Massage track planning method of physiotherapy robot
By using robotic arms and visual devices in physiotherapy robots, combined with deep learning models and physiotherapy courses, the problem of low automation of existing physiotherapy equipment is solved, and high-precision massage control is achieved to meet the needs of different patients.
Patent Information
- Application Number
- CN202411321975.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-23
- Publication Date
- 2024-11-19
AI Technical Summary
The existing physical therapy equipment is difficult to meet the specific needs of different patients when used, the degree of automation is low, manual assisted operation is required, and the efficiency is low, making it difficult to achieve theoretical conversion to actual massage control.
The physiotherapy robot includes a robotic arm and a visual device. The human body contour is determined by obtaining images, the acupoints and meridian positions are determined using preset deep learning models, and the robotic arm joint activity stroke is determined in combination with preset physiotherapy courses to achieve high-precision massage.
Automatic massage is realized, which can accurately control different users and needs, and improves the control accuracy of the power, position and speed of massage.
Smart Images

Figure CN118977250A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of physiotherapy equipment, and in particular to a massage trajectory planning method for a physiotherapy robot. Background Art
[0002] With the development of science and technology and people's increasing demand for health, physical therapy technology has received widespread attention. Traditional physical therapy methods mainly rely on manual operation. With the development of technology, it has become possible to use robots for physical therapy.
[0003] When some existing physiotherapy equipment is in use, it can only be operated through the experience and skills of the physiotherapist, which makes it difficult to meet the specific needs of different patients. The physiotherapy equipment has a low degree of automation and requires manual assistance, which has low efficiency and makes it difficult to transform theory into actual massage control. Summary of the invention
[0004] The main purpose of the present invention is to provide a massage trajectory planning method for a physiotherapy robot, aiming to solve the problem that it is difficult to convert the physiotherapy massage theory of traditional Chinese medicine into an executable physiotherapy massage trajectory.
[0005] To achieve the above object, the present invention provides a massage trajectory planning method for a physiotherapy robot, wherein the physiotherapy robot comprises a mechanical arm and a visual device and a massage device disposed at the end of the mechanical arm, and the massage trajectory planning method for the physiotherapy robot comprises: Acquire an image captured by the visual device, and determine a human body contour based on the image; Determining massage positions of acupuncture points and meridians in the human body contour according to a preset deep learning model; Determine the range of motion corresponding to each joint in the robotic arm according to the preset physiotherapy course and the massage position; The mechanical arm is controlled to start moving, and the massage device is controlled to start high-precision massage.
[0006] In one embodiment, the visual device includes a 3D camera and a 2D camera; The step of acquiring the image captured by the visual device and determining the human body contour according to the image comprises: Acquire a 3D image captured by the 3D camera, and assign coordinates to a plurality of pixel points in the 3D image; A 2D image captured by the 2D camera is acquired, and human body edge pixel points among a plurality of the pixel points are identified by an instance segmentation algorithm to obtain the human body contour.
[0007] In one embodiment, before acquiring the image captured by the visual device and determining the human body contour according to the image, the method further includes: Determining the base coordinates of the robotic arm; According to the base coordinates, the massage device is calibrated in a tool coordinate system and a hand-eye calibration; In one embodiment, before acquiring the human body image captured by the visual device and acquiring the human body contour according to the human body image, the method further includes: The mechanical arm is dynamically calibrated.
[0008] In one embodiment, the method for obtaining the preset deep learning model includes: Acquire a training human body contour image, and mark acupoints and meridian positions on the training human body contour image; The preset human body contour image and the corresponding acupuncture points and meridian positions are imported into the initial deep learning model for training to obtain the preset deep learning model.
[0009] In one embodiment, the step of importing the preset human body contour image and the corresponding acupuncture points and meridians into an initial deep learning model for training to obtain the preset deep learning model includes: When the training times reach the preset times, the training is stopped and the model at this time is used as the preset deep learning model; or, When the performance improvement value is less than a preset threshold, the training is stopped and the model at this time is used as the preset deep learning model.
[0010] In one embodiment, the physiotherapy robot further comprises a current loop provided on each joint of the robot arm; The step of controlling the massage device to start massage comprises: The current magnitude on the current loop is obtained, and the position, speed and force parameters of the robot arm are adjusted according to a first preset mapping relationship.
[0011] In one embodiment, the method for obtaining the preset physical therapy data includes: According to the user's needs, the preset physical therapy course is obtained according to the second preset mapping relationship.
[0012] In one embodiment, controlling the robotic arm to start moving includes: The movement of the robotic arm is controlled by force-position hybrid control technology and compliant control technology.
[0013] The present invention provides a massage trajectory planning method for a physiotherapy robot, wherein the physiotherapy robot comprises a mechanical arm and a visual device and a massage device arranged at the end of the mechanical arm, wherein an image is captured by the visual device so as to determine a human body contour according to the image, and after the human body contour is determined, the human body contour is brought into a preset deep learning model, and the massage positions of acupuncture points and meridians in the human body contour are calculated, so as to facilitate trajectory planning according to the massage positions, wherein when planning the path, the activity stroke corresponding to each joint in the mechanical arm is determined according to a preset physiotherapy course, the complex movement and operation of the mechanical arm in three-dimensional space is converted into the activity of each part of the mechanical arm, the mechanical arm is controlled to start the activity, and the massage device is controlled to start the massage, and an automated massage is formed by the physiotherapy robot, and different massage plans can be generated for different users and needs, so as to realize precise control of the force, position and speed of massage therapy. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] Figure 1 It is a flowchart of a massage trajectory planning method for a physiotherapy robot provided in an embodiment of the present invention.
[0015] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION
[0016] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0017] It should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.
[0018] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the meaning of "and / or" appearing in the full text includes three parallel schemes. Taking "A and / or B" as an example, it includes scheme A, or scheme B, or a scheme that satisfies both A and B. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.
[0019] See also Figure 1 The present invention provides a massage trajectory planning method for a physiotherapy robot, wherein the physiotherapy robot comprises a mechanical arm and a visual device and a massage device disposed at the end of the mechanical arm, and the massage trajectory planning method for the physiotherapy robot comprises: S10, acquiring an image captured by the visual device, and determining a human body contour according to the image; Among them, in this embodiment, the human body contour of the user is obtained through the visual device, and the human body contour in the target area is identified to provide basic data for subsequent physical therapy massage.
[0020] S20, determining massage positions of acupuncture points and meridians in the human body contour according to a preset deep learning model; After the human body contour is identified, the human body contour is brought into the preset deep learning model, and the positions of acupuncture points and meridians in the human body contour are obtained through the matching relationship with the preset deep learning model, and the massage path is planned based on the massage position. S30, determining the activity range corresponding to each joint in the robotic arm according to the preset physical therapy course and the massage position; Among them, the preset massage therapy courses are combined with the generated massage positions to determine the overall activity of the robotic arm, and then the complex movements and operations of the robotic arm in three-dimensional space are converted into the activity of each part of the robotic arm, so that the posture of the robotic arm is more natural and smooth. At the same time, the identified acupuncture points and meridians are combined to enable the robotic arm to perform physical therapy in accordance with traditional Chinese medicine theory.
[0021] It should be noted that the preset physiotherapy course is a physiotherapy plan prepared in advance based on traditional Chinese medicine physiotherapy massage theory.
[0022] S40, controlling the mechanical arm to start moving, and controlling the massage device to start high-precision massage.
[0023] In this embodiment, the physiotherapy robot includes a robotic arm and a visual device and a massage device arranged at the end of the robotic arm, wherein an image is captured by the visual device so as to determine the human body contour according to the image, and after the human body contour is determined, the human body contour is brought into a preset deep learning model, and the massage positions of the acupuncture points and meridians in the human body contour are calculated, so as to facilitate trajectory planning according to the massage positions, wherein when planning the path, the activity range corresponding to each joint in the robotic arm is determined according to the preset physiotherapy course, and the complex movement and operation of the robotic arm in three-dimensional space is converted into the activity of each joint of the robotic arm, the robotic arm is controlled to start the activity, and the massage device is controlled to start the massage, so that automated massage is formed by the physiotherapy robot, and different massage plans can be generated for different users and needs, so as to achieve precise control of the force, position and speed of massage therapy.
[0024] It should be noted that there are many ways to process images captured by the visual device in order to determine the human body contour, such as obtaining the contour by polarization processing after capturing the image with a camera; In addition, there are many ways to determine the active range corresponding to each joint in the robotic arm, such as extension and rotation of the robotic arm. In the present embodiment, the robotic arm is a six-axis collaborative robotic arm. Therefore, the step of determining the active range corresponding to each joint in the robotic arm includes determining the rotation angle corresponding to each joint in the robotic arm.
[0025] Specifically, the visual device includes a 3D camera and a 2D camera, and step S10 includes: S11, acquiring a 3D image captured by the 3D camera, and assigning coordinates to a plurality of pixel points in the 3D image; Among them, the image in the field of view is first extracted as pixel points by the 3D camera, and coordinates are assigned to all the pixel points to facilitate subsequent further identification.
[0026] S12, acquiring a 2D image captured by the 2D camera, identifying human body edge pixels among a plurality of the pixels through an instance segmentation algorithm, and obtaining the human body contour.
[0027] After the coordinates of the pixels are assigned, the human body contour in the target area is identified through the AI instance segmentation algorithm of the 2D image, providing basic data for subsequent physical therapy and massage.
[0028] It should be noted that, in this embodiment, the instance segmentation algorithm is intelligently implemented by AI to identify the human body contour in the target area without human intervention, and the implementation is simple and fast.
[0029] On the other hand, before step S10, the method further includes: S01, determining the base coordinates of the robotic arm; S02. According to the base coordinates, calibrate the tool coordinate system and the hand-eye calibration of the massage device.
[0030] In this embodiment, the tool coordinate system and hand-eye calibration of the massage device are performed through the base coordinates of the robotic arm to facilitate the use of the robotic arm and the massage device and provide basic parameters for the control of the robotic arm. Similarly, before step S10, the following steps are also included: S03. Perform dynamic calibration on the robotic arm.
[0031] The mechanical arm is dynamically calibrated to facilitate subsequent control of the mechanical arm to complete a specified action.
[0032] Furthermore, the method for obtaining the preset deep learning model includes: S21, obtaining a training human body contour image, and marking acupoints and meridian positions on the training human body contour image; First, based on traditional Chinese medicine theory, professional Chinese medicine physicians use annotation tools such as LabelImg to annotate the acupuncture points and meridians in the image dataset. Then engineers process the data and convert the data results into JOSN format. At the same time, they rotate, scale, and crop the images to increase the diversity of the dataset and improve the generalization ability of the model. S22. Importing the preset human body contour image and the corresponding acupuncture points and meridian positions into an initial deep learning model for training to obtain the preset deep learning model.
[0033] By increasing the depth of the convolutional layer and modifying the size and scale of the anchor box, the model structure of Faster R-CNN (regional convolutional neural network) is adjusted to better adapt to the characteristics of acupoints and meridians. The Faster R-CNN model is then trained using the prepared dataset and evaluated and adjusted to improve model performance.
[0034] In this embodiment, based on the advantages of deep learning technology in feature extraction, candidate region generation, classification and regression, the image is feature extracted through the convolution layer to quickly generate candidate regions, and the ROI Pooling (region of interest pooling) layer extracts feature vectors of fixed size. Finally, the classification and regression layer is used to accurately identify and locate acupoints and meridians. This process not only improves the recognition speed and accuracy, but also has strong generalization ability and robustness, providing a basis for subsequent trajectory planning.
[0035] Furthermore, step S22 includes: S221. When the number of training times reaches a preset number of times, stop training and use the model at this time as the preset deep learning model; S222. When the performance improvement value is less than a preset threshold, stop training and use the model at this time as the preset deep learning model.
[0036] In this embodiment, during the model training process, an Early Stopping strategy is triggered by setting a maximum number of iterations and a minimum performance improvement threshold to avoid overfitting during the training process.
[0037] On the other hand, step S40 includes: S41, obtaining the current magnitude on the current loop, and adjusting the position, speed and force parameters of the robotic arm according to a first preset mapping relationship.
[0038] In this embodiment, the load current change is fed back through the current loop to achieve timely response to the change of human body posture and ensure the safety of massage therapy.
[0039] Furthermore, the method for obtaining the preset physical therapy data includes: According to the user's needs, the preset physical therapy course is obtained according to the second preset mapping relationship.
[0040] In order to produce different massage courses according to different needs.
[0041] Further, controlling the robotic arm to start moving includes: The movement of the robotic arm is controlled by force-position hybrid control technology and compliant control technology.
[0042] By using the mechanical arm force-position hybrid control technology and the robot compliance control technology, the robot can simultaneously meet the precise control of position, force and speed when performing massage. These control strategies will dynamically adjust the movement trajectory, massage force and massage speed of the massage device based on the feedback information of the mechanical arm joint motor current loop, so that the massage device can always fit the contour of the human body with constant force.
[0043] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.
Claims
1. A massage trajectory planning method for a physiotherapy robot, characterized in that: The physiotherapy robot comprises a mechanical arm and a visual device and a massage device arranged at the end of the mechanical arm. The massage trajectory planning method of the physiotherapy robot comprises: Acquire an image captured by the visual device, and determine a human body contour based on the image; Determining massage positions of acupuncture points and meridians in the human body contour according to a preset deep learning model; Determining the range of motion corresponding to each joint in the robotic arm according to the preset physiotherapy course and the massage position; The mechanical arm is controlled to start moving, and the massage device is controlled to start high-precision massage.
2. The massage trajectory planning method of the physiotherapy robot according to claim 1, characterized in that: The visual device includes a 3D camera and a 2D camera; The step of acquiring the image captured by the visual device and determining the human body contour according to the image comprises: Acquire a 3D image captured by the 3D camera, and assign coordinates to a plurality of pixel points in the 3D image; A 2D image captured by the 2D camera is acquired, and human body edge pixel points among a plurality of the pixel points are identified by an instance segmentation algorithm to obtain the human body contour.
3. The massage trajectory planning method of the physiotherapy robot according to claim 1, characterized in that: Before acquiring the image captured by the visual device and determining the human body contour according to the image, the method further includes: Determining the base coordinates of the robotic arm; According to the base coordinates, the massage device is calibrated in the tool coordinate system and in the hand-eye calibration.
4. The massage trajectory planning method of the physiotherapy robot according to claim 3, characterized in that: Before acquiring the human body image captured by the visual device and acquiring the human body contour according to the human body image, the method further includes: The mechanical arm is dynamically calibrated.
5. The massage trajectory planning method of the physiotherapy robot according to claim 1, characterized in that: The method for obtaining the preset deep learning model includes: Acquire a training human body contour image, and mark acupoints and meridian positions on the training human body contour image; The preset human body contour image and the corresponding acupuncture points and meridian positions are imported into the initial deep learning model for training to obtain the preset deep learning model.
6. The massage trajectory planning method of the physiotherapy robot according to claim 5, characterized in that: The step of importing the preset human body contour image and the corresponding acupuncture points and meridians into the initial deep learning model for training to obtain the preset deep learning model comprises: When the training times reach the preset times, the training is stopped and the model at this time is used as the preset deep learning model; When the performance improvement value is less than a preset threshold, the training is stopped and the model at this time is used as the preset deep learning model.
7. The massage trajectory planning method of the physiotherapy robot according to claim 1, characterized in that: The physiotherapy robot also includes a current loop provided on each joint of the mechanical arm; The step of controlling the massage device to start massage comprises: The current magnitude on the current loop is obtained, and the position, speed and force parameters of the robot arm are adjusted according to a first preset mapping relationship.
8. The massage trajectory planning method of the physiotherapy robot according to claim 7, characterized in that: The controlling the robotic arm to start moving comprises: The movement of the robotic arm is controlled by force-position hybrid control technology and compliant control technology.
9. The massage trajectory planning method of the physiotherapy robot according to claim 1, characterized in that: The method for obtaining the preset physical therapy data includes: According to the user's needs, the preset physical therapy course is obtained according to the second preset mapping relationship.
Citation Information
Patent Citations
Massage robot, control method and massage method
CN117045497A
Camera device, system, and program
JP2024127407A
Cited By
Physiotherapy path regulation and control method and device of rehabilitation bed and rehabilitation bed
CN120233699A
A method and device for regulating a physiotherapy path of a rehabilitation bed and a rehabilitation bed
CN120233699B
Physiotherapy robot massage path generation method
CN120516680A
