Posture control apparatus and method for rehabilitation exercise robot

Abstract

Disclosed are a posture control apparatus and method for a rehabilitation exercise robot. The posture control apparatus for a rehabilitation exercise robot according to the present invention includes: a sensing unit which senses the posture of a user; an output unit for outputting feedback information for correcting the posture of the user; and a processor which collects the joint positions of the user on the basis of sensing information sensed by the sensing unit, determines whether the posture of the user is a compensatory movement on the basis of the joint positions, generates feedback information according to the determination result, and outputs the feedback information through the output unit.

Description

Posture control device and method for a rehabilitation exercise robot

[0001] The present invention relates to a posture control device and method for a rehabilitation exercise robot.

[0002] Patients with impaired upper limb motor function receive training to improve upper limb motor ability using a rehabilitation exercise robot.

[0003] For example, terminal device type rehabilitation exercise robots, which are in the form of a robot that the patient holds and moves by holding the handle of the robot, are widely used in clinical settings because they are small in size and easy to wear.

[0004] Meanwhile, patients often utilize other muscles or joints to perform movements when muscle weakness makes it difficult to perform desired movements. This is known as compensatory movement. A typical example is when a patient cannot extend their arm during a movement requiring arm extension, they lean forward to achieve the target.

[0005] When using end-effector robots, movement characteristics such as compensatory movements are frequently observed in patients. To encourage greater use of the arms during training, these movements must be reduced. Therefore, there is a need to reduce compensatory movements in patients to help correct their posture.

[0006] The background technology of the present invention is disclosed in Korean Patent Publication No. 10-2531025 (May 4, 2023), entitled “Robot device for muscle strengthening rehabilitation exercise of upper and lower limbs.”

[0007] The present invention was created to improve the above-mentioned problems, and an object of one aspect of the present invention is to provide a posture control device and method for a rehabilitation exercise robot that monitors the posture of a user in a robot exercise robot to determine whether it is a compensatory exercise and provides feedback so that the user can correct the posture based on the determination result.

[0008] A posture control device of a rehabilitation exercise robot according to one aspect of the present invention is characterized by including: a sensing unit that detects a user's posture; an output unit that outputs feedback information for correcting the user's posture; and a processor that collects the user's joint positions based on the detection information detected by the sensing unit, determines whether the user's posture is a compensatory movement based on the joint positions, and generates the feedback information based on the determination result and outputs it through the output unit.

[0009] The sensing unit of the present invention is characterized by including at least one of a camera for photographing a user and a radar for detecting the user's movement.

[0010] The sensing unit of the present invention is characterized in that it detects the user's posture in a non-contact manner without coming into contact with the user's body.

[0011] The feedback information of the present invention is characterized by including at least one of visual and auditory materials that guide the user to correct posture.

[0012] The joint location of the present invention is characterized by including at least one of the head, neck, torso, waist, left shoulder, left elbow, left wrist, left hand, right shoulder, right elbow, right wrist, and right hand.

[0013] The processor of the present invention is characterized in that it calculates a joint range of motion in a section defined by a start point and an end point of motion using the joint position, and calculates motion information within the joint range of motion.

[0014] The above movement information of the present invention is characterized in that it includes at least one of a forward movement distance of the left shoulder and right shoulder center points, a lateral movement distance of the torso rotation angle center point, an upward movement distance of the left shoulder or right shoulder coordinate, a torso forward tilt angle, a torso lateral tilt angle, a torso rotation angle, a shoulder abduction angle, a shoulder flexion angle, a shoulder horizontal abduction angle, an arm plane and torso angle, and an elbow angle.

[0015] The processor of the present invention is characterized in that it determines whether the user's posture is the compensatory movement by using a model that inputs the range of joint motion in a section defined by the start point and end point of the movement using the joint position.

[0016] The model of the present invention is characterized by including at least one of SVM (Support Vector Machine), K-NN (K-Nearest Neighbors), Random Forest, and Decision Tree.

[0017] A method for controlling posture of a rehabilitation exercise robot according to one aspect of the present invention is characterized by including: a step in which a sensing unit detects a user's posture; a step in which a processor collects the user's joint positions based on the detection information detected by the sensing unit, determines whether the user's posture is a compensatory movement based on the joint positions, and generates feedback information for correcting the user's posture based on the determination result; and a step in which an output unit outputs the feedback information.

[0018] The sensing unit of the present invention is characterized by including at least one of a camera for photographing a user and a radar for detecting the user's movement.

[0019] The sensing unit of the present invention is characterized in that it detects the user's posture in a non-contact manner without coming into contact with the user's body.

[0020] The feedback information of the present invention is characterized by including at least one of visual and auditory materials that guide the user to correct posture.

[0021] The joint location of the present invention is characterized by including at least one of the head, neck, upper body, waist, left shoulder, left elbow, left wrist, left hand, right shoulder, right elbow, right wrist, and right hand.

[0022] In the step of generating the feedback information of the present invention, the processor is characterized in that it calculates the joint range of motion in a section defined by the start point and the end point of the motion using the joint position, and calculates motion information within the joint range of motion.

[0023] The above movement information of the present invention is characterized in that it includes at least one of a forward movement distance of the left shoulder and right shoulder center points, a lateral movement distance of the torso rotation angle center point, an upward movement distance of the left shoulder or right shoulder coordinate, a torso forward tilt angle, a torso lateral tilt angle, a torso rotation angle, a shoulder abduction angle, a shoulder flexion angle, a shoulder horizontal abduction angle, an arm plane and torso angle, and an elbow angle.

[0024] In the step of generating the feedback information of the present invention, the processor is characterized in that it determines whether the user's posture is the compensatory movement by using a model that inputs the range of joint motion in a section defined by the start point and end point of the movement using the joint position.

[0025] The model of the present invention is characterized by including at least one of SVM (Support Vector Machine), K-NN (K-Nearest Neighbors), Random Forest, and Decision Tree.

[0026] A posture control device and method of a rehabilitation exercise robot according to one aspect of the present invention monitors the posture of a user in a rehabilitation exercise robot to determine whether it is a compensatory exercise and provides feedback so that the user can correct the posture based on the determination result.

[0027] A posture control device and method for a rehabilitation exercise robot according to another aspect of the present invention can provide an environment with improved user convenience and reduce the number of medical personnel required to monitor the user's posture.

[0028] A posture control device and method for a rehabilitation exercise robot according to another aspect of the present invention can provide a training environment capable of judging the complex posture of the entire upper body of a user.

[0029] A posture control device and method for a rehabilitation exercise robot according to another aspect of the present invention can enable a user to train in various postures and induce active posture correction of the user through feedback, thereby increasing the effectiveness of rehabilitation training for a rehabilitation subject.

[0030] Figure 1 is a block diagram of a posture control device of a rehabilitation exercise robot according to one embodiment of the present invention.

[0031] FIG. 2 is a drawing illustrating a rehabilitation exercise robot according to one embodiment of the present invention.

[0032] Figure 3 is a drawing showing a joint position according to one embodiment of the present invention.

[0033] Figures 4a to 4d are exemplary diagrams of unit sections according to one embodiment of the present invention.

[0034] Figure 5 is a drawing showing an anatomical plane according to one embodiment of the present invention.

[0035] Figure 6 is a diagram illustrating a principle for extracting motion information according to one embodiment of the present invention.

[0036] Figures 7a to 7c are diagrams showing a data acquisition environment according to one embodiment of the present invention.

[0037] FIG. 8 is a diagram illustrating an example of classification and feedback of a correction request posture according to an embodiment of the present invention.

[0038] Figure 9 is a flowchart of a posture control method of a rehabilitation exercise robot according to one embodiment of the present invention.

[0039] Hereinafter, a posture control device and method for a rehabilitation exercise robot according to an embodiment of the present invention will be described in detail with reference to the attached drawings. In this process, the thickness of lines and the sizes of components depicted in the drawings may be exaggerated for clarity and convenience. Furthermore, the terms described below are defined based on their functions in the present invention and may vary depending on the intentions or practices of the user or operator. Therefore, the definitions of these terms should be based on the contents throughout this specification.

[0040] The present invention can be implemented in various different forms and is not limited to the embodiments described herein. Furthermore, in the drawings, irrelevant parts have been omitted to clearly illustrate the present invention, and like reference numerals have been used throughout the specification to designate similar parts.

[0041] Throughout the specification, whenever a part is said to "include" a component, this does not mean that it excludes other components, but rather that it may include other components, unless otherwise stated.

[0042] The implementations described herein may be implemented, for example, as a method or process, a device, a software program, a data stream, or a signal. Even if discussed only in the context of a single form of implementation (e.g., discussed only as a method), the implementation of the discussed features may also be implemented in other forms (e.g., as a device or program). The devices may be implemented using suitable hardware, software, firmware, and the like. The methods may be implemented in a device such as a processor, which generally refers to a processing device including, for example, a computer, a microprocessor, an integrated circuit, or a programmable logic device.

[0043] FIG. 1 is a block diagram of a posture control device of a rehabilitation exercise robot according to an embodiment of the present invention, and FIG. 2 is a drawing illustrating a rehabilitation exercise robot according to an embodiment of the present invention.

[0044] Referring to FIG. 1, a posture control device of a rehabilitation exercise robot according to one embodiment of the present invention may include a sensing unit (100), a processor (200), and an output unit (300).

[0045] The sensing unit (100) can detect the user's posture.

[0046] The user may be a patient with reduced upper limb motor function who is receiving rehabilitation training to improve upper limb motor ability using a rehabilitation exercise robot (400).

[0047] The sensing unit (100) can detect the user's posture in a non-contact manner, without contacting the user's body. The non-contact sensing unit (100) minimizes contact with the user's body, thereby enhancing user convenience and reducing the preparation time for rehabilitation training.

[0048] The sensing unit (100) may be a 3D camera that captures the user's posture or a radar that detects the user's movements. The type of sensing unit (100) is not particularly limited.

[0049] The sensing unit (100) can be installed in a rehabilitation exercise robot (400). The installation location of the sensing unit (100) is not particularly limited as long as it is a location where the user's posture can be detected.

[0050] The rehabilitation exercise robot (400) may be a terminal device type upper limb rehabilitation exercise robot (400).

[0051] Referring to FIG. 2, a user can sit in front of a rehabilitation exercise robot (400), and an image output unit (310) can be mounted in front of the user. A sensing unit (100) can be mounted above or behind the image output unit (310).

[0052] While the user performs upper limb rehabilitation training using the front image output unit (310), the sensing unit (100) of the rehabilitation exercise robot (400) can detect the user's posture.

[0053] A rehabilitation exercise robot (400) may include a base (410), a main body (420), a robot arm (430), a support (440), and a table (450).

[0054] The base (410) can support the main body (420) while being placed on the floor.

[0055] The main body (420) can be installed on the upper part of the base (410).

[0056] The main body (420) can be moved up and down relative to the base (410), thereby allowing the height of the robot arm (140) to be adjusted. The main body (420) can be manually moved up and down by the user. In another embodiment, the main body (420) can be automatically moved up and down using a motor or the like.

[0057] The robot arm (430) protrudes forward from the main body (420). Here, the robot arm (430) can be driven by a mechanism that moves with at least two degrees of freedom. For example, the robot arm (430) can be configured to move with two degrees of freedom in the horizontal direction. To this end, the robot arm (430) can be configured with a link structure composed of two or more links.

[0058] The support member (440) is connected to the end of the robot arm (430) and can support the user's hand. Accordingly, the user can perform rehabilitation exercises while placing his or her hand on the support member (440).

[0059] The support member (440) may be formed with a handle that the user can hold. The user can hold the handle and move it in two degrees of freedom, and can also move it while the hand is fixed by a fixed band.

[0060] A table (450) can be installed between the main body (420) and the base (410), and the robot arm (430) can be positioned on the upper part of the table (450).

[0061] Accordingly, the user positions the forearm on the robot arm (430) to compensate for the gravity of the user's arm and moves the handle. The robot's handle moves in a plane, and the robot can detect the position of the robot's handle.

[0062] The video output unit (310) displays a cursor on the screen according to the movement of the handle when the user holds and moves the handle.

[0063] Through this process, the user is able to perform rehabilitation exercises.

[0064] Meanwhile, while the user is performing rehabilitation exercises, the sensing unit (100) can detect the user's posture. The sensing information detected by the sensing unit (100) may be image information and may be input to the processor (200).

[0065] The processor (200) can collect the user's joint positions based on the detection information detected by the sensing unit (100). Based on the joint positions, the processor (200) can determine whether the user's posture is a compensatory movement and generate feedback information to correct the user's posture. The processor (200) can output the generated feedback information through the output unit (300).

[0066] The processor (200) may include a posture collection unit (210), a posture analysis unit (220), a posture determination unit (230), and a posture feedback unit (240).

[0067] The posture collection unit (210) can collect the user's joint positions using the detection information detected by the sensing unit (100) and track them in three dimensions.

[0068] Figure 3 is a drawing showing a joint position according to one embodiment of the present invention.

[0069] Referring to FIG. 3, the joint locations may include Head, Neck, Torso, Waist, Left Shoulder, Left Elbow, Left Wrist, Left Hand, Right Shoulder, Right Elbow, Right Wrist, and Right Hand.

[0070] The joint positions shown in Fig. 3 are as shown in Table 1 below.

[0071] Joint numberJoint name1Head2Neck3Torso4Waist5Left shoulder6Left elbow7Left wrist8Left hand9Right shoulder10Right elbow11Right wrist12Right hand

[0072] The posture collection unit (210) can detect the elbow joint angle based on the elbow joint position and calculate the elbow joint angular velocity using this elbow joint angle. The posture collection unit (210) can determine whether or not it is a re-estimation section based on the elbow joint angular velocity.

[0073] An interval can be defined as the moment a user moves from point 1 to point 2, from the moment they leave position 1 to the moment they arrive at position 2.

[0074] The posture collection unit (210) can determine a section as a re-estimation section if the elbow joint angular velocity is greater than or equal to a preset reference angular velocity.

[0075] The posture collection unit (210) can improve the accuracy of the joint position by calculating the elbow joint position by dividing it into a re-estimation section and a non-re-estimation section.

[0076] If the section is determined not to be a re-estimation section, the posture collection unit (210) can calculate the elbow joint position using the robot end position and the first upper arm direction vector.

[0077] The first upper arm direction vector can be detected through a detection image captured by the sensing unit (100). The posture collection unit (210) can detect an elbow position vector and a shoulder position vector through detection information analysis, and can detect the first upper arm direction vector using the elbow position vector and the shoulder position vector.

[0078] If the section is determined to be a re-estimation section, the posture collection unit (210) can detect a second upper arm direction vector by applying the current shoulder position vector detected from the previous elbow position vector and the posture image of the current step to the process of detecting the upper arm direction vector.

[0079] In the re-estimation section, the posture collection unit (210) can detect the elbow joint position based on the second upper arm direction vector and the current robot end position information.

[0080] In the re-estimation section, the angular velocity of the elbow joint is high, so the accuracy of detection of the elbow position vector for the elbow joint through image analysis may be reduced. Therefore, by applying the previous elbow position vector detected in the previous step to the second upper arm direction vector, the problem of reduced accuracy of the joint position can be compensated for.

[0081] The posture analysis unit (220) can calculate the characteristics and joint angles of the arms and torso that can be used for determination based on the collected three-dimensional joint positions.

[0082] The posture analysis unit (220) can define the start and end points of movement to calculate joint movement information and calculate the range of motion (ROM) of the joint in the corresponding section (trial).

[0083] Meanwhile, as described above, a section can be defined as the moment a user leaves location 1 and arrives at location 2 when moving from point 1 to point 2, but the definition method may vary depending on the content.

[0084] Figures 4a to 4d are exemplary diagrams of unit sections according to one embodiment of the present invention.

[0085] Referring to FIGS. 4a to 4d, examples of unit sections are provided, and the method of defining sections may vary depending on the content.

[0086] An example of a unit section (223) may be a movement from the starting point (221) of the arrows in FIGS. 4A to 4D to the destination point (222). One section may typically be a movement of less than several seconds (s).

[0087] The posture analysis unit (220) moves during the above-mentioned time period (t goal ~t start ) and calculate the range of movement within that section.

[0088] The motion information that can be calculated within the range of motion defined as a section is as shown in Table 2 below. The posture analysis unit (220) can calculate the distance moved by a joint or the angle of joint motion in the section.

[0089] Distance / AngleMovement InformationDefinitionDistance (mm) Trunk lean-forwardForward movement of the center points of the left and right shouldersDistance (mm) Trunk lean-lateralSideward movement of the center points of the left and right shouldersDistance (mm)Shoulder elevationUpward movement of the coordinates of the left and right shouldersAngle (degree) Trunk flexionAngle of forward lean of the trunkAngle (degree) Trunk lateral flexionAngle of sideways lean of the trunkAngle (degree) Trunk rotationAngle (degree) Shoulder abductionAngle of shoulder abduction (angle between upper arm and trunk in coronal plane)Angle (degree) Shoulder flexionAngle of shoulder flexion (angle between upper arm and trunk in sagittal plane)Angle (degree) Shoulder horizontal abductionAngle of shoulder horizontal abduction (angle between upper arm and trunk in transverse plane)Angle (degree) Arm planeAngle between arm plane and trunk (arm plane: plane formed by shoulder, elbow, and wrist)Angle (degree) Elbow flexionAngle of elbow (angle between upper arm and angle of the lower arm)

[0090] The posture analysis unit (220) can calculate trunk lean-forward, trunk lean-lateral, and shoulder elevation using the distance that a specific point moves during the movement section. The posture analysis unit (220) can obtain trunk flexion angle, trunk lateral flexion angle, and trunk rotation angle by calculating the rotation matrix (R) of the trunk coordinate system. Once the rotation matrix is ​​calculated, the rotation angles γ, β, and α of the trunk can be calculated using the following mathematical equations (1) to (4). Here, γ is trunk flexion, β is trunk latreral flexion, and α is trunk rotation.

[0091] (1)

[0092] (2)

[0093] (3)

[0094] (4)

[0095] FIG. 5 is a drawing showing an anatomical plane according to an embodiment of the present invention, and FIG. 6 is a drawing showing a principle for extracting motion information according to an embodiment of the present invention.

[0096] Referring to FIGS. 5 and 6, the posture analysis unit (220) projects shoulder abduction, shoulder flexion, shoulder horizontal abduction, arm plane, and elbow flexion onto the anatomical plane at an angle () using the following mathematical equations (5) to (7). ; projection) or the angle between two vectors ( ; relative) can be obtained by calculating it.

[0097] The posture analysis unit (220) can calculate the joint range of motion by subtracting the minimum value from the maximum value among the distances or angles calculated as described above in the section. The spectator range of motion can be calculated using the following mathematical equation (8).

[0098] ROM = max(vaule)-min(value)

[0099] The posture determination unit (230) can determine whether the user's posture is a compensatory movement based on the features calculated by the posture analysis unit (220).

[0100] The posture determination unit (230) may include a classifier learned based on data that an actual medical professional has determined to require correction.

[0101] The classifier judges a large number of compensatory movement data, allowing for a variety of desirable postures during training.

[0102] In an embodiment, a non-linear support vector machine may be used as a classifier.

[0103] The data required for the classifier can be collected from both non-disabled and patients.

[0104] Figures 7a to 7c are diagrams showing a data acquisition environment according to one embodiment of the present invention.

[0105] Referring to Figures 7a to 7c, the experimental environment in which the reaching movement was performed was a center-out reaching movement performed on a plane, a total of 24 reaching movement targets in 8 directions x 3 distances were presented on the screen, and the user performed the reaching movement by moving the handle to move to the targets presented on the screen.

[0106] All subjects were required to perform reaching movements with minimal trunk movement, but stroke patients may unintentionally include compensatory movements.

[0107] Each subject performed the reaching exercise at least 196 times. RGB images were captured while the subject performed the reaching exercise. These images were later reviewed by a therapist and labeled to determine whether compensatory exercises were used or whether the posture required correction. Table 3 below provides examples.

[0108] Trial exercise interval(Trial) Whether compensation exercise is used subject 11O2O3X4X5O......194X195X196Osubject 21X2O3O......196Xsubject 3...............subject 4......

[0109] The therapist's judgment (labeling) results were used to train a classifier. The support vector machine (SVM) is a model that defines the optimal decision boundary that maximizes the margin between each group, and can be used to find the boundary between movement information using compensatory movement and movement information without compensatory movement.

[0110] The methods that can be used to create a classifier model are not limited to SVM, and K-NN (K-Nearest Neighbors), Random Forest, and Decision Tree can also be used.

[0111] The movement information calculated in the posture analysis unit (220) and the labeling results judged by the therapist can be used as input values ​​for the classifier learning data.

[0112] After scaling and normalizing the data, a linear kernel can be applied when applying linear SVM, and a Gaussian kernel can be applied when applying nonlinear SVM.

[0113] An example of applying linear SVM is as follows.

[0114] Linear SVM result example Prediction No reward motion With reward motion Actual No reward motion 106985 With reward motion 282266

[0115]

[0116] An example of applying nonlinear SVM is as follows.

[0117] Nonlinear SVM result example Prediction No compensation movement With compensation movement Actual No compensation movement 938215 With compensation movement 166382

[0118]

[0119] After dividing the training data and learning data in a 7:3 ratio, accuracy and sensitivity were calculated using 5-fold cross-validation, and the information required for the calculation is shown in Table 4 below.

[0120] Predicted Compensatory Exercise NoneCompensatory Exercise UsedActual Compensatory Exercise NoneTruenegative(TN)Falsepositive(FP)Compensatory Exercise UsedFalsenegative(FN)Truepositive(TP)

[0121] Accuracy represents the proportion of diagonal matrices that are predicted correctly out of the total.

[0122] Accuracy can be calculated using the mathematical formula 9 below.

[0123] (9)

[0124] Sensitivity represents the proportion of actual compensatory movements classified as compensatory movements.

[0125] Sensitivity can be calculated using the mathematical expression 10 below.

[0126] (10)

[0127] In linear SVM, the accuracy was 78.3% and the sensitivity was 48.5%, and in nonlinear SVM, the accuracy was 77.5% and the sensitivity was 69.6%.

[0128] To further improve accuracy and sensitivity, other classifiers can be used, such as K-Nearest Neighbors (K-NN), Random Forest, and Decision Tree, in addition to SVM. In addition, any classifier that can learn compensation motion data and determine classification in real time is also possible.

[0129] The detailed determination unit (230) can determine whether a trial is a section using compensatory movement by processing the calculated ROM value as an input value of a classifier model.

[0130] Accordingly, the posture determination unit (230) can transmit the ROM value used in classifier learning as input to the posture determination unit (230) and the determination result to the posture feedback unit (240).

[0131] The posture feedback unit (240) can generate feedback information for correcting the user's posture based on the determination result.

[0132] That is, the posture feedback unit (240) can generate feedback information to guide the part requiring correction by referring to the joint angle calculated by the analysis unit.

[0133] Feedback information may include visual and auditory aids that guide the user to correct their posture.

[0134] The output unit (300) can output feedback information received from the posture feedback unit (240).

[0135] The output unit (300) may include a video output unit (310) and an audio output unit (320). The video output unit (310) may output visual data to guide the user to correct their posture. The audio output unit (320) may output auditory data to guide the user to correct their posture.

[0136] FIG. 8 is a diagram illustrating an example of classification and feedback of a correction request posture according to an embodiment of the present invention.

[0137] As illustrated in FIG. 8, the output unit (300) can output feedback information such as “Move your body less”, “Stretch your arms more”, “Bend your arms more”, “I’ll try again”, “You moved your body less”, etc., in the form of visual data (text, improved posture guidance pictures, etc.) and auditory data (alarm sounds, voice guidance, etc.).

[0138] Accordingly, the user can receive guidance on posture correction through sound or visual data and can attempt active posture correction. At this time, the output unit (300) can provide feedback to the user to let them know whether the active posture correction was successful.

[0139] In this embodiment, to help understand the embodiment, the posture collection unit (210), the posture analysis unit (220), the posture determination unit (230), and the posture feedback unit (240) are described as separate components within the processor (200). However, depending on the embodiment, the processor (200) may be implemented as a configuration in which each sub-component is performed in an integrated manner.

[0140] Figure 9 is a flowchart of a posture control method of a rehabilitation exercise robot (400) according to one embodiment of the present invention.

[0141] Referring to Fig. 9, the sensing unit (100) can detect the user's posture (S100).

[0142] The posture collection unit (210) can collect the user's joint positions using the detection information detected by the sensing unit (100) (S200). The joint positions may include the head, neck, torso, waist, left shoulder, left elbow, left wrist, left hand, right shoulder, right elbow, right wrist, and right hand.

[0143] Additionally, the posture collection unit (210) can correct the joint position. In this case, the posture collection unit (210) can determine whether or not it is a re-estimation section based on the elbow joint angular velocity. If it is a re-estimation section according to the determination result, the posture collection unit (210) can correct the joint position by applying the previous elbow position vector detected in the previous step to the second upper arm direction vector (S300).

[0144] The posture analysis unit (220) can calculate the characteristics and joint angles of the arms and torso that can be used for determination based on the collected three-dimensional joint positions.

[0145] That is, the posture analysis unit (220) can calculate motion information that can be calculated within a range defined as a section, and can calculate the distance moved or the joint operation angle in the section (S400).

[0146] The posture analysis unit (220) can determine whether the user's posture is a compensatory movement based on the features calculated by the posture analysis unit (220) (S500).

[0147] In this case, the posture determination unit (230) and the posture analysis unit (220) may include a classifier learned based on compensatory movement data that an actual medical professional has determined requires correction.

[0148] The posture feedback unit (240) can generate feedback information for correcting the user's posture based on the determination results (S600). In this case, the posture feedback unit (240) can generate feedback information to guide the user to the areas requiring correction by referring to the joint angles calculated by the posture analysis unit (220).

[0149] The output unit (300) can output feedback information received from the posture feedback unit (240), for example, visual data or auditory data.

[0150] In this way, the posture control device and method of a rehabilitation exercise robot according to one embodiment of the present invention monitors the posture of a user in the rehabilitation exercise robot to determine whether it is a compensatory exercise and provides feedback so that the user can correct the posture based on the determination result.

[0151] The posture control device and method of a rehabilitation exercise robot according to one embodiment of the present invention can provide an environment with improved user convenience and reduce the number of medical personnel required to monitor the user's posture.

[0152] The posture control device and method of a rehabilitation exercise robot according to one embodiment of the present invention can provide a training environment capable of judging the complex posture of the entire upper body of a rehabilitation subject.

[0153] The posture control device and method of a rehabilitation exercise robot according to one embodiment of the present invention can enable a user to train in various postures and induce active posture correction of the user through feedback, thereby increasing the effectiveness of the user's rehabilitation training.

[0154] The implementations described herein may be implemented, for example, as a method or process, a device, a software program, a data stream, or a signal. Even if discussed only in the context of a single form of implementation (e.g., discussed only as a method), the implementation of the discussed features may also be implemented in other forms (e.g., as a device or program). The devices may be implemented using suitable hardware, software, firmware, and the like. The methods may be implemented in a device such as a processor, which generally refers to a processing device including, for example, a computer, a microprocessor, an integrated circuit, or a programmable logic device.

[0155] While the present invention has been described with reference to the embodiments illustrated in the drawings, these are merely exemplary, and those skilled in the art will understand that various modifications and equivalent embodiments are possible. Accordingly, the true technical protection scope of the present invention should be defined by the following claims.

Claims

1. A sensing unit that detects the user's posture; An output section that outputs feedback information to correct the user's posture; and A posture control device for a rehabilitation exercise robot, including a processor that collects the user's joint positions based on the detection information detected by the sensing unit, determines whether the user's posture is a compensatory movement based on the joint positions, and generates the feedback information based on the determination result and outputs it through the output unit.

2. In the first paragraph, the sensing unit A posture control device for a rehabilitation exercise robot, comprising at least one of a camera for photographing a user and a radar for detecting the user's movements.

3. In the first paragraph, the sensing unit A posture control device for a rehabilitation exercise robot that detects the user's posture in a non-contact manner without contacting the user's body.

4. In paragraph 1, the feedback information is A posture control device for a rehabilitation exercise robot including at least one of visual and auditory materials that guide a user to correct posture.

5. In the first paragraph, the joint position is A posture control device for a rehabilitation exercise robot including at least one of a head, a neck, a torso, a waist, a left shoulder, a left elbow, a left wrist, a left hand, a right shoulder, a right elbow, a right wrist, and a right hand.

6. In the first paragraph, the processor A posture control device of a rehabilitation exercise robot that calculates a joint range of motion in a section defined by the start and end points of movement using the above joint positions and calculates movement information within the above joint range of motion.

7. In paragraph 6, the exercise information is A posture control device for a rehabilitation exercise robot including at least one of a forward movement distance of the left shoulder and right shoulder center points, a lateral movement distance of the torso rotation angle center point, an upward movement distance of the left shoulder or right shoulder coordinates, a torso forward tilt angle, a torso lateral tilt angle, a torso rotation angle, a shoulder abduction angle, a shoulder flexion angle, a shoulder horizontal abduction angle, an arm plane and torso angle, and an elbow angle.

8. In the first paragraph, the processor A posture control device for a rehabilitation exercise robot that determines whether a user's posture is a compensatory exercise using a model that inputs the range of joint motion in a section defined by the start and end points of the exercise using the above joint positions.

9. In paragraph 8, the model A posture control device for a rehabilitation exercise robot including at least one of SVM (Support Vector Machine), K-NN (K-Nearest Neighbors), Random Forest, and Decision Tree.

10. Step in which the sensing unit detects the user's posture; A step in which the processor collects the user's joint positions based on the detection information detected by the sensing unit, determines whether the user's posture is a compensatory movement based on the joint positions, and generates feedback information for correcting the user's posture based on the determination result; and A method for controlling posture of a rehabilitation exercise robot, comprising a step of outputting the above feedback information.

11. In the 10th paragraph, the sensing unit A method for controlling posture of a rehabilitation exercise robot, the robot including at least one of a camera for photographing a user and a radar for detecting the user's movement.

12. In the 10th paragraph, the sensing unit A method for controlling the posture of a rehabilitation exercise robot that detects the user's posture in a non-contact manner without contacting the user's body.

13. In paragraph 10, the feedback information is A method for controlling posture of a rehabilitation exercise robot, the method including at least one of visual and auditory materials guiding a user to correct posture.

14. In the 10th paragraph, the joint position is A method for controlling posture of a rehabilitation exercise robot including at least one of a head, neck, upper body, waist, left shoulder, left elbow, left wrist, left hand, right shoulder, right elbow, right wrist, and right hand.

15. In the step of generating the feedback information in paragraph 10, A method for controlling posture of a rehabilitation exercise robot, wherein the processor calculates a range of joint motion in a section defined by the start and end points of the movement using the joint positions, and calculates movement information within the range of joint motion.

16. In paragraph 15, the exercise information is A posture control method of a rehabilitation exercise robot, comprising at least one of a forward movement distance of the left shoulder and right shoulder center points, a lateral movement distance of the torso rotation angle center point, an upward movement distance of the left shoulder or right shoulder coordinates, a torso forward tilt angle, a torso lateral tilt angle, a torso rotation angle, a shoulder abduction angle, a shoulder flexion angle, a shoulder horizontal abduction angle, an arm plane and torso angle, and an elbow angle.

17. In the step of generating the feedback information in paragraph 10, The above processor is a posture control method of a rehabilitation exercise robot that determines whether the user's posture is the compensation exercise by using a model that inputs the range of joint motion in a section defined by the start point and end point of the movement using the above joint positions.

18. In paragraph 17, the model A method for controlling posture of a rehabilitation exercise robot, comprising at least one of SVM (Support Vector Machine), K-NN (K-Nearest Neighbors), Random Forest, and Decision Tree.

Images