materials, perception and action is suitable for medical method therapy, the hydrogel material provides human-like elastic touch, and the built-in vibration, heating and touch force monitoring module can improve the efficiency and effect of therapy; 2) the integrated training glove machine hand can realize hand-to-hand high-precision training of the human-like machine hand to reproduce the therapy method, and after the training glove and the machine hand are separated, the training can be suitable for remote control method therapy training and reproduction, and the components are suitable for replacement and maintenance; 3) through the high-precision digital acquisition, storage and data fusion method of therapy method training, the digital reproduction of the doctor's expert method can be realized, and the ordinary doctor, patient and family therapy edema method can be realized. BRIEF DESCRIPTION OF DRAWINGS

[0029] Figure 1 is a schematic diagram of the integrated structure of the humanoid machine hand and the training glove according to the present application;

[0030] Figure 2 is a schematic diagram of the integrated structure of the humanoid machine hand and the training glove according to the present application;

[0031] Figure 3 is a schematic diagram of the internal function module of the palm side of the skeleton plate of the humanoid machine hand unit according to the present application;

[0032] Figure 4 is a schematic diagram of the palm side of the humanoid machine hand unit according to the present application;

[0033] Figure 5 Figure 1 is a schematic diagram of the back of the training glove unit of the present application;

[0034] Figure 6 Figure 2 is a schematic diagram of the back of the training glove unit of the present application;

[0035] Figure 7 Figure 3 is a schematic diagram of the palm of the training glove unit of the present application;

[0036] Figure 8 Figure 4 is a flow chart of the working method of the present application.

[0037] Figure 1 is a schematic diagram of the back of the training glove unit of the present application; DETAILED DESCRIPTION

[0038] The present application will be further described in conjunction with the accompanying drawings and specific embodiments, but the embodiments of the present application are not limited thereto.

[0039] The present embodiment provides a humanoid robot hand for lymphedema hand therapy integrated training gloves, including a humanoid robot hand unit, a training glove unit and intelligent terminal application software, a common integrated six-dimensional force sensor 7-DOF humanoid robot arm, the end of the installation of the humanoid robot hand system, can realize the high precision and high quality digital acquisition, reproduction and promotion of the process of the doctor's hand method, suitable for hospitals, rehabilitation centers and family environment.

[0040] The humanoid robot hand integrated with the training gloves performs the integrated training therapy, including the following:

[0041] 1. System structure and hardware configuration

[0042] 1) The humanoid robot hand unit (1) includes:

[0043] The robot hand palm (11): using hydrogel material (elastic modulus 50kPa, thickness 4mm) casting, covering the robot hand skeleton plate and functional modules, the robot hand palm muscle and the surface of the human hand muscle and skin;

[0044] Machine hand skeleton plate (12): made of aluminum alloy material, the palm skeleton and the five-finger skeleton are integrally formed, each finger has three degrees of freedom, driven by a micro-finger driving motor (121), the motor model is DSM-10A (diameter 10 mm, torque 0.8 N·m), and the finger adopts a connecting rod structure; the basic function module includes:

[0045] Machine hand array film stress sensor (122): a distributed point array fixed film stress sensor can monitor the touch force of the palm and finger surfaces of the machine hand in real time, including: 12 sensing points are arranged in the palm area to form a touch force matrix, each sensing point is a circular film stress sensor with a diameter of 5 mm, a thickness of 1 mm, a range of 0-20 N, and an accuracy of ±0.05 N; five circular film stress sensors are embedded in the middle of each knuckle of the finger area, 14 touch force sensing monitoring points (2 knuckles for the thumb and 3 knuckles for each of the other four fingers), each point sensing sheet has a diameter of 5 mm, a thickness of 0.3 mm, a range of 0-15 N, and each point sensing sheet is connected through a film circuit and outputs real-time monitoring touch force values through a data interface;

[0046] Heating sheet (123): controllable working temperature range 25-45℃, control accuracy ±0.5℃, temperature value set by intelligent terminal application software;

[0047] Three-dimensional vibration motor (124): two vibration motors with a diameter of 8 mm are fixed orthogonally in one body, the vibration frequency is set to 80 Hz and the amplitude is set to 2 mm by the intelligent terminal application software;

[0048] Mechanical and electrical interface (13): fixed at the wrist, connected with the end interface of the humanoid robot arm, uses Type-C interface for power supply and communication, supports general communication protocols;

[0049] Array snap male head (14): 24 magnetic snap male heads are arranged on the back side of the machine hand skeleton plate (12) for mechanical connection and fixation with the training gloves.

[0050] 2) The training glove unit (2) is made of cloth with a certain elasticity to adapt to different trainers' hands, including:

[0051] Finger joint bending degree sensor (21): a Five-Flex film bending degree sensor can be used, which is sewn and installed at the center of the training glove knuckle to detect the bending angles of 14 joints of five fingers in real time;

[0052] Pose IMU sensor (22): MPU-6050 can be used, with a sampling frequency of 100 Hz, outputting acceleration and angular velocity data, used for collecting the spatial pose of the training glove motion;

[0053] Communication control module (23): The main control chip can be STM32F407, integrated with WiFi and Bluetooth modules;

[0054] Function keys (24): including four keys of switch, training, delete and execution;

[0055] Array snap female seat (25): 24 snap male heads are arranged on the palm surface of the training glove, which are one-to-one corresponding to the 24 magnetic snap female seats on the back of the humanoid robot hand, used for connecting and fixing the training glove and the humanoid robot hand for high-precision fitting training;

[0056] Training glove array film stress sensor (26): arranged below the snap array seat on the palm surface of the training glove, the film stress sensor is distributedly installed in the form of point array, which can monitor the touch force feedback of the palm surface and the finger pulp surface of the training glove in real time, including: 12 sensing points are arranged in the palm area to form a touch force matrix, each sensing point is a circular film stress sensor with a diameter of 5 mm, a thickness of 1 mm, a range of 0-20 N and an accuracy of ±0.05 N; one circular film stress sensor is embedded in the middle of each joint of the finger pulp in the finger area, and there are 14 touch force sensing monitoring points (2 joints for the thumb and 3 joints for each of the other four fingers), each point sensing sheet has a diameter of 5 mm, a thickness of 0.3 mm, a range of 0-15 N, and each point sensing sheet is connected through a film circuit and outputs real-time monitoring touch force values through a data interface, which is the same as the film stress sensor array (122) in the humanoid robot hand.

[0057] 3) Intelligent terminal application software, which can run on smart phones, tablets or other computers, communicates with the training glove and the main control computer of the humanoid robot arm through Bluetooth or WiFi modules, controls and manages functions including data acquisition, storage, visualization, model training and control instruction issuing, provides an interface for setting all parameters, and issues them to the robot hand or robot arm main control computer for execution; interacts with the expert skill database or expert skill large model of the cloud data center to generate new treatment data, the expert skill database of the cloud data center adopts SQLite relational database, and the expert skill training large model is constructed based on Transformer.

[0058] 4) Matching equipment, using a 7-DOF robot arm commonly used on the market with an end integrated six-dimensional force sensor, the end running track accuracy is less than 0.2 mm.

[0059] 2. Connection relationship

[0060] The array of snap male heads (14) on the back of the humanoid robot hand unit (1) is fixed to the array of snap female sockets (25) on the palm of the training glove unit (2) by pressing the two together, and the two are integrated into one, and can be separated by applying a certain separation force to the snap; the humanoid robot hand and the humanoid robot arm are connected and fixed by a standard mechanical interface, and the standard power supply and data interface of the two can be connected, and the training glove, the humanoid robot hand, and the robot arm can collect complete data during the operation of the hand therapy.

[0061] 3. Working method steps

[0062] S1. System preparation

[0063] The training glove and the humanoid robot hand are connected and integrated, the power of the robot arm is turned on, the back switch of the training glove is pressed, the communication connection is initialized, the trainer wears the training glove, and the hand-to-hand training of the robot hand is prepared, and the operation is carried out according to the doctor's planned treatment path trajectory;

[0064] S2. Manipulation training and data collection

[0065] The trainer (which can be a doctor, an expert, a patient, and a family member) presses the "training" button, and the robot hand is trained by hand-to-hand operation, and the system collects all the data in real time, including:

[0066] 1) Robot hand data:

[0067] Robot hand trajectory: The Cartesian coordinates P(t) = (x(t), y(t), z(t)) and the attitude quaternion Q1(t) = [q w1 , q x1 , q y1 , q z1 ] of the center of the robot hand installed at the end of the robot arm are calculated and output synchronously from the joint angle data of the robot arm;

[0068] Robot hand force: F arm (t) and torque M arm (t): obtained from the six-axis force sensor integrated at the end of the robot arm;

[0069] Robot hand feedback force: 12-point F hi (t) and finger 14-point F fi (t) of the palm, obtained from the thin film stress sensor array;

[0070] Settable parameters: The vibration frequency f v (t) and heating temperature T h (t) of the robot hand are set by the intelligent terminal software, and the system default values are not set;

[0071] 2) Training glove data includes:

[0072] Finger bending angle: detected by film bending sensor θ fi (t) (i=1…14, 14 joints of five fingers);

[0073] Glove posture: output posture quaternion Q2(t) = [q w2 , q x2 , q y2 , q z2 ] by IMU sensor at the center of the back of the glove;

[0074] 3) The trainer guides the robot hand by hand, gently and closely to the patient's body surface to plan the marker treatment path according to the treatment method, and runs each path for a running time t∈[0,T], the system monitors the force and feedback force applied by the robot hand in real time, and the alarm will sound and light when it exceeds the safety range;

[0075] 4) After training a path, press the "training" button to automatically store the training data package D i to the memory of the intelligent terminal or mechanical arm control computer, including timestamp and real-time data;

[0076] According to the planned treatment path sequence, train all the treatment path data of the patient's body surface markers one by one, add the trainer ID and the code of the patient's edema location / degree / cause, and form the patient's total treatment path training data D, which is automatically stored locally and uploaded to the cloud storage. The software gives information and voice prompts, and the trainer takes out the training gloves;

[0077] S3. Repetition of multiple long-term treatments

[0078] Select the stored total treatment path training data D in the intelligent terminal application software, set the treatment parameters including: treatment path training data D repetition number N, each treatment path D i repetition number N i , each path running interval time Δt, treatment method moving speed adjustment, treatment method safety force range, if not set, the default value is used;

[0079] Press the "treatment" button on the training glove, the mechanical arm drives the robot hand to reproduce according to the training method data, and completes the treatment according to the set parameters; when the trainer wants to improve his own practice method to the expert level, the method quality can be improved;

[0080] S4. Treatment method level improvement

[0081] 1) Trainer selects patient treatment path training data D from the cloud, and the system automatically matches expert techniques for similar conditions in the database according to the patient edema occurrence location / degree / cause information in D, sorts the list according to the similarity of the condition, and selects one of the expert technique treatment data D expert , retaining the hand method path trajectory coordinates and posture data in the training data D, and fusing the intensity, speed, and rhythm of the expert hand method treatment data D through a data fusion algorithm to generate new hand method treatment data D that has been improved to the expert level expert newE：

[0082] D newE = α × D + (1 - α) × D expert

[0083] Where: α ∈ [0, 1] is the adaptive fusion coefficient, when α is set to 0, it is the intensity, speed, and rhythm of the expert method, and for other values, it is the corresponding data fusion ratio of the trainer and the expert;

[0084] 2) The trainer can also choose to send the training data D to the cloud expert method pre-training large model, directly generate new treatment data package D that reaches the expert level of intensity, speed, and rhythm under the condition of retaining the hand method running trajectory coordinates and posture data in the training treatment data newG , downloaded to the robot host computer memory, and drive the robot hand to perform expert-level hand treatment.

[0085] When the integrated training gloves are separated from the robot hand, the built-in Bluetooth or WIFI module in the training gloves can be used to transmit data to achieve wireless short-range remote control of the training robot hand; the system supports remote control of the training robot hand with another training glove of the same type, and the built-in Bluetooth or WIFI module in the training glove can be used to connect to a 4G / 5G network to transmit data, achieving wireless remote control of the training robot hand, and the other implementation methods are the same, including:

[0086] 1) The trainer wears the training gloves and performs remote control of the robot hand according to the treatment path on the patient's body, and the IMU posture data and bending sensor data in the training gloves are transmitted to the host computer of the robot arm in real time to drive the robot arm to move synchronously at the same ratio.

[0087] 2) The pose T hand (t) of the robot hand is synchronized with the pose T glove (t) of the glove through the following formula:

[0088] T hand (t) = T glove (t) × T offset

[0089] In the formula: T offset ​The coordinate transformation matrix determined for the initial calibration stage is used to align the initial pose of the glove and the robot hand.

[0090] 3) The expected value of the force applied by the robot hand is set by the trainer through the application software, the value of the six-dimensional force sensor at the end of the mechanical arm is monitored in real time, the real-time touch force of the robot hand is monitored, and the values are displayed in real time on the intelligent terminal application software interface for observation and adjustment by the remote control trainer.

[0091] 4) The remaining steps are the same as the integration training of the glove and the humanoid robot hand.

[0092] Implementation effect

[0093] The embodiment fully verifies the realizability of the technical scheme of the application: high-precision and high-quality digital acquisition of expert treatment process data, reproducible expert-level treatment operation trajectory error less than 1mm, treatment force error less than 0.2N, treatment speed error less than 2%, supporting automatic and intelligent expert-level treatment improvement, supporting remote control training, and meeting the safety, digitization, intelligence, and expertization requirements of clinical treatment lymphatic drainage for tumor edema.

[0094] The above is only the preferred specific embodiment of the application, but the protection scope of the application is not limited thereto, any person skilled in the art can make equivalent replacement or change within the technical range disclosed by the application according to the technical scheme and inventive concept of the application, which should be covered within the protection scope of the application.

Claims

1. A humanoid robotic hand system for manual therapy integrated training glove, characterized in that, The application relates to a human-shaped robot hand unit, a training glove unit and intelligent terminal application software. The human-shaped robot hand unit and the training glove unit are detachably connected through a magnetic attraction snap array; when connected, the trainer wears the glove and simultaneously and spatially trains the robot hand; when separated, the trainer wears the glove and remotely controls the robot hand to synchronously train; and the same type of training gloves can be connected to remotely control the robot hand to synchronously train. The human-shaped robot hand unit comprises a robot hand skeleton plate and a robot hand palm, the wrist part of the robot hand is provided with a standard mechanical interface and a power supply data interface, is used for connecting an end human-imitating mechanical arm with a six-dimensional force sensor, realizes data communication and integrated linkage control, and supports direct replacement of a commonly-used mechanical dexterous hand in the market. The training glove unit is integrated with a multi-mode sensor module, a communication control module and a man-machine interaction module. The intelligent terminal application software is used for data acquisition and storage, communication control, networking cloud expert hand method database or expert hand method training large model optimization treatment data, and the intelligent terminal can be an intelligent mobile phone, a tablet computer, a notebook computer, a desktop computer and an industrial computer.

2. The system of claim 1, wherein, The robot hand skeleton plate is made of metal or plastic hard material and comprises a palm skeleton and a finger skeleton; the inside is integrated with a finger driving motor, human-shaped finger mechanical parts and a functional module, can realize adaptive deformation actions such as grabbing, holding and pinching, and the outside is provided with an array snap female seat for connecting and fixing the training glove through threads or glue joint.

3. The system of claim 2, wherein, The functional module installed on the palm side of the robot hand skeleton plate comprises: an array thin film stress sensor module fixed on the palm area and the finger pulp area and used for real-time detection of contact force in the treatment process; a three-dimensional vibration motor module integrated with two same type micro vibration motors fixed on a rigid base in a quadrature mode, the motor rotating speed and phase difference are set through the intelligent terminal application software, a spiral three-dimensional vibration force is generated, and the treatment edema effect is improved; a heating sheet module with a temperature control heating range of 25-45 DEG C and used for low temperature environment and auxiliary treatment.

4. The system of claim 2, wherein, The robot hand palm is made of a water gel material imitating human muscles and skin, is poured on the robot hand skeleton plate and the functional module through a real hand guide mold process; The robot hand skeleton plate supports direct replacement of a commonly-used mechanical dexterous hand in the market, the functional module is integrated, the elastic touch palm of the high human-imitating hand is made of the water gel material.

5. The system of claim 1, wherein, The multi-mode sensor module of the training glove unit comprises a thin film inductive bending degree sensor module integrated on the five finger pulp sides or the finger back sides and used for real-time detection of the bending angles of the five finger joints; an array thin film type stress sensor module integrated on the glove palm and the five finger pulp parts and used for real-time detection of training method strength; an IMU position sensor module integrated on the glove back and used for real-time detection of training method space attitude data.

6. The system of claim 1, wherein, The communication control module of the training glove unit is integrated on the glove back, comprises a microprocessor, a memory, a wireless communication module, a charging battery and an electrical data interface, is automatically switched into wired communication and power supply when connected with the robot hand unit, is automatically switched into wireless communication and battery power supply when separated, the man-machine interaction module comprises magic tape binding belts and functional buttons, the buttons comprise system switch, start training, delete data and execute treatment buttons.

7. A method of treating a condition using the system of any one of claims 1-6, wherein the method comprises administering the system to the subject. Comprising the following steps: S1. System preparation: connect the anthropomorphic robot hand integrated with training gloves to the anthropomorphic robot arm with six-dimensional force sensors, the trainer wears gloves and presses the switch to prepare for training; S2. Data acquisition: press the training button, the trainer drives the robot hand to run along the planned treatment path, the system collects multi-modal data in real time, stops collecting and stores after completing a single path, and repeats the entire path training; S3. Reproduce the method: configure path repetition, duration and other parameters through the intelligent terminal, and drive the robot hand to reproduce the training method; S4. Method improvement: keep the path trajectory and posture data unchanged, adjust the treatment parameters, or match the cloud expert data fusion to generate new treatment data, or send data to the cloud big model to generate expert-level data, and drive the robot hand to execute the treatment.

8. The method of claim 7, wherein, The multi-modal data in step S2 includes: finger bending angle, pose data of training gloves, motion trajectory coordinates, attitude, force, vibration frequency and temperature control data of robot hand, and the data is stored in the storage of intelligent terminal or mechanical arm host computer.

9. The method of claim 7, wherein, The treatment data fusion formula in step S4 is D newE = α × D + (1 - α) × D expert , the machine hand trajectory, posture data in the training data D are kept unchanged, and the force, speed and rhythm data in the expert hand method data are fused, wherein α ∈ [0, 1] is a set fusion coefficient, D is training data, D expert is an expert hand method data automatically matched in the cloud expert hand method database, and D newE is new treatment data of fused expert level generated.

10. The method of claim 7, wherein, During remote training, the IMU pose of the training glove and the finger bending sensor data are transmitted to the master computer of the robot arm, and the pose T hand (t) of the robot hand at time t is calculated according to the formula T glove (t) = T hand (t) = T glove (t) x T offset , T offset is the coordinate transformation matrix of the initial calibration.

Images