Wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical cooperative driving and rehabilitation training method thereof

Abstract

The invention discloses a wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical cooperative driving, the wrist and finger joint rehabilitation robot comprises a base and a finger joint flexion and extension mechanism, an elbow adjusting mechanism is fixedly arranged on the base, a pronation and supination mechanism is fixed on the elbow adjusting mechanism, a wrist fixing mechanism is fixed on the pronation and supination mechanism, and the finger joint flexion and extension mechanism is fixed on the wrist fixing mechanism. An ulnar-radioulnar deviation mechanism is fixed to the wrist fixing mechanism, a flexion and extension mechanism is fixed to the ulnar-radioulnar deviation mechanism, an electric needle stimulation mechanism is further fixed to the ulnar-radioulnar deviation mechanism, a visual recognition device is further arranged on the ulnar-radioulnar deviation mechanism, an electric air pump is further connected to the finger joint flexion and extension mechanism, and a strain gauge bending sensor is further arranged on the finger joint flexion and extension mechanism. And the electromyographic signal collector is arranged on the forearm of the user. The problem that equipment in the prior art generally lacks combination of traditional Chinese medicine acupoint stimulation and western medicine exercise rehabilitation is solved. The invention further discloses a rehabilitation training method of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical cooperative driving.

Description

Technical Field

[0001] The present invention belongs to the technical field of wrist and finger joint rehabilitation equipment, and relates to a wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive. The present invention also relates to a rehabilitation training method using the above-mentioned wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive. Background Art

[0002] Traditional methods for wrist and finger joint evaluation and training usually rely on manual operation, and the treatment effect is significantly affected by the experience of therapists. Rehabilitation therapists perform joint mobilization, muscle strength training, etc. through manual techniques. Such rehabilitation means are difficult to accurately stimulate specific acupoints on the hand, mostly for large-area massage or simple electrical stimulation, and cannot achieve personalized and precise rehabilitation treatment. It is difficult to ensure the consistency of daily training, and there are problems such as unstable training intensity, high labor costs, and difficult quantification of curative effects, resulting in uneven rehabilitation effects.

[0003] Huazhong University of Science and Technology developed a portable wearable two-degree-of-freedom rehabilitation device based on a hybrid drive of pneumatic muscles and motors (Huang Ming, Huang Xinhan, Yuan Yong, etc.; Proxy sliding mode control method for a 2-DOF wrist joint rehabilitation robot [J]. Journal of Huazhong University of Science and Technology: Natural Science Edition, 2015 (S1): 298-301.). Among them, the pneumatic muscle mainly drives the flexion and extension movements of the wrist joint, and the motor completes the drive of the ulnar deviation and radial deviation movements of the wrist joint. The system uses Hall sensors and pressure sensors. This robot is light to wear and enables the rehabilitator to perform passive rehabilitation training according to a predetermined trajectory. However, due to the characteristics of the pneumatic muscle itself, the motion linearity of this device is insufficient, the accuracy is low, and the control is relatively difficult.

[0004] Harbin Institute of Technology developed a 5-DOF wrist joint rehabilitation device (Luo Yang. Research on a post-operative rehabilitation exoskeleton system for the wrist joint based on myoelectric signal control [D]. Harbin Institute of Technology, 2018.). This device uses a driving method in which the motor winds and unwinds a rope to pull the robot joint, making the robot have good flexibility. It uses two passive degrees of freedom to compensate for the deviation between the wrist joint and the robot rotation axis, and realizes passive, active, and impedance control of the user through measuring and sEMG signals. The circumduction of the wrist of this device uses a telescopic universal hinge for transmission, and there will be interference with the human body at a certain angle.

[0005] In summary, existing devices generally lack the combination of traditional Chinese medicine acupoint stimulation and Western medicine exercise rehabilitation, and cannot give full play to the advantages of collaborative treatment. Summary of the Invention

[0006] The object of the present invention is to provide a wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive, which solves the problem that existing devices generally lack the combination of traditional Chinese medicine acupoint stimulation and Western medicine movement rehabilitation in the prior art.

[0007] The present invention also discloses a rehabilitation training method using a wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive.

[0008] The technical solution adopted by the present invention is that a wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive includes a base and a finger joint flexion and extension mechanism. An elbow adjustment mechanism is fixedly arranged on the base, a pronation and supination mechanism is fixed on the elbow adjustment mechanism, a wrist fixation mechanism is fixed on the pronation and supination mechanism, a ulnar and radial deviation mechanism is fixed on the wrist fixation mechanism, a flexion and extension mechanism is fixed on the ulnar and radial deviation mechanism, an electroacupuncture stimulation mechanism is also fixed on the ulnar and radial deviation mechanism, a visual recognition device is also arranged on the ulnar and radial deviation mechanism, the finger joint flexion and extension mechanism is also connected with an electric air pump, a strain gauge bending sensor is also arranged on the finger joint flexion and extension mechanism, an electromyogram signal collector arranged on the user's forearm, the strain gauge bending sensor and the electromyogram signal collector are electrically connected to a data acquisition card through wires, and the data acquisition card, the visual recognition device, and the electric air pump are respectively electrically connected to a control unit through wires.

[0009] The feature of the first technical solution of the present invention is further that:

[0010] The finger joint flexion and extension mechanism is a bellows-type pneumatic rehabilitation glove, which includes a glove body. Actuators are arranged at the positions corresponding to each finger on the back of the glove body, and each actuator is adapted to the length of the corresponding finger. The actuator includes a flexible restraint layer arranged on the back of each finger, the flexible restraint layer is distributed along the length of the finger, and a plurality of fixed connectors are evenly arranged on the flexible restraint layer. A section of bellows is connected between adjacent two fixed connectors. Air channels are arranged inside the plurality of fixed connectors, and the two ends of the air channels are connected to the bellows connected to the two ends of the fixed connectors. One end of the fixed connector farthest from the fingertip, which faces away from the fingertip, is also connected with an intake pipe. The intake pipe is connected to one end of the air channel of the corresponding fixed connector, and the intake pipe is connected to the electric air pump through a pipeline. Each electric air pump is electrically connected to the control unit through a wire. The position of the flexible restraint layer corresponding to each bellows is set as a square wave structure, and a strain gauge bending sensor is embedded between the flexible restraint layer corresponding to each bellows and the glove body.

[0011] The electroacupuncture stimulation mechanism includes a servo fixed on the ulnar radial deflection mechanism on the side close to the palm, a connecting plate is fixedly connected to the output shaft of the servo, an electric push rod is fixed on the connecting plate, a linear motor is fixed on the other end of the electric push rod, a connecting plate a is fixedly connected to the extended end of the linear motor, the linear motor moves in the direction close to the inner side of the palm and away from the inner side of the palm, a sleeve is fixedly connected to the side of the connecting plate a close to the palm, one end of a spring a is fixed to the bottom of the sleeve, an electroacupuncture is fixedly connected to the other end of the spring a, a spherical contact is used for the electroacupuncture head of the electroacupuncture and the surface is silver-plated, the servo, the electric push rod, the electroacupuncture and the linear motor are all electrically connected to a control unit through wires, and a visual recognition device includes a binocular camera and a visual recognition module arranged on the ulnar radial deflection mechanism, the binocular camera faces the side of the palm, the binocular camera is electrically connected to the visual recognition module through wires, and the visual recognition module is connected to the control unit through wires.

[0012] The elbow adjustment mechanism includes a bracket installed at one end of the base, the bracket is provided with a bracket a fixedly provided and horizontally provided, the bracket a is provided with two guide rails parallel to each other, the guide rails are slidably connected with a slider, an adjustment box is fixed on the slider, a transmission mechanism is provided in the adjustment box, an elbow bracket is connected to the transmission mechanism, an elbow support pad is provided on the upper surface of the elbow bracket, the transmission mechanism includes a height adjustment shaft rotatably installed in the adjustment box, one end of the height adjustment shaft extends out of the adjustment box and is connected with a height adjustment knob, a gear is fixedly installed at a position where the height adjustment shaft is located in the adjustment box, the gear is meshed with a vertically provided rack, the elbow bracket is fixed at the upper end of the rack, a vertically distributed slide groove is also provided on the inner wall of the adjustment box, a slider a is slidably connected in the slide groove, the upper end of the slider a is connected to the bottom of the elbow bracket, a mounting bracket is fixedly provided on the end of the bracket away from its installation on the base, and the pronation and supination mechanism is installed on the mounting bracket;

[0013] The pronation and supination mechanism comprises an external gear slewing bearing, the inner ring of the external gear slewing bearing is fixed on the mounting bracket, the outer gear ring of the external gear slewing bearing is meshed with a gear a, a motor is also fixed on the bracket a through a motor support, the gear a is fixedly connected to the gear shaft, the gear shaft is fixedly connected to the output shaft of the motor through a coupling, and the wrist fixing mechanism is fixed on the side of the external gear ring away from the elbow bracket; the sliding direction of the slider on the guide rail is parallel to the axis of the external gear slewing bearing;

[0014] The wrist fixing mechanism includes a lower support adjusting frame fixed on the side of the external gear ring away from the elbow bracket. The lower support adjusting frame is horizontally arranged. A pedestal bearing is fixedly arranged on the upper surface of the lower support adjusting frame. A vertically arranged lead screw is inserted through the pedestal bearing. The lower end of the lead screw passes through the lower support adjusting frame and is connected with a height adjusting knob a through a coupling a. The upper end of the lead screw is fixedly connected with a horizontally arranged lower bracket through a lead screw nut pair. Guide rails a vertically arranged are also fixedly connected to the positions on both sides of the lower bracket on the lower support adjusting frame. Sliders b are fixedly connected to the positions of the lower bracket corresponding to the guide rails a on both sides. The sliders b are slidably connected to the corresponding guide rails a on one side. Left brackets and right brackets are respectively fixed on both sides of the lower bracket. An upper bracket is fixedly connected to the left bracket. The upper bracket is in an inverted "L" shape. A lower cushion, an upper cushion and a right cushion are respectively arranged on the upper surface of the lower bracket, the lower surface of the upper bracket and the left side surface of the right bracket. A plurality of springs are evenly arranged below the right cushion on the side of the right bracket close to the left bracket. The ulnar and radial deviation mechanism is fixed on the lower support adjusting frame;

[0015] A horizontally arranged lead screw penetrates through the lower bracket. The left bracket and the right bracket are fixed on the lead screws extending out of both sides of the lower bracket through nuts; A plurality of adjusting holes are evenly arranged on the left bracket from top to bottom. A plurality of adjusting holes are also arranged on the lower end of the upper bracket from top to bottom. The left bracket and the upper bracket are fixed through pins inserted through the corresponding adjusting holes; One end of the lead screw is also connected with a wrist joint clamping adjusting knob through a coupling b;

[0016] The ulnar and radial deviation mechanism includes a lower left bracket and a lower right bracket fixed at both ends of the upper surface of the lower support adjusting frame. The upper ends of the lower left bracket and the lower right bracket are respectively connected with a middle left bracket and a middle right bracket. The upper end of the middle left bracket is rotatably connected with an upper left bracket through a transmission shaft. The upper end of the middle right bracket is rotatably connected with an upper right bracket through a rotating shaft. The other ends of the upper left bracket and the upper right bracket are jointly fixedly connected with a connecting plate. A driving motor is installed on the middle left bracket through a motor support a. A bevel gear is installed on the output shaft of the driving motor. A bevel gear a is installed on the transmission shaft installed on the middle left bracket and the upper left bracket. The bevel gear a and the bevel gear are meshed. The axial direction of the output shaft of the driving motor is parallel to the axial direction of the external tooth type slewing bearing. The flexion and extension mechanism is connected to the connecting plate; The binocular camera is fixed at the upper end of the middle right bracket; The servo motor is installed on the side of the middle right bracket close to the palm center;

[0017] A linear guide rail is provided on the connecting plate. A waist-shaped chute is also provided on the connecting plate. The extending direction of the chute is the same as that of the linear guide rail. As shown in the figure, the flexion and extension mechanism includes a bracket a. A slider c is correspondingly arranged below the bracket a and corresponding to the linear guide rail. The bracket a and the connecting plate are slidably connected through the cooperation of the slider c and the linear guide rail. A transmission shaft b is rotatably installed on the bracket a. One end of the transmission shaft b extends out of the upper surface of the bracket a and is provided with a synchronous pulley. The other end of the transmission shaft b extends out of the lower surface of the bracket a and the chute on the connecting plate in sequence and is connected with a flexion and extension connecting rod. A chute is provided on the connecting plate. When the bracket a slides on the linear guide rail, the transmission shaft b slides in the chute. The flexion and extension connecting rod extends towards the side away from the external tooth type slewing bearing, and a guide rail b is provided on the upper surface of the flexion and extension connecting rod. A connecting slider is provided on the guide rail b. A handle is fixedly connected below the slider. A flexion and extension motor is provided on the lower surface of the end of the bracket a far away from the transmission shaft b. The output shaft of the flexion and extension motor passes through the bracket a and is connected with a synchronous pulley a. The synchronous pulley and the synchronous pulley a are connected by a synchronous belt in transmission. A plurality of pin holes are evenly and correspondingly arranged on the connecting plate along the sliding direction of the bracket a. The bracket a and the connecting plate are fixed by inserting a pin c through the corresponding pin holes.

[0018] The second technical solution adopted by the present invention is: a wrist and finger joint rehabilitation training method based on dynamic acupoint visual recognition and electrical collaborative drive. Using the above-mentioned wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive, including finger flexion and extension training. The specific process is as follows:

[0019] According to the thickness and length of the user's arm, adjust the elbow adjustment mechanism, wrist fixation mechanism, pronation and supination mechanism, ulnar and radial deviation mechanism, and flexion and extension mechanism to suitable positions, so that the user's forearm is fixed on the robot. And according to the actual situation of the user's fingers, set the inflation rate and inflation volume of the electric air pump corresponding to each finger and the safety threshold of the strain gauge bending sensor corresponding to each finger. Then put the glove body on the user's hand. The control unit controls each electric air pump to inflate at a certain speed according to the preset parameters, so that the fingers gradually bend to achieve fist clenching training. The control unit controls each electric air pump to deflate reversely to achieve extension training. During fist clenching training and extension training, the strain gauge bending sensor monitors the finger bending degree in real time. If it exceeds the safety threshold, the corresponding electric air pump immediately stops working and alarms.

[0020] The feature of the second technical solution of the present invention also lies in: it also includes electrostimulation rehabilitation, specifically:

[0021] According to the thickness and length of the user's arm, adjust the elbow adjustment mechanism, wrist fixation mechanism, pronation and supination mechanism, ulnar and radial deviation mechanism, and flexion and extension mechanism to suitable positions, so that the user's forearm is fixed to the robot. Then, the servo motor rotates to drive the electric push rod to rotate. Combining with the elongation and shortening range of the electric push rod, the electro-acupuncture needle can stimulate all acupoints on the palm. Then, the binocular camera collects the image of the inner side of the palm and uploads it to the visual recognition module. The visual recognition module identifies the positions of the key points of the hand, and then calculates the corresponding acupoints according to the positions of the key points of the hand. Then, it feeds back the corresponding acupoint coordinates to the control unit. The control unit controls the movement of the servo motor and the electric push rod, moves the electro-acupuncture needle to the corresponding acupoint position, and then controls the linear motor to move towards the palm side, so that the electro-acupuncture needle pricks the corresponding acupoint, and then controls the electro-acupuncture needle to discharge for electrostimulation.

[0022] A strain gauge is set on the spherical contact head at the end of the electro-acupuncture needle to monitor the contact pressure in real time. The pressure in the treatment window is 0.2 - 0.8N. If the pressure > 1N, the control unit controls the linear motor to retract the electro-acupuncture needle by 0.5mm to prevent overpressure damage. If the pressure < 0.2N, it is considered that the stimulation is not effective, and the control unit needs to continue to control the linear motor to push the electro-acupuncture needle forward.

[0023] When performing electrostimulation rehabilitation training, set the safety threshold of the output current of the electro-acupuncture needle on the control unit. An electromyogram signal collector is installed on the user's forearm. The electromyogram signal collector collects the electromyogram signals of the target muscle in real time during electrostimulation, calculates the integrated electromyogram value iEMG and the root mean square value RMS, quantifies the degree of muscle activation, and dynamically adjusts the output current of the electro-acupuncture needle. Specifically, when the electromyogram amplitude < 50μV, it is considered that the stimulation is insufficient, and the control unit controls the output current of the electro-acupuncture needle to increase by 10%. When the electromyogram amplitude > 150μV, it is considered that the activation is excessive, and the output current of the electro-acupuncture needle is reduced to the safety threshold to avoid muscle fatigue.

[0024] The visual recognition module identifies the positions of the key points of the hand, and then the specific process of calculating the corresponding acupoints according to the positions of the key points of the hand is as follows:

[0025] The visual recognition module uses the MediaPipe Hands model to detect the key points of the hand according to the collected image, identifies 21 key points of the hand, and obtains the normalized coordinates of the 21 key points of the hand;

[0026] Calculate the positions of the corresponding key points in the pixel coordinate system according to the normalized coordinates. Specifically:

[0027] x 像素 = x 归一化 × w; y 像素 = y 归一化 × h

[0028] Among them, x 归一化 and y归一化 are the horizontal position of the corresponding key point in the image and the vertical position of the key point in the image, where x 归一化 increases from left to right, and y 归一化 increases from top to bottom; x 像素 and y 像素 are respectively the coordinate values in the pixel coordinate system after the conversion of x 归一化 and y 归一化 ; w and h are respectively the pixel sizes of the corresponding width and height under the images collected by the binocular camera;

[0029] Next, boundary detection is performed. Using min(max(...)) to limit the pixel coordinates within a reasonable range, specifically: min(max(x 像素 , w - 1)); min(max(y 像素 , h - 1)), that is, limit x 像素 within the range of [0, w - 1], and limit y 像素 within the range of [0, h - 1];

[0030] Then, calculate the acupoint pixel coordinates of Hegu acupoint, Laogong acupoint, Dazhui acupoint, Shaoshang acupoint, and Zhongzhu acupoint according to the pixel coordinates of the corresponding key points after boundary detection, specifically:

[0031] Hegu acupoint: According to the Euclidean distance d between the thumb fingertip THUMB_TIP and the index finger fingertip INDEX_FINGER_TIP, if d < 0.03, then calculate the pixel coordinates of the Hegu acupoint in the following way:

[0032]

[0033] where, x THUMB_TIP and y THUMB_TIP are respectively the pixel coordinates corresponding to the conversion of the horizontal position and vertical position coordinates of the thumb fingertip THUMB_TIP in the image, x INDEX_TIP and y INDEX_TIP are respectively the pixel coordinates corresponding to the conversion of the horizontal position and vertical position coordinates of the index finger fingertip INDEX_FINGER_TIP in the image, x 合谷 and y 合谷 are respectively the pixel coordinates of the Hegu acupoint in the horizontal position and vertical position directions;

[0034] If d ≥ 0.03, then prompt the user to adjust the posture, move the hand slightly away from the camera, and then recalculate until d < 0.03;

[0035] Laogong acupoint:

[0036] where, x劳宫 and y 劳宫 are the pixel coordinates of the Laogong acupoint in the horizontal and vertical position directions respectively; and are the pixel coordinates corresponding to the conversion of the horizontal and vertical position coordinates of the INDEX_FINGER_MCP of the index finger metacarpophalangeal joint in the image, x MIDDLE_MCP and y MIDDLE_MCP are the pixel coordinates corresponding to the conversion of the horizontal and vertical position coordinates of the MIDDLE_FINGER_MCP of the middle finger metacarpophalangeal joint in the image respectively;

[0037] Dazhui acupoint: x 大椎 = x PINKY_MCP ; y 大椎 = y PINKY_MCP

[0038] wherein, x 大椎 and y 大椎 are the pixel coordinates of the Dazhui acupoint in the horizontal and vertical position directions respectively; x PINKY_MCP and y PINKY_MCP are the pixel coordinates corresponding to the conversion of the horizontal and vertical position coordinates of the PINKY_MCP of the little finger metacarpophalangeal joint in the image respectively;

[0039] Shaoshang acupoint: x 少商 = x INDEX_TIP ; y 少商 = y INDEX_TIP

[0040] wherein, x 少商 and y 少商 are the pixel coordinates of the Shaoshang acupoint in the horizontal and vertical position directions respectively;

[0041] Zhuzhu acupoint:

[0042] wherein, x 中渚 and y 中渚 are the pixel coordinates of the Zhuzhu acupoint in the horizontal and vertical position directions respectively; x PINKY_MCP and y PINKY_MCP are the pixel coordinates corresponding to the conversion of the horizontal and vertical position coordinates of the PINKY_MCP of the little finger metacarpophalangeal joint in the image respectively; x RING_MCP and y RING_MCP are the pixel coordinates corresponding to the conversion of the horizontal and vertical position coordinates of the RING_FINGER_MCP of the ring finger metacarpophalangeal joint in the image respectively;

[0043] Then, convert the pixel coordinates of the acupoints of Hegu (LI4), Laogong (PC8), Dazhui (GV14), Shaoshang (LU11), and Zhongzhu (SJ3) into coordinates in the mechanical coordinate system. Then, through inverse kinematics calculation, drive the rotation angle of the servo motor and the telescopic amount of the electric push rod, and send them to the control unit.

[0044] It also includes wrist and finger joint movement function training, specifically: according to the thickness and length of the user's arm, jointly adjust the axis of the forearm through the elbow adjustment mechanism and the wrist fixation mechanism to make it coincide with the pronation and supination movement axis of the wrist joint, adjust the distance between the elbow bracket and the pronation and supination mechanism, then fix the user's forearm on the robot through the wrist fixation mechanism, and realize the ulnar and radial deviation movement of the wrist joint through the ulnar and radial deviation mechanism, and realize the flexion and extension movement function through the flexion and extension mechanism. During the movement process, collect the output data of the relevant muscles during the movement of the wrist joint through the electromyogram signal collector and upload it to the control unit.

[0045] The beneficial effects of the present invention are:

[0046] The present invention combines training with motion measurement and evaluation, can realize the pronation and supination movement, ulnar and radial deviation movement, flexion and extension movement of the wrist joint, and the flexion and extension movement of the finger joints, includes seven adjustable degrees of freedom, realizes the adjustment of the mechanical structure size according to the arm sizes of different individuals, and compensates for the joint misalignment of the human-machine matching.

[0047] In the present invention, the finger joint flexion and extension mechanism passes positive pressure gas or negative pressure gas through an electric air pump to realize the bending / extension movement when the finger joint rotates, and according to the feedback of the strain gauge bending sensor, the control board adjusts the air pump pressure and the air charging and discharging rate in real time to realize precise control of the flexion and extension angle.

[0048] The present invention uses visual recognition and bio-signal feedback to realize the precise matching of dynamic tracking of hand acupoints and electroacupuncture stimulation, breaks through the limitation of large positioning deviation of traditional equipment, and enables the therapeutic effect to directly reach the target.

[0049] The finger joint flexion and extension mechanism of the present invention simulates the natural movement trajectory of the human body based on the bionic principle, avoids the compression injury of the joint by the rigid mechanical structure while providing effective mechanical assistance, and greatly improves the training comfort of patients.

[0050] The present invention enables the integration mechanism of traditional Chinese and Western medicine to couple the stimulation of hand acupoints with the rehabilitation training of motor function, not only strengthens the activation of local muscles and nerves, but also promotes the overall qi and blood circulation, forming a multi-dimensional therapeutic effect. At the same time, it can sense the mechanical state and physiological response of the hand in real time, dynamically adjust the training intensity and stimulation parameters, and balance the rehabilitation efficiency and safety. Through progressive training, patients gradually reconstruct the physiological movement pattern, reduce abnormal compensation, and finally realize the synchronous optimization of function recovery and movement coordination, providing a more natural, precise and safe solution for the rehabilitation of wrist and finger joints. Description of the Drawings

[0051] Figure 1 It is a schematic structural diagram of the dorsal side of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0052] Figure 2 It is a schematic structural diagram of the palm side of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0053] Figure 3 It is a schematic structural diagram of the flexion and extension mechanism of the middle finger joint of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0054] Figure 4 It is a schematic structural diagram of the electroacupuncture stimulation mechanism of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0055] Figure 5 It is a schematic structural diagram of the elbow joint adjustment mechanism of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0056] Figure 6 It is a schematic structural diagram of the pronation and supination mechanism of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0057] Figure 7 It is a schematic structural diagram of the wrist fixation mechanism of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0058] Figure 8 It is a schematic structural diagram of the ulnar and radial deviation mechanism of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0059] Figure 9 It is a schematic structural diagram of the flexion and extension mechanism structure of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0060] Figure 10a It is a state diagram of the negative air pressure of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0061] Figure 10b It is a state diagram of the non-inflated wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention;

[0062] Figure 10c It is a state diagram of the positive air pressure of the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive of the present invention.

[0063] In the figure: 1. Base, 2. Elbow adjustment mechanism, 3. Pronation and supination mechanism, 4. Wrist fixation mechanism, 5. Ulnar and radial deviation mechanism, 6. Flexion and extension mechanism structure, 7. Finger joint flexion and extension mechanism, 8. Electroacupuncture stimulation mechanism, 9. Bracket a, 10. Slide block, 11. Bracket, 11-1. Mounting bracket, 12. Guide rail, 13. Slide block a, 14. Elbow bracket, 15. Elbow cushion, 16. Rack, 17. Gear, 18. Adjustment box, 19. Height adjustment knob, 20. Gear a, 21. Motor, 22. Motor support, 23. Coupling, 24. Outer gear ring, 25. Inner ring, 26. Height adjustment knob a, 27. Coupling a, 28. Bearing with housing, 29. Lower support adjustment frame, 30. Lead screw, 31. Screw rod, 32. Nut, 33. Lower bracket, 34. Left bracket, 35. Plug, 36. Upper bracket, 37. Upper cushion, 38. Right cushion, 39. Right bracket, 40. Spring, 41. Clamping adjustment knob, 42. Coupling, 43. Guide rail a, 44. Slide block b, 45. Lower cushion, 46. Lower left bracket, 47. Lower middle bracket, 48. Motor support, 49. Driving motor, 50. Bevel gear, 51. Transmission shaft, 51-1. Rotating shaft, 52. Bevel gear a, 53. Upper left bracket, 54. Connecting plate, 55. Linear guide rail, 56. Upper right bracket, 57. Upper middle bracket, 58. Lower right bracket, 59. Handle, 60. Connecting slide block, 61. Flexion and extension connecting rod, 62. Guide rail b, 63. Chute, 65. Bracket a, 66. Plug c, 67. Slide block c, 68. Transmission shaft b, 69. Synchronous belt pulley, 70. Synchronous belt, 71. Synchronous belt pulley a, 72. Flexion and extension motor, 73. Glove body, 74. Air inlet pipe, 75. Bellows, 76. Fixed connector, 77. Flexible restriction layer, 78. Air passage, 79. Electroacupuncture needle, 80. Sleeve, 81. Spring a, 82. Electric push rod, 83. Binocular camera, 84. Connecting plate, 85. Steering gear, 86. Linear motor, 87. Connecting plate a. Specific implementation manner

[0064] The following is a detailed description in combination with the specific implementation manner.

[0065] Embodiment 1

[0066] The wrist and finger joint rehabilitation robot of the present invention based on dynamic acupoint visual recognition and electrical collaborative drive has a structure as shown in Figure 1 and 2As shown in the figure, it includes a base 1 and a knuckle flexion and extension mechanism 7. An elbow adjustment mechanism 2 is fixedly arranged on the base 1. A pronation and supination mechanism 3 is fixed on the elbow adjustment mechanism 2. A wrist fixation mechanism 4 is fixed on the pronation and supination mechanism 3. A ulnar and radial deviation mechanism 5 is fixed on the wrist fixation mechanism 4. A flexion and extension mechanism 6 is fixed on the ulnar and radial deviation mechanism 5. An electroacupuncture stimulation mechanism 8 is also fixed on the ulnar and radial deviation mechanism 5. A visual recognition device is also arranged on the ulnar and radial deviation mechanism 5. The knuckle flexion and extension mechanism 7 is also connected to an electric air pump. A strain gauge bending sensor is also arranged on the knuckle flexion and extension mechanism 7. An electromyogram signal collector arranged on the user's forearm, the strain gauge bending sensor and the electromyogram signal collector are electrically connected to a data acquisition card through wires. The data acquisition card, the visual recognition device, and the electric air pump are respectively electrically connected to a control unit through wires.

[0067] As Figure 3 shown, the knuckle flexion and extension mechanism 7 is a bellows-type pneumatic rehabilitation glove, including a glove body 73. Actuators are arranged at the positions corresponding to each finger on the back of the glove body 73. Each actuator is adapted to the length of the corresponding finger. The actuator includes a flexible restraint layer 77 arranged on the back of each finger. The flexible restraint layer 77 is distributed along the length of the finger, and a plurality of fixed connectors 76 are evenly arranged on the flexible restraint layer 77. A section of bellows 75 is connected between two adjacent fixed connectors 76. Air channels 78 are arranged inside a plurality of fixed connectors 76. Both ends of the air channel 78 and the bellows 75 connected to both ends of the fixed connector 76. One end of the fixed connector 76 farthest from the fingertip facing away from the fingertip is also connected to an air inlet pipe 74. The air inlet pipe 74 is connected to one end of the air channel 78 of the corresponding fixed connector 76. The air inlet pipe 74 is connected to the electric air pump through a pipeline. Each electric air pump is electrically connected to the control unit through a wire. The position of the flexible restraint layer 77 corresponding to each bellows 75 is set as a square wave-like structure. A strain gauge bending sensor is embedded between the flexible restraint layer 77 corresponding to each bellows 75 and the glove body 73. The flexible restraint layer 77 and the fixed connector 76 are made of the same material, and are both made by 3D printing with photosensitive resin material.

[0068] Its working principle is: As Figure 10a - Figure 10c shown, after positive pressure gas is introduced into the actuator in the hollow air channel, each bellows chamber will expand and elongate. In the case of input negative air pressure, the gas in the bellows chamber is sucked out, and the bellows chamber will contract and shorten. By restricting the length of one side of the bellows by the restraint layer, under the state of loading air pressure, a deformation difference is generated between the side with unchangeable length and the side that can elongate and shorten, so that the bellows actuator can generate forward elongation bending and reverse contraction bending motions under positive and negative air pressures, corresponding to the bending / extension motions when the finger joint rotates, realizing the active bidirectional motion of the actuator. At the same time, according to the data feedback of the strain gauge bending sensor, the control unit adjusts the pressure and the air charging and discharging rate of the electric pump in real time to achieve precise control of the flexion and extension angle.

[0069] As Figure 4 shown, the electroacupuncture stimulation mechanism 8 includes a servo motor 85 fixed on the ulnar and radial deviation mechanism 5 near the palm side. A connecting plate 84 is fixedly connected to the output shaft of the servo motor 85. An electric push rod 82 is fixed on the connecting plate 84. The other end of the electric push rod 82 is fixedly connected to a linear motor 86. The extending end of the linear motor 86 is fixedly connected to a connecting plate a87. The linear motor 86 moves in the direction close to the inner side of the palm and away from the inner side of the palm. A sleeve 80 is fixedly connected to the side of the connecting plate a87 close to the palm center. One end of a spring a81 is fixed to the inner bottom of the sleeve 80. The other end of the spring a81 is fixedly connected to an electroacupuncture needle 79. The electroacupuncture head of the electroacupuncture needle 79 uses a spherical contact and its surface is silver-plated. The servo motor 85, the electric push rod 82, the electroacupuncture needle 79, and the linear motor 86 are all electrically connected to a control unit through wires. The visual recognition device includes a binocular camera 83 and a visual recognition module arranged on the ulnar and radial deviation mechanism 5. The binocular camera 83 faces the palm center side. The binocular camera 83 is electrically connected to the visual recognition module through a wire. The visual recognition module is connected to the control unit through a wire. When the electroacupuncture needle 79 contacts the skin, the spring a81 automatically compresses according to the skin curvature to ensure that the needle tip is vertically attached to the acupoint surface.

[0070] Embodiment 2

[0071] On the basis of Embodiment 1, as Figure 5 shown, the elbow adjustment mechanism 2 includes a bracket 11 installed at one end of the base 1. A horizontally arranged bracket a9 is fixedly arranged on the bracket 11. Two mutually parallel guide rails 12 are arranged on the bracket a9. A slider 13 is slidably connected to the guide rails 12. An adjustment box 18 is fixed on the slider 13. A transmission mechanism is arranged in the adjustment box 18. An elbow bracket 14 is connected to the transmission mechanism. An elbow pad 15 is arranged on the upper surface of the elbow bracket 14. The transmission mechanism includes a height adjustment shaft rotatably installed in the adjustment box 18. One end of the height adjustment shaft extends out of the adjustment box 18 and is connected to a height adjustment knob 19. A gear 17 is fixedly installed at the position of the height adjustment shaft located inside the adjustment box 18. The gear 17 meshes with a vertically arranged rack 16. The elbow bracket 14 is fixed to the upper end of the rack 16. A vertically distributed chute is also arranged on the inner wall of the adjustment box 18. A slider a13 is slidably connected in the chute. The upper end of the slider a13 is connected to the bottom of the elbow bracket 14. An installation bracket 11-1 is fixedly arranged at the end of the bracket 11 away from its installation on the base 1. The pronation and supination mechanism 3 is installed on the installation bracket 11-1;

[0072] As Figure 6As shown in the figure, the pronation and supination mechanism 3 includes an external gear type slewing bearing. The inner ring 25 of the external gear type slewing bearing is fixed on the mounting bracket 11-1. The external gear ring 24 of the external gear type slewing bearing meshes with gear a20. A motor 21 is also fixed on the bracket a9 through a motor support 22. Gear a20 is fixedly connected to a gear shaft. The gear shaft is fixedly connected to the output shaft of the motor 26 through a coupling 23. The wrist fixing mechanism 4 is fixed on the side of the external gear ring 24 away from the elbow bracket 14. The sliding direction of the slider 13 on the guide rail 12 is parallel to the axis of the external gear type slewing bearing.

[0073] As Figure 7 shown in the figure, the wrist fixing mechanism 4 includes a lower support adjusting frame 29 fixed on the side of the external gear ring 24 away from the elbow bracket 14. The lower support adjusting frame 29 is horizontally arranged. A pedestal bearing 28 is fixedly arranged on the upper surface of the lower support adjusting frame 29. A vertically arranged lead screw 30 is inserted through the pedestal bearing 28. The lower end of the lead screw 30 passes through the lower support adjusting frame 29 and is connected to a height adjusting knob a26 through a coupling a27. The upper end of the lead screw 30 is fixedly connected to a horizontally arranged lower bracket 33 through a lead screw nut pair. Guide rails a43 vertically arranged are also fixedly connected to the positions on both sides of the lower bracket 33 on the lower support adjusting frame 29. Sliders b44 are fixedly connected to the positions on both sides of the lower bracket 33 corresponding to the guide rails a43. The sliders b44 are slidably connected to the corresponding guide rails a43 on one side. Left brackets 34 and right brackets 39 are respectively fixed on both sides of the lower bracket 33. An upper bracket 36 is fixedly connected to the left bracket 34. The upper bracket 36 is in an inverted "L" shape. A lower cushion 45, an upper cushion 37 and a right cushion 38 are respectively arranged on the upper surface of the lower bracket 33, the lower surface of the upper bracket 36 and the left side surface of the right bracket 39. A plurality of springs 40 are evenly arranged below the right cushion 38 on the side of the right bracket 39 close to the left bracket 34. The ulnar and radial deviation mechanism 5 is fixed on the lower support adjusting frame 29.

[0074] A horizontally arranged lead screw 31 passes through the lower bracket 33. The left bracket 34 and the right bracket 39 are fixed on the lead screws 31 extending out of both sides of the lower bracket 33 through nuts 32. A plurality of adjusting holes are evenly arranged on the left bracket 34 from top to bottom. A plurality of adjusting holes are also arranged on the lower end of the upper bracket 36 from top to bottom. The left bracket 34 and the upper bracket 36 are fixed through pins 35 passing through the corresponding adjusting holes. One end of the lead screw 31 is also connected to a wrist joint clamping adjusting knob 41 through a coupling b42.

[0075] As Figure 8As shown, the ulna-radial deflection mechanism 5 includes a left lower bracket 46 and a right lower bracket 58 fixed at both ends of the upper surface of the lower support adjustment frame 29, the upper ends of the left lower bracket 46 and the right lower bracket 58 are respectively connected to the left middle bracket 47 and the right middle bracket 57, the upper end of the left middle bracket 47 is rotatably connected to the left upper bracket 53 through a transmission shaft 51, and the upper end of the right middle bracket 57 is rotatably connected to the right upper bracket 56 through a rotation shaft 51-1, and the other ends of the left upper bracket 53 and the right upper bracket 56 are commonly fixedly connected to a connecting plate 54, and the upper end of the left middle bracket 47 is connected to the left upper bracket 53 through a transmission shaft 51-1. The motor support a48 is installed with a driving motor 49, and a bevel gear 50 is installed on the output shaft of the driving motor 49. A bevel gear a52 is installed on the transmission shaft 51 installed on the left middle bracket 47 and the left upper bracket 53. The bevel gear a52 is meshed with the bevel gear 50. The axial direction of the output shaft of the driving motor 49 is parallel to the axial direction of the external gear slewing support bearing. The flexion and extension mechanism 6 is connected to the connecting plate 54; the binocular camera 83 is fixed on the upper end of the right middle bracket 57; the steering gear 85 is installed on the side of the right middle bracket 57 close to the palm;

[0076] The connecting plate 54 is provided with a linear guide rail 55, and the connecting plate 54 is also provided with a waist-shaped slide groove 63, and the extension direction of the slide groove 63 is the same as the extension direction of the linear guide rail 55. Figure 9 As shown, the flexion and extension mechanism 6 includes a bracket a65, a slider c67 is arranged below the bracket a65 corresponding to the linear guide rail 55, the bracket a65 and the connecting plate 54 are slidably connected through the slider c67 and the linear guide rail 55, a transmission shaft b68 is rotatably installed on the bracket a65, one end of the transmission shaft b68 extends out of the upper surface of the bracket a65 and is installed with a synchronous pulley 69, and the other end of the transmission shaft b68 extends out of the lower surface of the bracket a65 and the slide groove 63 on the connecting plate 54 is connected to the flexion and extension connecting rod 61, and a slide groove is provided on the connecting plate 54. When the bracket a65 slides on the linear guide rail 55, the transmission shaft b68 slides in the slide groove, and the flexion and extension connecting rod 61 extends toward the side away from the external gear slewing support bearing and a guide rail b62 is provided on the upper surface of the flexion and extension connecting rod 61, a connecting slider 60 is provided on the guide rail b62, a handle 59 is fixedly connected below the slider 60, a flexion and extension motor 72 is provided on the lower surface of the end of the bracket a65 away from the transmission shaft b68, the output shaft of the flexion and extension motor 72 passes through the bracket a65 and is connected to the synchronous pulley a71, the synchronous pulley 69 and the synchronous pulley a71 are connected through the synchronous belt 70, a plurality of pin holes are evenly arranged on the connecting plate 54 along the sliding direction of the bracket a65, and the bracket a65 and the connecting plate 54 are fixed by a latch c66 passing through the corresponding pin holes.

[0077] In this embodiment, the working principles of the elbow adjustment mechanism 2, the pronation and supination mechanism 3, the wrist fixing mechanism 4, the ulnar and radial deviation mechanism 5, and the flexion and extension mechanism 6 are as follows:

[0078] Due to the varying thicknesses and lengths of human arms, the elbow adjustment mechanism 2 and the wrist fixation mechanism 4 are required to jointly adjust the axis of the forearm so that it coincides with the axis of pronation and supination movement of the wrist joint. According to the distance between the elbow and wrist of the tester, the horizontal adjustment of the position of the elbow bracket can be achieved by sliding the slider 10 on the guide rail 12, thereby adjusting the position of the adjustment box 18 on the bracket a9, and then adjusting the distance between the elbow bracket 14 and the pronation and supination mechanism 3. Then, the vertical adjustment and fixation are achieved by rotating the height adjustment knob 19. Specifically: rotating the height adjustment knob 19 drives the gear 17 to rotate. Through the meshing transmission between the gear 17 and the rack 16, the gear 17 drives the rack fixedly connected to the elbow bracket 14 to perform vertical displacement. At the same time, the slider a13 is slidably connected to the inner wall slide rail of the adjustment box 18 to achieve the vertical adjustment of the elbow bracket 14.

[0079] The elbow pad 15 is used to support the user's elbow and is a soft pad made of rubber.

[0080] When the motor 21 rotates, due to the meshing and interaction relationship between the gear a20 and the external gear ring 24, the entire external gear ring 24 and the wrist fixation mechanism fixedly connected thereto are driven to rotate around its axis, realizing the function of pronation and supination movement.

[0081] The wrist fixation mechanism 4 comfortably fixes the front end of the user's forearm to the robot, thereby avoiding the misalignment between the axis of pronation and supination movement of the human wrist and the axis of the robot due to relative displacement. By rotating the clamping adjustment knob 41 and the nut 39, the distance between the adjustment left bracket 34 and the right bracket 39 is adjusted. For the convenience of the user to place the wrist on this mechanism, the left bracket 34 and the upper bracket 36 are fixed by inserting the pin 35 through the corresponding adjustment holes, and the position of the upper bracket 36 is adjusted and fixed up and down, so as to jointly fix the user's wrist joint with other brackets. By rotating the height adjustment knob a26, and then driving the coupling a27 and the lead screw 30 to adjust, the lower bracket 33 drives the lower cushion 45 to move up and down along the guide rail a43, and then its position is fixed by the lead screw nut pair. By cooperating with the elbow adjustment mechanism 2, the height position of the user's forearm is adjusted so that the rotation axis of the pronation and supination movement of the human forearm is aligned with the circumferential movement axis of the pronation and supination mechanism of the robot. The wrist fixation mechanism 4 is fixedly connected to the external gear ring 24 of the pronation and supination mechanism 3 through the lower support adjustment frame 29. The lower cushion 45, the upper cushion 37 and the right cushion 38 are all soft pads made of rubber, so the human arm will contact the rehabilitation robot with a more comfortable touch.

[0082] For the ulnar and radial deviation movements of the wrist joint, they are realized by the ulnar and radial deviation mechanism 5. The lower left bracket 46 and the lower right bracket 58 are connected to the lower support adjustment bracket 29, playing the role of supporting the ulnar and radial deviation mechanism 5. Since the upper ends of the middle left bracket 47 and the middle right bracket 57 are respectively rotationally connected to the upper left bracket 53 and the upper right bracket 56 through the transmission shaft 51 and the rotating shaft 51-1, and the other ends of the upper left bracket 53 and the upper right bracket 56 are fixedly connected together with the connecting plate 54, the connecting plate 54 can swing on the middle left bracket 47 and the middle right bracket 57 around the ulnar and radial deviation movement axis. The bevel gear 50 is installed on the output shaft of the driving motor 49. The bevel gear a52 is installed on the transmission shaft 51, and the two bevel gears are meshed for transmission. When the driving motor 49 rotates, its output torque is reversed through the bevel gear system, and finally drives the upper left bracket 53 to perform a swinging movement around the ulnar and radial deviation movement axis, realizing the ulnar and radial deviation movement function.

[0083] Considering the anatomical feature of the human wrist joint that the flexion and extension movement axis and the ulnar and radial deviation movement axis do not coincide, the cooperation mode between the flexion and extension mechanism 6 and the connecting plate 54 is not fixedly connected, but movable. The bracket a65 on the flexion and extension mechanism 6 can slide on the linear guide 55 on the connecting plate 54 of the ulnar and radial deviation mechanism 5, and its position is fixed by the pin c66 to adjust the distance between the human flexion and extension movement axis and the ulnar and radial deviation movement axis. When the output shaft of the flexion and extension motor 72 rotates, the transmission shaft b68 is driven through the synchronous pulley-synchronous belt transmission mechanism, thereby driving the flexion and extension connecting rod 61 to rotate around the flexion and extension transmission shaft b68, realizing the flexion and extension movement function.

[0084] The connecting slider 60 is used to install the handle 59. Since the glove is independent, after the patient holds the handle, the five fingers curl up in a semi-helical wrapped shape, making the palm fit completely with the curved surface of the handle, and then the wrist joint rehabilitation training can be carried out.

[0085] Embodiment 3

[0086] On the basis of Embodiment 2, all the motors in the elbow joint adjustment mechanism 2, the pronation and supination mechanism 3, the wrist fixation mechanism 4, the ulnar and radial deviation mechanism 5, and the flexion and extension mechanism 6 are controlled by the control unit.

[0087] The motor 21 includes a reducer, a torque sensor and an encoder. The output shaft of the motor 21 is connected to the gear a25, and the gear a20 meshes with the external gear ring 24 for transmission. The motor 21 drives the external gear ring 24 to perform a rotary motion. Therefore, the torque sensor and the encoder in the motor 21 can measure the motion angle and torque during the pronation and supination movement of the human wrist joint and feedback them to the control unit in real time.

[0088] The drive motor 49 includes a speed reducer, a torque sensor, and an encoder. The output shaft of the drive motor 49 is connected to the ulnar and radial deviation drive shaft 51 through a bevel gear commutation system. The ulnar and radial deviation movement can be achieved through the drive motor 49. Therefore, the torque sensor and encoder in the drive motor 49 can measure the movement angle and torque when the human wrist joint makes ulnar and radial deviations, and feedback them to the control unit in real time.

[0089] The flexion and extension motor 72 includes a speed reducer, a torque sensor, and an encoder. The output shaft of the flexion and extension motor 72 is connected to the flexion and extension link 61 through a synchronous belt. The flexion and extension movement of the wrist joint can be achieved through the flexion and extension motor 72.

[0090] The present invention realizes the pronation and supination movement, ulnar and radial deviation movement, flexion and extension movement of the wrist joint, and flexion and extension movement of the finger joints. It includes seven adjustable degrees of freedom, realizes the adjustment of the mechanical structure size according to the arm sizes of different individuals, and compensates for the function of misalignment of the human-machine matching joints. The ergonomic adaptation mechanism (elbow adjustment mechanism 2, wrist fixation mechanism 4), multi-degree-of-freedom movement mechanism (pronation and supination mechanism 3, ulnar and radial deviation mechanism 5, flexion and extension mechanism 6), pneumatic flexion and extension rehabilitation mechanism (finger joint flexion and extension mechanism 7 and electroacupuncture stimulation mechanism 8).

[0091] Before any rehabilitation training, the user needs to wash their hands, sit comfortably in front of the rehabilitation robot, and place their hand in the camera shooting area.

[0092] Embodiment 4

[0093] The wrist and finger joint rehabilitation training method of the present invention based on dynamic acupoint visual recognition and electrical collaborative drive adopts the wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive in Embodiment 3, including finger flexion and extension training. The specific process is as follows:

[0094] According to the thickness and length of the user's arm, adjust the elbow adjustment mechanism 2, wrist fixation mechanism 4, pronation and supination mechanism 3, ulnar and radial deviation mechanism 5, and flexion and extension mechanism 6 to suitable positions, so that the user's forearm is fixed on the robot. Set the inflation rate and inflation volume of the electric air pump corresponding to each finger according to the actual situation of the user's fingers, and set the safety threshold of the strain gauge bending sensor corresponding to each finger according to the specific situation of use. Then put the glove body 73 on the user's hand. The control unit controls each electric air pump to inflate at a preset parameter at a certain speed, so that the fingers gradually bend to achieve fist clenching training. The control unit controls each electric air pump to deflate in the reverse direction to achieve extension training. During fist clenching training and extension training, the strain gauge bending sensor monitors the finger bending degree in real time. If it exceeds the safety threshold, the corresponding electric air pump immediately stops working and alarms.

[0095] Embodiment 5

[0096] Based on Embodiment 4, it further includes electrical stimulation rehabilitation, specifically:

[0097] According to the thickness and length of the user's arm, the elbow adjustment mechanism 2, wrist fixation mechanism 4, pronation and supination mechanism 3, ulnar and radial deviation mechanism 5, and flexion and extension mechanism 6 are adjusted to suitable positions, so that the user's forearm is fixed to the robot. Then, the servo motor 85 rotates to drive the electric push rod 82 to rotate. Combining the extension and contraction range of the electric push rod 82, the electro-acupuncture needle 79 can stimulate all acupoints on the palm. Then, the binocular camera 83 collects the image of the inner side of the palm and uploads it to the visual recognition module. The visual recognition module identifies the positions of the key points of the hand, then calculates the corresponding acupoints according to the positions of the key points of the hand, and then feeds back the corresponding acupoint coordinates to the control unit. The control unit controls the movement of the servo motor 85 and the electric push rod 82, moves the electro-acupuncture needle 79 to the corresponding acupoint position, and then controls the linear motor 86 to move towards the palm side, so that the electro-acupuncture needle 79 needles at the corresponding acupoint, and then controls the electro-acupuncture needle 79 to discharge for electrical stimulation.

[0098] A strain gauge is arranged on the spherical contact head at the end of the electro-acupuncture needle 79 to monitor the contact pressure in real time. The pressure in the treatment window is 0.2 - 0.8N. If the pressure > 1N, the control unit controls the linear motor 86 to retract the electro-acupuncture needle 79 by 0.5mm to prevent overpressure damage. If the pressure < 0.2N, it is considered that the stimulation is not effective, and the control unit needs to continue to control the linear motor 86 to push the electro-acupuncture needle 79 forward.

[0099] When performing electrical stimulation rehabilitation training, a safety threshold for the output current of the electro-acupuncture needle 79 is set on the control unit. An electromyogram signal collector is installed on the user's forearm. The electromyogram signal collector collects the electromyogram signals of the target muscle in real time during electrical stimulation, calculates the integrated electromyogram value iEMG and the root mean square value RMS to quantify the muscle activation degree, and dynamically adjusts the output current of the electro-acupuncture needle. Specifically: when the electromyogram amplitude < 50μV, it is considered that the stimulation is insufficient, and the control unit controls the output current of the electro-acupuncture needle 79 to increase by 10%; when the electromyogram amplitude > 150μV, it is considered that the activation is excessive, and the output current of the electro-acupuncture needle 79 is decreased to the safety threshold to avoid muscle fatigue.

[0100] The visual recognition module identifies the positions of the key points of the hand, and then the specific process of calculating the corresponding acupoints according to the positions of the key points of the hand is as follows:

[0101] The visual recognition module detects the key points of the hand using the MediaPipe Hands model based on the collected image, identifies 21 key points of the hand, and obtains the normalized coordinates of the 21 key points.

[0102] Calculate the positions of the corresponding key points in the pixel coordinate system according to the normalized coordinates. Specifically:

[0103] x 像素 = x归一化 × w; y 像素 = y 归一化 × h

[0104] where x 归一化 and y 归一化 are respectively the horizontal position of the corresponding key point in the image and the vertical position of the key point in the image. Among them, x 归一化 increases from left to right, and y 归一化 increases from top to bottom; x 像素 and y 像素 are respectively the coordinate values in the pixel coordinate system after the conversion of x 归一化 and y 归一化 ; w and h are respectively the pixel sizes of the corresponding width and height under the images collected by the binocular camera;

[0105] Next, boundary detection is performed. The pixel coordinates are limited within a reasonable range using min(max(...)), specifically:

[0106] min(max(x 像素 , w - 1)); min(max(y 像素 , h - 1))

[0107] That is, x 像素 is limited within the range of [0, w - 1], and y 像素 is limited within the range of [0, h - 1];

[0108] Then, according to the pixel coordinates of the corresponding key points after boundary detection, the pixel coordinates of the Hegu acupoint, Laogong acupoint, Dazhui acupoint, Shaoshang acupoint, and Zhongzhu acupoint are calculated. Specifically:

[0109] Hegu acupoint: According to the Euclidean distance d between the two points of the thumb tip THUMB_TIP and the index finger tip INDEX_FINGER_TIP, if d ≥ 0.03, the user is prompted to adjust the posture, move the hand slightly away from the camera, and then recalculate until d < 0.03; when d < 0.03, the pixel coordinates of the Hegu acupoint are calculated in the following way:

[0110]

[0111] where x THUMB_TIP and y THUMB_TIP are respectively the pixel coordinates corresponding to the conversion of the horizontal position and vertical position coordinates of the thumb tip THUMB_TIP in the image, x INDEX_TIP and y INDEX_TIP are respectively the pixel coordinates corresponding to the conversion of the horizontal position and vertical position coordinates of the index finger tip INDEX_FINGER_TIP in the image, x 合谷 and y合谷 They are the pixel coordinates of Hegu acupoint in the horizontal and vertical directions respectively;

[0112] Laogong acupoint:

[0113] Among them, x 劳宫 and y 劳宫 They are the pixel coordinates of Laogong acupoint in the horizontal and vertical directions respectively; and They are the pixel coordinates corresponding to the horizontal and vertical position coordinates of the INDEX_FINGER_MCP of the index finger after conversion in the image. x MIDDLE_MCP and y MIDDLE_MCP They are the pixel coordinates corresponding to the horizontal and vertical position coordinates of the MIDDLE_FINGER_MCP of the middle finger after conversion in the image;

[0114] Dazhui acupoint: x 大椎 = x PINKY_MCP ; y 大椎 = y PINKY_MCP

[0115] Among them, x 大椎 and y 大椎 They are the pixel coordinates of Dazhui acupoint in the horizontal and vertical directions respectively; x PINKY_MCP and y PINKY_MCP They are the pixel coordinates corresponding to the horizontal and vertical position coordinates of the PINKY_MCP of the little finger after conversion in the image;

[0116] Shaoshang acupoint: x 少商 = x INDEX_TIP ; y 少商 = y INDEX_TIP

[0117] Among them, x 少商 and y 少商 They are the pixel coordinates of Shaoshang acupoint in the horizontal and vertical directions respectively;

[0118] Zhuzhu acupoint:

[0119] Among them, x 中渚 and y 中渚 They are the pixel coordinates of Zhuzhu acupoint in the horizontal and vertical directions respectively; x PINKY_MCP and y PINKY_MCP They are the pixel coordinates corresponding to the horizontal and vertical position coordinates of the PINKY_MCP of the little finger after conversion in the image; x RING_MCP and y RING_MCPThey are the pixel coordinates corresponding to the horizontal and vertical position coordinates of the metacarpophalangeal joint of the ring finger (RING_FINGER_MCP) after coordinate conversion in the image. Then, the pixel coordinates of the acupoints Hegu, Laogong, Dazhui, Shaoshang, and Zhongzhu are converted into coordinates in the mechanical coordinate system, and then the rotation angles of the driving servos and the telescopic amounts of the electric push rods are calculated through inverse kinematics and sent to the control unit.

[0120] Embodiment 6

[0121] On the basis of Embodiment 5, it further includes the training of the wrist and finger joint movement functions, specifically: according to the thickness and length of the user's arm, the axis of the forearm is jointly adjusted by the elbow adjustment mechanism 2 and the wrist fixing mechanism 4 to make it coincide with the pronation and supination movement axis of the wrist joint, the distance between the elbow bracket 14 and the pronation and supination mechanism 3 is adjusted, and then the user's forearm is fixed to the robot by the wrist fixing mechanism 4, and the ulnar and radial deviation movement of the wrist joint is realized through the ulnar and radial deviation mechanism 5, and the flexion and extension movement functions are realized through the flexion and extension mechanism 6. During the movement process, the output data of the relevant muscles during the wrist joint movement are collected by the electromyogram signal collector and uploaded to the control unit.

[0122] Embodiment 7

[0123] On the basis of Embodiment 5, after normalizing the coordinates in the present invention, after converting the normalized coordinates into pixel coordinates, boundary checking is performed to prevent accessing pixel coordinates outside the image size range, ensuring that the pixel coordinates are within the ranges of [0, width - 1] and [0, height - 1]. In this way, when drawing points and marking acupoints, coordinates outside the image will never be accessed. This can prevent the calculated coordinates from exceeding the boundaries of the image. During the subsequent calculation of acupoints, the function will limit them within the image size, avoiding the situation where acupoint markings exceed the image due to calculation errors. This processing reduces the display problems caused by image size mismatches, making the acupoint display always within the image. A display screen is set in the control unit. After obtaining the coordinates of the acupoints, text and color markings are given to each acupoint for display, and each acupoint is distinguished by color, which helps with visual verification. Through intuitive markings, errors can be further adjusted and verified manually to ensure the correctness of the position of each acupoint.

[0124] Embodiment 8

[0125] Based on Embodiment 5, the present invention further sets up dynamic tracking compensation. The binocular camera 83 captures hand images. When the hand moves, the coordinates of the key acupoints can be updated in a timely manner. The vision system updates the acupoint coordinates in real time and sends a correction instruction to the control unit of the electro-acupuncture through the UART protocol (baud rate 115200bps), with a response delay ≤ 50ms. The binocular camera 83 of the present invention uses a global shutter industrial camera with a resolution ≥ 720P (frame rate ≥ 30fps), equipped with a ring-shaped fill light and a polarization filter to effectively eliminate ambient light interference and ensure clear and stable hand images. The vision recognition module of the present invention serves as a computing unit, equipped with an NVIDIA Jetson Nano or an equivalent embedded GPU module, and transmits image data to the electro-acupuncture control unit through USB3.0 or Gigabit Ethernet.

Claims

1. A wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive, characterized in that It includes a base (1) and a knuckle flexion and extension mechanism (7). An elbow adjustment mechanism (2) is fixedly arranged on the base (1). A pronation and supination mechanism (3) is fixed on the elbow adjustment mechanism (2). A wrist fixation mechanism (4) is fixed on the pronation and supination mechanism (3). A ulnar and radial deviation mechanism (5) is fixed on the wrist fixation mechanism (4). A flexion and extension mechanism (6) is fixed on the ulnar and radial deviation mechanism (5). An electroacupuncture stimulation mechanism (8) is also fixed on the ulnar and radial deviation mechanism (5). A visual recognition device is also arranged on the ulnar and radial deviation mechanism (5). The knuckle flexion and extension mechanism (7) is also connected to an electric air pump. A strain gauge bending sensor is also arranged on the knuckle flexion and extension mechanism (7). An electromyogram signal collector arranged on the user's forearm, the strain gauge bending sensor and the electromyogram signal collector are electrically connected to a data acquisition card through wires. The data acquisition card, the visual recognition device, and the electric air pump are respectively electrically connected to a control unit through wires.

2. The wrist and finger joint rehabilitation robot based on dynamic acupoint vision recognition and electrical collaborative drive according to claim 1, characterized in that The knuckle flexion and extension mechanism (7) is a bellows-type pneumatic rehabilitation glove, including a glove body (73). Actuators are arranged at the positions corresponding to each finger on the back of the glove body (73). Each actuator is adapted to the length of the corresponding finger. The actuator includes a flexible restraint layer (77) arranged on the back of each finger. The flexible restraint layer (77) is distributed along the length of the finger, and a plurality of fixed connectors (76) are evenly arranged on the flexible restraint layer (77). A section of bellows (75) is connected between two adjacent fixed connectors (76). Air channels (78) are arranged inside the plurality of fixed connectors (76). The two ends of the air channel (78) are connected to the bellows (75) connected to the two ends of the fixed connector (76). One end of the fixed connector (76) farthest from the fingertip and facing away from the fingertip is also connected to an intake pipe (74). The intake pipe (74) is connected to one end of the air channel (76) of the corresponding fixed connector (76). The intake pipe (74) is connected to the electric air pump through a pipeline. Each electric air pump is electrically connected to the control unit through a wire. The position of the flexible restraint layer (77) corresponding to each bellows (75) is set as a square wave-like structure. The strain gauge bending sensor is embedded between the flexible restraint layer (77) corresponding to each bellows (75) and the glove body (73).

3. The wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive according to claim 2, characterized in that, The electroacupuncture stimulation mechanism (8) comprises a steering gear (85) fixed on the radial-ulnar deflection mechanism (5) close to the palm side, the output shaft of the steering gear (85) is fixedly connected to a connecting plate (84), the connecting plate (84) is fixed to an electric push rod (82), the other end of the electric push rod (82) is fixed to a linear motor (86), the extended end of the linear motor (86) is fixedly connected to a connecting plate a (87), the linear motor (86) moves in a direction close to the inner side of the palm and away from the inner side of the palm, the connecting plate a (87) is fixedly connected to a sleeve (80) on the side close to the palm center, the inner bottom of the sleeve (80) is fixed to a One end of a spring a (81), the other end of the spring a (81) is fixedly connected to an electric needle (79), the electric needle head of the electric needle (79) adopts a spherical contact and the surface is silver-plated, the steering gear (85), the electric push rod (82), the electric needle (79) and the linear motor (86) are all electrically connected to a control unit through a wire, and the visual recognition device comprises a binocular camera (83) and a visual recognition module arranged on the ulnar radial deflection mechanism (5), the binocular camera (83) faces the side of the palm, the binocular camera (83) is electrically connected to the visual recognition module through a wire, and the visual recognition module is connected to the control unit through a wire.

4. The wrist and finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical collaborative drive according to claim 3, wherein The elbow adjustment mechanism (2) comprises a bracket (11) mounted on one end of the base (1), the bracket (11) being provided with a bracket a (9) fixedly mounted horizontally, the bracket a (9) being provided with two mutually parallel guide rails (12), the guide rails (12) being slidably connected with a slider (13), an adjustment box (18) being fixedly mounted on the slider (13), a transmission mechanism being arranged in the adjustment box (18), an elbow bracket (14) being connected to the transmission mechanism, an elbow support pad (15) being arranged on the upper surface of the elbow bracket (14), the transmission mechanism comprising a height adjustment shaft rotatably mounted in the adjustment box (18), one end of the height adjustment shaft extending out of the adjustment box (18) 18) and connected with a height adjustment knob (19); a gear (17) is fixedly installed at a position of the height adjustment shaft located in the adjustment box (18); the gear (17) is meshed with a vertically arranged rack (16); the elbow bracket (14) is fixed to the upper end of the rack (16); a vertically distributed slide groove is also arranged on the inner wall of the adjustment box (18); a slider a (13) is slidably connected in the slide groove; the upper end of the slider a (13) is connected to the bottom of the elbow bracket (14); a mounting bracket (11-1) is fixedly arranged on the end of the bracket (11) away from the mounting of the bracket on the base (1); the pronation and supination mechanism (3) is installed on the mounting bracket (11-1); The pronation and supination mechanism (3) includes an externally toothed slewing bearing. The inner ring (25) of the externally toothed slewing bearing is fixed on the mounting bracket (11-1). The external gear ring (24) of the externally toothed slewing bearing meshes with a gear a (20). A motor (21) is also fixed on the bracket a (9) through a motor support (22). The gear a (20) is fixedly connected to a gear shaft, and the gear shaft is fixedly connected to the output shaft of the motor (21) through a coupling (23). The wrist fixing mechanism (4) is fixed on the side of the external gear ring (24) away from the elbow bracket (14). The sliding direction of the slider (13) on the guide rail (12) is parallel to the axis of the externally toothed slewing bearing. The wrist fixing mechanism (4) includes a lower support adjusting frame (29) fixed on the side of the external gear ring (24) away from the elbow bracket (14). The lower support adjusting frame (29) is horizontally arranged. A pedestal bearing (28) is fixedly arranged on the upper surface of the lower support adjusting frame (29). A vertically arranged lead screw (30) is penetrated in the pedestal bearing (28). The lower end of the lead screw (30) passes through the lower support adjusting frame (29) and is connected with a height adjusting knob a (26) through a coupling a (27). The upper end of the lead screw (30) is fixedly connected with a horizontally arranged lower bracket (33) through a lead screw nut pair. Guide rails a (43) arranged vertically are also fixedly connected at positions on both sides of the lower bracket (33) on the lower support adjusting frame (29). Sliders b (44) are fixedly connected at positions corresponding to the guide rails a (43) on both sides of the lower bracket (33). The sliders b (44) are slidably connected with the corresponding guide rails a (43). Left brackets (34) and right brackets (39) are respectively fixed on both sides of the lower bracket (33). An upper bracket (36) is fixedly connected to the left bracket (34). The upper bracket (36) is arranged in an inverted "L" shape. A lower cushion (45), an upper cushion (37) and a right cushion (38) are respectively arranged on the upper surface of the lower bracket (33), the lower surface of the upper bracket (36) and the left side surface of the right bracket (39). A plurality of springs (40) are evenly arranged below the right cushion (38) on the side of the right bracket (39) close to the left bracket (34). The radioulnar deviation mechanism (5) is fixed on the lower support adjusting frame (29). A horizontally arranged lead screw (31) penetrates through the lower bracket (33). The left bracket (34) and the right bracket (39) are fixed on the lead screws (31) extending out of both sides of the lower bracket (33) through nuts (32). A plurality of adjusting holes are evenly arranged from top to bottom on the left bracket (34). A plurality of adjusting holes are also arranged from top to bottom at the lower end of the upper bracket (36). The left bracket (34) and the upper bracket (36) are fixed by inserting pins (35) through the corresponding adjusting holes. One end of the lead screw (31) is also connected with a wrist joint clamping adjusting knob (41) through a coupling b (42). The ulnar and radial deviation mechanism (5) includes a lower left support (46) and a lower right support (58) fixed at both ends of the upper surface of the lower support adjusting frame (29). The upper ends of the lower left support (46) and the lower right support (58) are respectively connected with a middle left support (47) and a middle right support (57). The upper end of the middle left support (47) is rotatably connected with an upper left support (53) through a transmission shaft (51). The upper end of the middle right support (57) is rotatably connected with an upper right support (56) through a rotating shaft (51-1). The other ends of the upper left support (53) and the upper right support (56) are fixedly connected with a connecting plate (54). A driving motor (49) is installed on the middle left support (47) through a motor support a (48). A bevel gear (50) is installed on the output shaft of the driving motor (49). A bevel gear a (52) is installed on the transmission shaft (51) installed on the middle left support (47) and the upper left support (53). The bevel gear a (52) meshes with the bevel gear (50). The axial direction of the output shaft of the driving motor (49) is parallel to the axial direction of the external tooth type slewing support bearing. The flexion and extension mechanism (6) is connected to the connecting plate (54); The binocular camera (83) is fixed at the upper end of the middle right support (57); The steering gear (85) is installed on one side of the middle right support (57) close to the palm center; The connecting plate (54) is provided with a linear guide rail (55), and the connecting plate (54) is also provided with a waist-shaped slide groove (63), and the extension direction of the slide groove (63) is the same as the extension direction of the linear guide rail (55). The flexion and extension mechanism (6) includes a bracket a (65), and a slider c (67) is provided below the bracket a (65) and corresponding to the linear guide rail (55). The bracket a (65) and the connecting plate (54) are connected by the slider c (67). 7) and the linear guide rail (55) are slidably connected, a transmission shaft b (68) is rotatably installed on the bracket a (65), one end of the transmission shaft b (68) extends out of the upper surface of the bracket a (65) and is installed with a synchronous pulley (69), the other end of the transmission shaft b (68) sequentially extends out of the lower surface of the bracket a (65), the slide groove (63) on the connecting plate (54) is connected to the flexion and extension connecting rod (61), the connecting plate (54) is provided with a slide groove, the bracket When the bracket a (65) slides on the linear guide rail (55), the transmission shaft b (68) slides in the slide groove, the flexion-extension link (61) extends toward the side away from the external gear slewing bearing, and the upper surface of the flexion-extension link (61) is provided with a guide rail b (62), a connecting slider (60) is provided on the guide rail b (62), and a handle (59) is fixedly connected to the lower side of the slider (60), and the lower surface of the bracket a (65) away from the transmission shaft b (68) is provided with a connecting slider (60). A flexion-extension motor (72) is provided, the output shaft of the flexion-extension motor (72) passes through the bracket a (65) and is connected to a synchronous pulley a (71), the synchronous pulley (69) and the synchronous pulley a (71) are connected by a synchronous belt (70), a plurality of pin holes are evenly arranged on the connecting plate (54) along the sliding direction of the bracket a (65), and the bracket a (65) and the connecting plate (54) are fixed by a latch pin c (66) passing through the corresponding pin holes.

5. A wrist and finger joint rehabilitation training method based on dynamic acupoint visual recognition and electrical collaborative drive, characterized in that, The wrist-finger joint rehabilitation robot based on dynamic acupoint visual recognition and electrical coordinated drive as described in claim 4 includes finger flexion and extension training, and the specific process is as follows: According to the thickness, length and width of the user's arm, the elbow adjustment mechanism (2), the wrist fixing mechanism (4), the pronation and supination mechanism (3), the ulnar radial deviation mechanism (5) and the flexion and extension mechanism (6) are adjusted to suitable positions so that the user's forearm is fixed on the robot, and the inflation rate and inflation amount of the electric air pump corresponding to each finger and the safety threshold of the strain gauge bending sensor corresponding to each finger are set according to the actual situation of the user's fingers, and then the glove body (73) is put on the user's hand, and the control unit controls each electric air pump to inflate at a certain speed according to preset parameters so that the fingers gradually bend to achieve fist training, and the control unit controls each electric air pump to deflate in the opposite direction to achieve stretching training. During fist training and stretching training, the strain gauge bending sensor monitors the degree of finger bending in real time. If the safety threshold is exceeded, the corresponding electric air pump immediately stops working and alarms.

6. The wrist and finger joint rehabilitation training method based on dynamic acupoint visual recognition and electrical collaborative drive according to claim 5, wherein It also includes electrical stimulation rehabilitation, specifically: According to the thickness and length of the user's arm, adjust the elbow adjustment mechanism (2), wrist fixation mechanism (4), pronation and supination mechanism (3), ulnar and radial deviation mechanism (5), and flexion and extension mechanism (6) to suitable positions, so that the user's forearm is fixed to the robot. Then, the servo motor (85) rotates to drive the electric push rod (82) to rotate. Combining the extension and contraction range of the electric push rod (82), the electric acupuncture needle (79) can stimulate all acupoints on the palm. Then, the binocular camera (83) collects the image of the inner side of the palm and uploads it to the visual recognition module. The visual recognition module identifies the positions of the hand key points, then calculates the corresponding acupoints according to the positions of the hand key points, and then feeds back the corresponding acupoint coordinates to the control unit. The control unit controls the movement of the servo motor (85) and the electric push rod (82), moves the electric acupuncture needle (79) to the corresponding acupoint position, and then controls the linear motor (86) to move towards the palm side, so that the electric acupuncture needle (79) needles at the corresponding acupoint, and then controls the electric acupuncture needle (79) to discharge for electrostimulation.

7. The wrist and finger joint rehabilitation training method based on dynamic acupoint visual recognition and electrical collaborative drive according to claim 6, characterized in that A strain gauge is arranged on the spherical contact head at the end of the electric acupuncture needle (79) to monitor the contact pressure in real time. The pressure in the treatment window is 0.2 - 0.8N. If the pressure > 1N, control the linear motor (86) to retract the electric acupuncture needle (79) by 0.5mm to prevent overpressure damage. If the pressure < 0.2N, it is considered that the stimulation is not effective, and the linear motor (86) needs to be continuously controlled to push the electric acupuncture needle (79).

8. The wrist and finger joint rehabilitation training method based on dynamic acupoint visual recognition and electrical collaborative drive according to claim 6, wherein When performing electrostimulation rehabilitation training, set the safety threshold of the output current of the electric acupuncture needle (79) on the control unit. An electromyogram signal collector is installed on the user's forearm. The electromyogram signal collector collects the electromyogram signal of the target muscle in real time during electrostimulation, calculates the integral electromyogram value iEMG and the root mean square value RMS, quantifies the muscle activation degree, and dynamically adjusts the output current of the electric acupuncture needle. Specifically: when the electromyogram amplitude < 50μV, it is considered that the stimulation is insufficient, and the output current of the electric acupuncture needle (79) is controlled to increase by 10%; when the electromyogram amplitude > 150μV, it is considered that the activation is excessive, and the output current of the electric acupuncture needle (79) is decreased to the safety threshold to avoid muscle fatigue.

9. The wrist and finger joint rehabilitation training method based on dynamic acupoint visual recognition and electrical collaborative drive according to claim 7, wherein The specific process of the visual recognition module identifying the positions of the hand key points and then calculating the corresponding acupoints according to the positions of the hand key points is as follows: The visual recognition module uses the MediaPipe Hands model to detect the hand key points according to the collected image, identifies 21 hand key points, and obtains the normalized coordinates of the 21 hand key points. Calculate the positions of the corresponding key points in the pixel coordinate system according to the normalized coordinates. Specifically: x pixel = x normalized × w y 像素 = y 归一化 × h where x 归一化 and y 归一化 are respectively the horizontal position of the corresponding key point in the image and the vertical position of the key point in the image, where x 归一化 increases from left to right, and y 归一化 increases from top to bottom; x 像素 and y 像素 are respectively the coordinate values of x 归一化 and y 归一化 in the pixel coordinate system after conversion, and w and h are respectively the pixel sizes of the corresponding width and height under the images collected by the binocular camera; Then perform boundary detection, and use min(max(...)) to limit the pixel coordinates within a reasonable range. Specifically: min(max(x 像素 , w - 1)) min(max(y 像素 ,h - 1)) That is, x 像素 is limited within the range of [0, w - 1], and y 像素 is limited within the range of [0, h - 1]; Then, calculate the acupoint pixel coordinates of Hegu acupoint, Laogong acupoint, Dazhui acupoint, Shaoshang acupoint, and Zhongzhu acupoint according to the pixel coordinates of the corresponding key points after boundary detection. Specifically: Hegu acupoint: According to the Euclidean distance d between the two points of the thumb fingertip THUMB_TIP and the index fingertip INDEX_FINGER_TIP, if d < 0.03, the pixel coordinates of the Hegu acupoint are calculated in the following manner: where x THUMB_TIP and y THUMB_TIP are respectively the pixel coordinates corresponding to the converted horizontal position and vertical position coordinates of the thumb fingertip THUMB_TIP in the image, x INDEX_TIP and y INDEX_TIP are respectively the pixel coordinates corresponding to the converted horizontal position and vertical position coordinates of the index finger fingertip INDEX_FINGER_TIP in the image, x 合谷 and y 合谷 are respectively the pixel coordinates of the Hegu acupoint in the horizontal position and vertical position directions; If d ≥ 0.03, the user is prompted to adjust the posture, move the hand slightly away from the camera, and then recalculate until d < 0.03; Laogong acupoint: where x 劳宫 and y 劳宫 are the pixel coordinates of the Laogong point in the horizontal and vertical position directions, respectively; and are the pixel coordinates corresponding to the conversion of the horizontal and vertical position coordinates of the INDEX_FINGER_MCP of the index finger metacarpophalangeal joint in the image, respectively. x MIDDLE_MCP and y MIDDLE_MCP are the pixel coordinates corresponding to the conversion of the horizontal and vertical position coordinates of the MIDDLE_FINGER_MCP of the middle finger metacarpophalangeal joint in the image, respectively; Dazhui acupoint: x Dazhui = x PINKY_MCP y_Dazhui = y PINKY_MCP where x 大椎 and y 大椎 are the pixel coordinates of the Dazhui acupoint in the horizontal and vertical directions of the pixel coordinates respectively; x PINKY_MCP and y PINKY_MCP are the pixel coordinates corresponding to the horizontal and vertical position coordinates of the PINKY_MCP of the little finger metacarpophalangeal joint after conversion in the image respectively; Shaoshang acupoint: x abscissa = x INDEX_TIP yShaoshang = y INDEX_TIP Where xShaoshang and yShaoshang are the pixel coordinates of the Shaoshang acupoint in the horizontal and vertical position directions, respectively; Zhongzhu acupoint: Among them, x Zhongzhu and y Zhongzhu are the pixel coordinates of the Zhongzhu acupoint in the horizontal and vertical directions respectively; x PINKY_MCP and y PINKY_MCP are the pixel coordinates corresponding to the converted horizontal and vertical position coordinates of the PINKY_MCP of the little finger in the image respectively; x RING_MCP and y RNNG_MCP are the pixel coordinates corresponding to the converted horizontal and vertical position coordinates of the RING_FINGER_MCP of the ring finger in the image respectively; Then, the pixel coordinates of the acupoints of Hegu, Laogong, Dazhui, Shaoshang, and Zhongzhu are converted into coordinates in the mechanical coordinate system, and then the rotation angles of the driving servos and the telescopic amounts of the electric push rods are solved through inverse kinematics and sent to the control unit.

10. The wrist and finger joint rehabilitation training method based on dynamic acupoint visual recognition and electrical collaborative drive according to claim 5, wherein It also includes wrist and finger joint movement function training, specifically: according to the thickness and length of the user's arm, the axis of the forearm is jointly adjusted by the elbow adjustment mechanism (2) and the wrist fixation mechanism (4) to coincide with the pronation and supination movement axis of the wrist joint. The distance between the elbow bracket (14) and the pronation and supination mechanism (3) is adjusted, and then the user's forearm is fixed to the robot through the wrist fixation mechanism (4). The ulnar and radial deviation movement of the wrist joint is realized through the ulnar and radial deviation mechanism (5), and the flexion and extension movement function is realized through the flexion and extension mechanism (6). During the movement, the output data of the relevant muscles during the wrist joint movement are collected by the electromyogram signal collector and uploaded to the control unit.

Images