Upper limb mirror image exercise robot control method and system based on virtual reality
Through virtual reality technology, the healthy side movement data is collected and reference trajectories are generated on the affected side. Combined with the auxiliary force field, the problem of failure to fully utilize the active movement of the affected side in the prior art is solved, and the efficiency and effect of upper limb mirror rehabilitation training is improved.
Patent Information
- Application Number
- CN202411028468.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-07-29
AI Technical Summary
The existing upper limb mirror rehabilitation robots fail to fully consider the active movement of the patient's affected side during the spasm and recovery periods, which weakens the patient's active participation in rehabilitation training and affects the rehabilitation effect.
Using the upper limb mirror exercise robot control method based on virtual reality, the healthy side motion data is collected through virtual reality equipment, and a smooth reference trajectory is generated, and displayed to the affected side through mirror conversion, combining tangential guidance force, normal correction force and gravity compensation. Strive to build an auxiliary force field to help patients exercise in virtual reality scenarios.
It improves the active participation and training efficiency of patients, significantly improves the stability and accuracy of rehabilitation training, and enhances the rehabilitation effect.
Smart Images

Figure CN118903787A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of mirror image rehabilitation robots, and in particular to a control method and system of an upper limb mirror image training robot based on virtual reality. Background Art
[0002] Stroke is the second leading cause of death and the leading cause of disability in my country; nearly 80% of stroke survivors will suffer from limb disability, among which the upper limb dysfunction is more significant; in recent years, mirror therapy has been proven to be an effective training method to promote upper limb motor function rehabilitation in patients with hemiplegia caused by stroke. Traditional mirror therapy uses a plane mirror to copy the activities of the healthy side to the affected side, and uses visual illusions and rehabilitation training to significantly improve upper limb motor function. However, this method cannot provide arm support, the therapeutic training efficiency is low, and the training results are unstable.
[0003] The upper limb mirror rehabilitation robot can solve the problems encountered in traditional mirror rehabilitation training and provide support and guidance for the patient's affected arm to achieve better rehabilitation effect; however, the existing upper limb mirror rehabilitation robot mainly extracts joint position point information from the healthy side and uses PD control for the joints to convert it into the rotation angle and angular velocity of each joint of the robot on the affected side; this control method fails to take into account the active movement of the affected side during the spasm and recovery periods, weakens the patient's active participation in the affected limb in rehabilitation training, and affects the rehabilitation effect. Summary of the invention
[0004] The purpose of the present invention is to provide a virtual reality-based upper limb mirror training robot control method and system, which can improve the patient's active participation and training efficiency by improving the robot system and control strategy.
[0005] Based on the above purpose, the present invention adopts the following technical solution:
[0006] The upper limb mirror training robot control method based on virtual reality includes the following steps:
[0007] S1. Formulate training plan: The therapist formulates training plan according to the patient's condition and plans training movements and paths for the patient. The therapist formulates a movement plan suitable for the patient according to the patient's needs.
[0008] S2, collecting healthy side sample points: the healthy side of the patient controls the operating handle of the virtual reality device and repeats the training action planned in S1; the n-dimensional position vector of the healthy hand of the patient in the Cartesian coordinate system is collected through the operating handle, and the sample point data is obtained through downsampling; the virtual reality device collects the position vector of the healthy hand in real time through a constant sampling frequency, records the motion trajectory and generates sample points;
[0009] S3. Generate healthy side reference trajectory: Use Bezier curve interpolation method to generate a smooth healthy side reference trajectory based on sample point data; Bezier curve interpolation method generates a smooth and continuous curve through a series of control points to ensure the continuity and smoothness of the trajectory. This generated curve, as the healthy side reference trajectory, provides an accurate movement path, allowing patients to perform effective repetitive exercises according to the plan formulated by the therapist during rehabilitation training, thereby improving the effectiveness and efficiency of rehabilitation training;
[0010] S4, generating an affected-side reference trajectory: generating an affected-side reference trajectory through mirror transformation on the sagittal plane according to the healthy-side reference trajectory generated in S3; the patient wears a head-mounted display device of a virtual reality device, and displays the affected-side reference trajectory in a virtual scene displayed by the head-mounted display device; by generating the affected-side reference trajectory in real time in the virtual scene, the patient can intuitively see and follow the trajectory to exercise, and the intuitive visual feedback can help the patient to imitate and train movements more accurately, and effectively improve the rehabilitation effect;
[0011] S5. Auxiliary training: When the patient uses the training robot to train along the reference trajectory of the affected side, the training robot generates an auxiliary force field composed of tangential guiding force, normal correction force and gravity compensation force in real time to help the patient's affected hand move along the reference trajectory of the affected side; during the training process, the patient's healthy side moves along the healthy side reference trajectory according to prompts in the virtual reality scene, and at the same time, the training robot on the affected side constructs an auxiliary force field composed of guiding force, correction force and gravity compensation force, wherein the tangential guiding force guides the affected side to move along the reference trajectory of the affected side, the normal correction force provides correction force according to the tracking error, and assists the affected side to return to the reference trajectory, and the gravity compensation force ensures that the patient is not affected by gravity during exercise; through the auxiliary force of the terminal traction robot, the patient's affected side can move along the reference trajectory of the affected side under the auxiliary force field of the terminal traction robot, thereby improving the training effect and rehabilitation efficiency.
[0012] Preferably, the process of S2 collecting healthy side sample points includes:
[0013] The n-dimensional position vector P∈R in the Cartesian coordinate system is calculated at fixed time intervals n Sampling is performed, and then the collected sample points are downsampled; m equally distributed key points are extracted and used as control points of the Bezier curve. The formula is as follows:
[0014]
[0015] In the above formula, Ω represents the downsampled data set, represents a set of m n-dimensional position vectors; where P k represents the kth position vector, k is 1-m, R nRepresents n-dimensional real space, that is, each vector P k are all n-dimensional vectors; ∈R n×n It means that this set is a subset of the n-by-m real matrix, and can be regarded as a column vector of a matrix with n rows and m columns.
[0016] The above sampling points are further downsampled. The specific steps of downsampling are as follows: calculate the total trajectory length L between the original sample points, divide the total length L into m segments on average, and the length of each segment is The distance is accumulated along the original trajectory. When the accumulated distance reaches the length of each segment, the point is selected as a key point. The above process is repeated until m equally distributed key points are selected.
[0017] Preferably, the process of S3 generating the healthy side reference trajectory includes:
[0018] According to the control points and order, the Bezier curve interpolation algorithm is used to generate points on the curve and obtain the healthy side reference trajectory; the r-th order Bezier curve is a curve generated by r+1 control points, and its formula is as follows:
[0019]
[0020] In the formula, B r (t) represents the value of the r-th order Bezier curve at parameter t, where r is the order of the Bezier curve; the value range of the interpolation parameter t is [0, 1], P j represents the jth control point, B j,r is the Bernstein basis function.
[0021] The construction of the Bezier curve is based on the Bernstein polynomial. The r-th order Bernstein polynomial is as follows:
[0022]
[0023] In the above formula, B j,r (t) is the value of the Bernstein polynomial, which represents the jth Bernstein multi-basis function of the rth order Bezier curve. The exponent of the polynomial The binomial coefficient represents the number of combinations of j elements taken from r elements. The calculation formula is
[0024] Furthermore, by adjusting the value range of parameter t, a smooth and continuous trajectory can be generated on the definition domain of the Bezier curve; the specific steps for adjusting parameter t are:
[0025] Initialize t to 0, set Δt to 0 (for example, Δt = 0.01), and use the Bezier curve formula to calculate B r (t):
[0026]
[0027] Record the calculated B r (t) is used as the trajectory point, and t is increased by Δt
[0028] t=t+Δt
[0029] Repeat the steps until t reaches or exceeds 1.
[0030] Preferably, the process of generating the reference trajectory of the affected side in S4 includes:
[0031] The control point set of the healthy side reference trajectory is P = {(x i ,y i , z i )|i=0,1,...,r},where (x i ,y i , z i ) represents the coordinates of the i-th control point;
[0032] For each control point (x i ,y i , z i ), calculate its mirror image point about the sagittal plane (-x i ,y i , z i ), obtain the control point set P′ after mirror transformation, and obtain the reference trajectory of the affected side:
[0033] B′ r (t) = (-x(t), y(t), z(t))
[0034] In the formula, B′ r (t) represents the value of the Bezier curve on the affected side at parameter t, and the value range of the interpolation parameter t is [0, 1].
[0035] Preferably, the process of S5 auxiliary training includes:
[0036] An auxiliary force field is constructed to help guide the patient's affected hand to move along the affected reference trajectory. The formula is as follows:
[0037] F assist =F t +F n +F g_comp
[0038] In the above formula, F assist represents the comprehensive auxiliary force field, F t represents the tangential guiding force, F n Denotes the normal correction force, F g_comp Represents the gravity compensation force.
[0039] Preferably, the tangential guiding force F t Obtained by the following method:
[0040]
[0041] In the above formula, F t is the tangential guiding force, Represents the unit tangent vector of the nearest point, N ILC is the distance between the patient's current position and the nearest point on the reference trajectory, N tgt is a preset distance threshold used to determine whether to apply a tangential guiding force. When the distance between the patient's current position and the nearest point on the reference trajectory is less than this threshold, a tangential guiding force is applied. The magnitude of the tangential guiding force is f t It changes with the change of hysteresis Δt, and its calculation formula is as follows
[0042]
[0043] In the above formula, Δt represents the time lag between the patient’s current position and the target position on the reference trajectory, L t is the hysteresis threshold, f t_set It is the preset maximum guiding force value. When the hysteresis reaches or exceeds the hysteresis threshold, the tangential guiding force reaches this value.
[0044] Preferably, the normal correction force F n Obtained by the following method:
[0045]
[0046] In the above formula, F n is the normal correction force, P IL represents the position coordinates of the patient's limbs at the current moment, P ILC represents the position coordinates of the point on the reference trajectory closest to the patient's current position, ||P IL -P ILC || is a scalar that represents the distance between the patient’s current position and the nearest point on the reference trajectory, which is used to calculate the direction of the correction force and the magnitude of the normal correction force f n It depends on the deviation between the patient's current position and the closest point on the reference trajectory, calculated as follows:
[0047]
[0048] In the above formula, R s is the adjustable radius. When the distance between the current position and the nearest point on the reference trajectory is within this range, it means that the affected side has excellent trajectory tracking ability and no normal correction force is applied; where R his the distance threshold at which the patient's affected side cannot complete trajectory tracking. The distance between the current position and the nearest point on the reference trajectory is R s and R h When the distance between the current position and the nearest point on the reference trajectory exceeds R h When the normal correction force is composed of two parts, one part is k h Δk represents a constant correction force, and the other part is k s (||P IL -P ILC ||-R h ) represents the normal correction force and exceeds R h increases in direct proportion to the distance.
[0049] Preferably, the steps for generating the gravity compensation force are specifically as follows:
[0050] The gravity compensation force is used to offset the gravity of the affected upper limb so that the patient is not affected by gravity during exercise. The gravity compensation force F g_comp The calculation formula is as follows:
[0051] F g_comp =m·g
[0052] In the above formula, F g_comp The force used to offset the effect of gravity on the patient's limbs, m is the mass of the patient's limb, and g is the acceleration due to gravity.
[0053] An upper limb mirror training robot system, comprising:
[0054] A healthy side motion information acquisition module, the healthy side motion information acquisition module comprising a handle device; the motion trajectory and position data of the healthy side of the patient are captured in real time through the handle device of the virtual reality device, the captured data are processed, and a smooth reference trajectory is generated after Bezier curve interpolation, and the processed data is transmitted to the visual feedback module and the affected side auxiliary module for use in training;
[0055] A visual feedback module, wherein the visual feedback module includes a head display device; the head display device of the virtual reality device is used to mirror the healthy side reference trajectory about the sagittal plane to generate the affected side reference trajectory, and the reference trajectory is displayed in real time in the virtual reality scene, so that the patient can intuitively see and follow the reference trajectory for training;
[0056] An affected-side auxiliary module comprises an end-traction upper limb training robot cooperating with the patient's affected side, and a six-dimensional force / torque sensor and a handheld controller are arranged at the end of the end-traction upper limb training robot: auxiliary forces such as tangential guiding force, normal correction force and gravity compensation force are provided by the end-traction upper limb rehabilitation robot, the tangential guiding force provides guiding force along the trajectory direction, the normal correction force corrects when the movement deviates, and the gravity compensation force provides vertical supporting force to help the patient's affected upper limb move along the reference trajectory, ensuring that the patient can accurately follow the affected-side reference trajectory during training and promote the rehabilitation effect.
[0057] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements any step in the above-mentioned upper limb mirror training robot control method.
[0058] An electronic device includes a memory and a processor. The memory stores a computer program. The processor executes the computer program. When the processor executes the computer program, any step in the control method of the upper limb mirror training robot is implemented.
[0059] The beneficial effects of the present invention are as follows:
[0060] The present invention uses an end-traction upper limb training robot to provide guidance and assistance for the patient's affected hand through a multi-dimensional force field composed of tangential guiding force, normal correction force and gravity compensation force; the tangential guiding force ensures that the affected upper limb moves along a reference trajectory, the normal correction force provides corrective force when there is movement deviation, and the gravity compensation force offsets the influence of limb gravity; the combined effect of these force fields significantly improves the stability and accuracy of training, ensuring that patients can perform rehabilitation training according to a predetermined trajectory, thereby effectively improving the rehabilitation effect.
[0061] The present invention combines virtual reality technology, and patients can wear virtual reality head display devices to enter virtual reality scenes; the virtual reality head display device generates a healthy side reference trajectory, and mirrors it about the sagittal plane, generates an affected side reference trajectory and displays it in the virtual reality scene; in the virtual reality scene, the patient can intuitively see and follow the reference trajectory for training. Through this visual illusion, the patient believes that the affected side has achieved the same exercise effect as the healthy side, which can effectively enhance the patient's enthusiasm and help improve the participation and rehabilitation effect of rehabilitation training.
[0062] The present invention integrates three modules, namely, visual feedback, healthy-side motion information collection and affected-side assistance, and can provide all-round support for postoperative rehabilitation of stroke patients; the healthy-side motion information collection module captures the motion trajectory and position data of the patient's healthy-side upper limb in real time, generates a smooth reference trajectory, and transmits it to other modules; the visual feedback module uses a virtual reality head display device to mirror the healthy-side reference trajectory about the sagittal plane, generates an affected-side reference trajectory, and displays it in real time in a virtual reality scene; the affected-side assistance module helps the affected-side upper limb move along the reference trajectory through an end-traction upper limb rehabilitation robot; through the synergistic effect of these three modules, the patient's affected and healthy sides can be trained synchronously, forming a systematic and scientific closed loop of rehabilitation training, which greatly improves the efficiency and effect of rehabilitation. BRIEF DESCRIPTION OF THE DRAWINGS
[0063] Figure 1 This is a flow chart of the upper limb mirror training robot control method according to Embodiment 1 of the present invention;
[0064] Figure 2 This is a basic control principle diagram of the upper limb mirror training robot according to Embodiment 1 of the present invention;
[0065] Figure 3 Schematic diagram of a method for finding the closest point according to Embodiment 1 of the present invention;
[0066] Figure 4 This is a schematic diagram of a third-order Bezier curve in Example 1 of the present invention;
[0067] Figure 5 This is a schematic diagram of the rehabilitation training task strategy in Example 1 of the present invention;
[0068] Figure 6 This is a schematic diagram of the deployment of the upper limb mirror training robot system in Example 2 of the present invention;
[0069] Figure 7 This is a conceptual diagram of the auxiliary force field of Example 2 of the present invention.
[0070] In the figure: head display device 1; operating handle 2; collaborative robotic arm 3; six-dimensional force / torque sensor 4; handheld controller 5; control computer 6; virtual reality computing device 7. DETAILED DESCRIPTION
[0071] Example 1
[0072] The following is a further explanation of the present invention in conjunction with specific embodiments. Figure 1 and Figure 2 As shown, this embodiment provides a method for controlling an upper limb mirror training robot based on virtual reality, comprising the following steps:
[0073] S1. Develop a training plan: The therapist develops a training plan based on the patient's condition and plans training movements and paths for the patient; the therapist develops a movement plan suitable for the patient's training based on the patient's needs.
[0074] S2. Collect healthy side sample points: The healthy side of the patient controls the operating handle of the virtual reality device and repeats the training action planned in S1; the patient holds the handle device on the healthy side and collects healthy side sample points according to the training path; the specific steps of healthy side sample point collection are as follows:
[0075] The n-dimensional position vector P∈R in the Cartesian coordinate system is calculated at fixed time intervals n Sampling is performed, and then the collected sample points are downsampled; m equally distributed key points are extracted and used as control points of the Bezier curve. The formula is as follows:
[0076]
[0077] In the above formula, Ω represents the downsampled data set, represents a set of m n-dimensional position vectors; where P k represents the kth position vector, k is 1-m, R n Represents n-dimensional real space, that is, each vector P k are all n-dimensional vectors; ∈R n×m It means that this set is a subset of the n-by-m real matrix, and can be regarded as a column vector of a matrix with n rows and m columns.
[0078] The above sampling points are further downsampled. The specific steps of downsampling are as follows: calculate the total trajectory length L between the original sample points, divide the total length L into m segments on average, and the length of each segment is The distance is accumulated along the original trajectory. When the accumulated distance reaches the length of each segment, the point is selected as a key point. The above process is repeated until m equally distributed key points are selected.
[0079] When collecting healthy side sample points in S2, in order to avoid multiple closest points and closest point jumps in the method of traversing path points, the following method can be used: Figure 3 The method shown determines the closest point on the path;
[0080] Where P clo and P clo ′ is the nearest point. The nearest point on the path can be found by using the forward tangent vector τ of the path. Pi , from P i To P act Vector The dot product is determined as follows:
[0081]
[0082] By checking the change in the sign of the dot product, the corresponding closest point P on the path can be identified clo .
[0083] Then, the virtual reality device collects the position vector of the healthy hand in real time through a constant sampling frequency, records the motion trajectory and generates sample points.
[0084] S3. Generate the healthy side reference trajectory based on the collected sample points:
[0085] The rth-order Bezier curve is a curve generated by r+1 control points, and the interpolation formula is:
[0086]
[0087] In the above formula, r is the order of the Bezier curve, B r (t) represents the value of the r-th order Bezier curve at parameter t. The value range of the interpolation parameter t is [0, 1]. j represents the jth control point, B j,r is the Bernstein basis function.
[0088] The construction of the Bezier curve is based on the Bernstein polynomial. The r-th order Bernstein polynomial is as follows:
[0089]
[0090] In the above formula, B j,r (t) is the value of the Bernstein polynomial, which represents the jth Bernstein multi-basis function of the rth order Bezier curve. The exponent of the polynomial The binomial coefficient represents the number of combinations of j elements taken from r elements. The calculation formula is
[0091] Take the third-order Bezier curve as an example, Figure 4 As shown, Figure 4 A third-order Bezier curve can be generated by passing through four control points P0, P1, P2, and P3; auxiliary points A0, A1, and A2 are taken on P0P1, P1P2, and P2P3 respectively, so that ||P0A0||: ||P0P1||=||P1A1||: ||P1P2||=||P2A2||: ||P2P2||=t, and ensure that there is a point B on A0A1 that satisfies the condition ||A0B0||: ||A0A1||=||A1B1||: ||A1A2||=t, then 0 satisfies ||B0O||: ||B0B1||=t, that is, the trajectory curve formula of point 0 is as follows:
[0092] O(t)=(1-t)3 P0+3t(1-t) 2 +3t 2 (1-t)P2+t 3 P3
[0093] By adjusting the value range of parameter t, a smooth and continuous trajectory can be generated on the domain of the Bezier curve. The specific steps for adjusting parameter t are:
[0094] Initialize t to 0, set Δt to 0 (for example, Δt = 0.01), and use the Bezier curve formula to calculate B r (t):
[0095]
[0096] Record the calculated B r (t) as the trajectory point, increase t by Δt:
[0097] t=t+Δt
[0098] Repeat the above steps until t reaches or exceeds 1.
[0099] S4. Generate the reference trajectory of the affected side according to the reference trajectory of the healthy side:
[0100] According to the control points in the healthy side reference trajectory, the healthy side reference trajectory control point set P is generated, and the formula is as follows:
[0101] P = {(x i ,y i , z i )|i=0,1,...,r}
[0102] Where (x i ,y i , z i ) represents the coordinates of the ith control point. For each control point (x i ,y i , z i ), calculate its mirror image point about the sagittal plane (-x i ,y i , z i ), use the control point set after mirror transformation as P′, and get the Bezier curve B′ of the affected side r The specific formula is as follows:
[0103] B′ r (t) = (-x(t), y(t), z(t))
[0104] According to the obtained Bezier curve formula of the affected side, a reference trajectory of the affected side can be generated in the virtual reality scene, providing the patient with a basis for the exercise path of the affected side.
[0105] The patient wears a virtual reality headset and displays the reference trajectory of the affected side in the virtual scene displayed by the headset. By generating the reference trajectory of the affected side in real time in the virtual scene, the patient can intuitively see and follow the trajectory to exercise. Intuitive visual feedback can help the patient imitate and train movements more accurately and effectively improve the rehabilitation effect.
[0106] S5. Perform auxiliary training based on the reference trajectory of the affected side:
[0107] During the patient's exercise, the affected hand cooperates with the end-traction upper limb exercise robot to move along the reference trajectory of the affected side. At this time, an auxiliary force field is constructed based on the data collected in real time by each sensor to assist the patient's affected hand in moving. The auxiliary force field includes tangential guiding force, normal correction force and gravity compensation force. The tangential guiding force is used to guide the patient's affected hand to move along the trajectory direction. The normal correction force provides correction force when the patient's affected hand deviates from the trajectory to assist the affected hand to return to the correct trajectory. The gravity compensation force is used to offset the gravity of the limbs to ensure that the patient is not affected by gravity during exercise. The formula for constructing the auxiliary force field is as follows:
[0108] F assist =F t +F n +F g_comp
[0109] In the above formula, F assist is the comprehensive auxiliary force field, F t is the tangential guiding force, F n is the normal correction force, F g_comp is the gravity compensation force.
[0110] The specific steps for generating the tangential guiding force are as follows:
[0111] The tangential guiding force is used to guide the patient's affected upper limb movement along the tangent direction of the reference trajectory to ensure that it moves along the correct trajectory. The tangential guiding force F t The calculation formula is as follows:
[0112]
[0113] In the above formula, F t is the tangential guiding force, Represents the unit tangent vector of the nearest point, N ILC is the distance between the patient's current position and the nearest point on the reference trajectory, N tgt is a preset distance threshold used to determine whether to apply a tangential guidance force. When the distance between the patient's current position and the nearest point on the reference trajectory is less than this threshold, a tangential guidance force is applied. The tangential guidance force F tThe size of f t It changes with the change of hysteresis Δt, and its calculation formula is as follows:
[0114]
[0115] In the above formula, Δt represents the time lag between the patient’s current position and the target position on the reference trajectory, L t is the hysteresis threshold, f t_set It is the preset maximum guiding force value. When the hysteresis reaches or exceeds the hysteresis threshold, the tangential guiding force reaches this value.
[0116] The specific steps for generating the normal correction force are:
[0117] like Figure 5 As shown, the normal correction force is used to provide a corrective force when the patient deviates from the trajectory, pulling him back to the correct movement trajectory to ensure the accuracy of the movement. The normal correction force F n The calculation formula is as follows:
[0118]
[0119] In the above formula, F n is the normal correction force, P IL represents the position coordinates of the patient's limbs at the current moment, P ILC represents the position coordinates of the point on the reference trajectory closest to the patient's current position, ||P IL -P ILC || is a scalar that represents the distance between the patient’s current position and the nearest point on the reference trajectory, which is used to calculate the direction of the correction force, the normal correction force F n The size of f n It depends on the deviation between the patient's current position and the closest point on the reference trajectory, calculated as follows:
[0120]
[0121] In the above formula, R s is the adjustable radius. When the distance between the current position and the nearest point on the reference trajectory is within this range, it means that the IL (affected side) has excellent trajectory tracking ability and no normal correction force is applied. h is the distance threshold at which the patient's affected side cannot complete trajectory tracking. When the distance between the current position and the nearest point on the reference trajectory is within R s and R h If the distance between the current position and the nearest point on the reference trajectory exceeds R, the robot will provide a normal correction force, and the normal correction force provided by the tracking error will increase in proportion to the error distance. h When the normal correction force is composed of two parts, one part is k hΔk represents a constant correction force, and the other part is k s (||P IL -P ILC ||-R h ), indicating the magnitude of the normal correction force and the value exceeding R h increases in direct proportion to the distance.
[0122] The specific steps for generating gravity compensation force are as follows:
[0123] The gravity compensation force is used to offset the gravity of the affected upper limb so that the patient is not affected by gravity during exercise. The gravity compensation force Fg _ The comp calculation formula is as follows:
[0124] F _ co mp =m·g
[0125] In the above formula, Fg _ co mp The force used to offset the effect of gravity on the patient's limbs, m is the mass of the patient's limb, and g is the acceleration due to gravity.
[0126] Overall rehabilitation training task strategy Figure 5 As shown in the figure, the two expected path curves represent the predetermined training paths, HL represents the healthy side and IL represents the affected side; the real-time position PHL of the healthy arm of the patient is captured using a virtual reality device, and the nearest point P on the expected path of the healthy side is calculated based on the captured healthy arm position. HLC , P HLC The mirror image position of IL is used as the target point Ptgt of IL along the path. The real-time position PIL of the patient's affected arm is captured by the sensor, and the nearest point P on the expected path of the affected side is calculated. ILC , an assistive force is applied to the affected arm through the end-traction upper limb rehabilitation robot to help it move along the reference path.
[0127] Example 2
[0128] The difference between this embodiment and embodiment 1 is that Figure 6 As shown, this embodiment provides an upper limb mirror training robot system based on virtual reality, including a healthy side movement information acquisition module, a visual feedback module and an affected side auxiliary module.
[0129] The visual feedback module includes a head display device 1 of a virtual reality device, which is used to display a virtual reality scene for the patient; the healthy side movement information acquisition module includes an operating handle 2 of the virtual reality device, which is used to collect the healthy side movement trajectory of the patient; the affected side auxiliary module includes an end-traction upper limb exercise robot, which mainly includes a collaborative robotic arm 3 (such as a UR5e collaborative robot), and a six-dimensional force / torque sensor 4 is arranged at the end of the collaborative robotic arm 3 to collect various data of the patient's affected hand in real time, as well as a handheld controller 5 for cooperating with the patient's affected hand, which can cooperate and fix the patient's affected hand with the end of the collaborative robotic arm 3; in addition, a control computer 6 and a virtual reality computing device 7 are also provided for data processing, path planning, virtual reality scene generation and the like.
[0130] In this embodiment, the healthy side motion information acquisition module can capture the patient's healthy side motion trajectory and position data in real time through the operating handle 2; the captured data is processed and a smooth healthy side reference trajectory is generated using the Bezier curve interpolation method, and then the processed data is transmitted to the visual feedback module and the affected side auxiliary module for use in training.
[0131] The visual feedback module can use the head display device 1 to mirror the generated healthy side reference trajectory about the sagittal plane to generate an affected side reference trajectory; the generated affected side reference trajectory can be displayed in real time in a virtual reality scene, so that the patient can intuitively see the trajectory path with the help of the head display device 1, and drive the affected side hand to move along the affected side reference path for training.
[0132] The affected-side auxiliary module provides auxiliary forces such as tangential guidance force, normal correction force and gravity compensation force to the patient's affected-side hand through an end-traction upper limb training robot that is fixedly matched with the affected-side hand. Under the action of the auxiliary force, the patient's affected side can move more accurately along the affected-side reference trajectory, thereby improving the effect of rehabilitation training.
[0133] The tangential guiding force provides guiding force along the trajectory direction, the normal correction force corrects when the movement deviates, and the gravity compensation force provides vertical support force. Combining the above three auxiliary forces, an auxiliary force field is constructed. The force field is designed to guide the patient's affected side to move along the affected side reference trajectory, as shown below:
[0134] F assist =F t +F n +F g_comp
[0135] In the above formula, F assist is the comprehensive auxiliary force field, F t is the tangential guiding force, F n is the normal correction force, F g_conp is the gravity compensation force.
[0136] In this embodiment, a schematic diagram of an auxiliary force field composed of multiple auxiliary forces is shown in FIG. Figure 7 As shown, for the convenience of expression, the three-dimensional curve is simplified to a plane representation; IL Represents the current position of IL, P ILC represents the closest point of IL on the path, P tgt represents the target point of IL on the path, along the path from P ILC To the target point P tgt The tangential guiding force F t , from P IL To P ILC Normal correction force F n ; The figure does not show the vertical upward gravity compensation force F applied to IL in order to improve training comfort. g_comp .
[0137] The tangential guiding force is used to guide the patient's affected upper limb movement along the tangent direction of the reference trajectory to ensure that it moves along the correct trajectory. The tangential guiding force F t The calculation formula is as follows:
[0138]
[0139] In the above formula, F t is the tangential guiding force, Represents the unit tangent vector of the nearest point, N ILC is the distance between the patient's current position and the nearest point on the reference trajectory, N tgt is a preset distance threshold used to determine whether to apply a tangential guiding force. When the distance between the patient's current position and the nearest point on the reference trajectory is less than this threshold, a tangential guiding force is applied. The magnitude of the tangential guiding force is f t It changes with the change of hysteresis Δt, and its calculation formula is as follows
[0140]
[0141] In the above formula, Δt represents the time lag between the patient’s current position and the target position on the reference trajectory, L t is the hysteresis threshold, f t_set is the preset maximum guiding force value. When the hysteresis reaches or exceeds the hysteresis threshold, the tangential guiding force reaches this value. The deviation error of IL relative to the predetermined path (||Δl n ||) can represent the motion accuracy of IL, which is P in the Cartesian coordinate system. IL and P ILC The difference between them is calculated as follows:
[0142]
[0143] The normal correction force is used to provide a corrective force when the patient deviates from the trajectory, pulling him back to the correct trajectory to ensure the accuracy of the movement. The normal correction force F n The calculation formula is as follows:
[0144]
[0145] In the above formula, F n is the normal correction force, P IL represents the position coordinates of the patient's limbs at the current moment, P ILC represents the position coordinates of the point on the reference trajectory closest to the patient's current position, ||P IL -P ILC || is a scalar that represents the distance between the patient’s current position and the nearest point on the reference trajectory, which is used to calculate the direction of the correction force and the magnitude of the normal correction force f n It depends on the deviation between the patient's current position and the closest point on the reference trajectory, calculated as follows:
[0146]
[0147] In the above formula, R s is the adjustable radius. When the distance between the current position and the nearest point on the reference trajectory is within this range, it means that the IL has excellent trajectory tracking ability and does not apply normal correction force. h is the distance threshold at which the patient's affected side cannot complete trajectory tracking. When the distance between the current position and the nearest point on the reference trajectory is within R s and R h If the distance between the current position and the nearest point on the reference trajectory exceeds R, the robot will provide a normal correction force, and the normal correction force provided will increase in proportion to the distance according to the tracking error. h When the normal correction force is composed of two parts, one part is k h Δk represents a constant correction force, and the other part is k s (||P IL -P ILC ||-R h ) represents the normal correction force and exceeds R h increases in direct proportion to the distance;
[0148] The specific steps for generating gravity compensation force are as follows:
[0149] The gravity compensation force is used to offset the gravity of the affected upper limb so that the patient is not affected by gravity during exercise. The gravity compensation force F g_comp The calculation formula is as follows:
[0150] F g_comp =m·g
[0151] In the above formula, F g_comp The force used to offset the effect of gravity on the patient's limbs, m is the mass of the patient's limb, and g is the acceleration due to gravity.
[0152] Example 3
[0153] The difference between this embodiment and embodiment 2 is that this embodiment is a computer-readable storage medium on which a computer program is stored, and the computer program includes the upper limb mirror rehabilitation robot system of the above-mentioned embodiment 2; or / and when the computer program is executed by the processor, any step in the control method of the upper limb mirror rehabilitation robot of the above-mentioned embodiment 1 is implemented.
[0154] Example 4
[0155] The difference between this embodiment and embodiment 3 is that this embodiment is an electronic device, including a memory and a processor, the memory stores a computer program, the computer program includes the upper limb mirror rehabilitation robot system of the above-mentioned embodiment 2; or / and the processor implements any step in the control method of the upper limb mirror rehabilitation robot of the above-mentioned embodiment 1 when executing the computer program.
[0156] The above description is only a further explanation of the present invention in combination with specific embodiments. All descriptions do not limit the scope of protection of the present invention. Any changes or replacement schemes that can be easily thought of by any technician in this field within the technical scope disclosed by the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.
Claims
1. A method for controlling an upper limb mirror training robot based on virtual reality, characterized in that: The following steps are involved: S1. Develop a training plan: Develop a training plan based on the patient's condition, and plan the training movements and paths; S2, collecting healthy side sample points: the healthy side of the patient controls the operating handle of the virtual reality device and repeats the training action planned in S1; the n-dimensional position vector of the healthy side hand of the patient in the Cartesian coordinate system is collected through the operating handle, and the sample point data is obtained by downsampling; S3, generating a healthy side reference trajectory: using Bezier curve interpolation method to generate a smooth healthy side reference trajectory based on the sample point data; S4, generating an affected-side reference trajectory: generating an affected-side reference trajectory through a mirror image transformation on a sagittal plane according to the healthy-side reference trajectory generated in S3; S5. Auxiliary training: When the patient uses the training robot to train according to the reference trajectory of the affected side, the training robot generates an auxiliary force field composed of tangential guiding force, normal correction force and gravity compensation force in real time to help the patient's affected hand move along the reference trajectory of the affected side.
2. The upper limb mirror training robot control method based on virtual reality according to claim 1 is characterized in that: In step S2, the process of collecting healthy side sample points includes: The n-dimensional position vector P∈R in the Cartesian coordinate system is calculated at fixed time intervals n Sampling is performed, and then the collected sample points are downsampled; m equally distributed key points are extracted as control points of the Bezier curve, as follows: In the formula, Ω represents the downsampled data set, represents a set of m n-dimensional position vectors; where P k represents the kth position vector, k is 1-m, R n Represents n-dimensional real space, that is, each vector P k are all n-dimensional vectors; ∈R n×m It means that this set is a subset of the n-by-m real matrix, and can be regarded as a column vector of a matrix with n rows and m columns.
3. The upper limb mirror training robot control method based on virtual reality according to claim 2 is characterized in that: In step S3, the process of generating the healthy side reference trajectory includes: According to the control points and order, the Bezier curve interpolation algorithm is used to generate points on the curve and obtain the healthy side reference trajectory; among them, the r-th order Bezier curve is a curve generated by r+1 control points: In the formula, B r (t) represents the value of the r-th order Bezier curve at parameter t, where r is the order of the Bezier curve; the value range of the interpolation parameter t is [0,1], P j represents the jth control point, B j,r (t) is the value of the Bernstein polynomial, representing the j-th Bernstein multibasis function of the r-th order Bezier curve.
4. The upper limb mirror training robot control method based on virtual reality according to claim 3 is characterized in that: The process of generating the reference trajectory of the affected side in S4 includes: The control point set of the healthy side reference trajectory is P = {(x i ,y i ,z i )|i=0,1,…,r}, where (x i ,y i ,z i ) represents the coordinates of the i-th control point; For each control point (x i ,y i ,z i ), calculate its mirror image point about the sagittal plane (-x i ,y i ,z i ), obtain the control point set P′ after mirror transformation, and obtain the reference trajectory of the affected side: B′ r (t)=(-x(t),y(t),z(t)) In the formula, B r '(t) represents the value of the Bezier curve on the affected side at parameter t, and the value range of the interpolation parameter t is [0,1].
5. The upper limb mirror training robot control method based on virtual reality according to claim 1, characterized in that: The S5 assisted training process includes: Construct an auxiliary force field to help guide the patient's affected hand to move along the affected reference trajectory: F assist =F t +F n +F g_comp In the above formula, F assist represents the comprehensive auxiliary force field, F t represents the tangential guiding force, F n Denotes the normal correction force, F g_comp Represents the gravity compensation force.
6. The upper limb mirror training robot control method based on virtual reality according to claim 5, characterized in that: The tangential guiding force F t Obtained by the following method: In the above formula, F t is the tangential guiding force, Represents the unit tangent vector of the nearest point, N ILC is the distance between the patient's current position and the nearest point on the reference trajectory, N tgt is a preset distance threshold used to determine whether to apply a tangential guiding force. When the distance between the patient's current position and the nearest point on the reference trajectory is less than this threshold, a tangential guiding force is applied. The magnitude of the tangential guiding force is f t It changes with the hysteresis Δt.
7. The upper limb mirror training robot control method based on virtual reality according to claim 5, characterized in that: The normal correction force F n Obtained by the following method: In the above formula, F n is the normal correction force, P IL represents the position coordinates of the patient's limbs at the current moment, P ILC represents the position coordinates of the point on the reference trajectory closest to the patient's current position, ||P IL -P ILC || is a scalar representing the distance between the patient's current position and the closest point on the reference trajectory, which is used to calculate the direction of the correction force.
8. An upper limb mirror training robot system, characterized in that: include: A healthy side motion information acquisition module, the healthy side motion information acquisition module comprising a handle device; the motion trajectory and position data of the healthy side of the patient are captured in real time through the handle device of the virtual reality device, the captured data are processed, and a smooth reference trajectory is generated after Bezier curve interpolation, and the processed data is transmitted to the visual feedback module and the affected side auxiliary module for use in training; A visual feedback module, wherein the visual feedback module includes a head display device; the head display device of the virtual reality device is used to mirror the healthy side reference trajectory about the sagittal plane to generate the affected side reference trajectory, and the reference trajectory is displayed in real time in the virtual reality scene, so that the patient can intuitively see and follow the reference trajectory for training; An affected-side auxiliary module comprises an end-traction upper limb training robot cooperating with the patient's affected side, and a six-dimensional force / torque sensor and a handheld controller are arranged at the end of the end-traction upper limb training robot: auxiliary forces such as tangential guiding force, normal correction force and gravity compensation force are provided by the end-traction upper limb rehabilitation robot, the tangential guiding force provides guiding force along the trajectory direction, the normal correction force corrects when the movement deviates, and the gravity compensation force provides vertical supporting force to help the patient's affected upper limb move along the reference trajectory, ensuring that the patient can accurately follow the affected-side reference trajectory during training and promote the rehabilitation effect.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, any step in the upper limb mirror training robot control method as described in any one of claims 1-7 is implemented.
10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program and the processor executes the computer program, wherein: When the processor executes the computer program, any step in the control method of the upper limb mirror training robot as described in any one of claims 1-7 is implemented.
Citation Information
Patent Citations
Upper limb both-arm rehabilitation training human-computer interaction method and system
CN110215676A
Advanced prediction control method for limb movement rehabilitation based on bilateral cooperation
CN113223666A
Upper limb rehabilitation training system for stroke patients
WO2023087954A1
Cited By
Mechanical arm control method and system based on self-adaptive force field
CN121200029A
