KR20220072026A - Rehabilitation Training Method and System Using Rehabilitation Robot - Google Patents

Abstract

The present invention relates to a rehabilitation training system and method, wherein the system according to the present invention includes a sensing unit that detects bio-signal data and motion data of a patient undergoing rehabilitation training using a rehabilitation robot device, and bio-signal data detected by the sensing unit. and a classification unit that receives motion data and classifies the patient's rehabilitation training state and motion using a pre-trained classification algorithm, and a coaching unit that provides coaching based on the result of classifying the patient's rehabilitation training status and motion do. The classification algorithm may be trained using learning data including trainer biosignal data, trainee motion data, and coaching information of a therapist for the trainee collected in a rehabilitation training process using the rehabilitation robot device. According to the present invention, it is possible to provide robot parameters suitable for the patient based on the robot parameter data and the patient's bio-signals and motion data and to correct the patient's posture, thereby reducing the burden on the physical therapist and providing the patient's bio-signals and motion data. Because it is monitored, there is an advantage that safe rehabilitation training is possible.

Description

Rehabilitation Training Method and System Using Rehabilitation Robot

The present invention relates to a rehabilitation training method and system using a rehabilitation robot, and to a rehabilitation training method and system using a rehabilitation robot that can collaborate with a physical therapist by proposing robot parameters and patient rehabilitation training motion coaching based on artificial intelligence .

With the spread of various lower extremity rehabilitation robots, the role of physical therapists in performing lower extremity rehabilitation training has increased and the importance has increased. A physical therapist who performs lower extremity rehabilitation training continuously interacts with the patient and plays a role of safely and successfully leading training using the lower extremity rehabilitation robot. In the case of rehabilitation training using the already developed and disseminated lower extremity rehabilitation robots, the patient's condition changes suddenly because it repeatedly performs only the fixed joint trajectory provided by the lower extremity rehabilitation robot, rather than operating the robot using the patient's biometric data. Because it is difficult to cope with this, and monitoring of the patient's training status is not provided, the success or failure of the patient's rehabilitation training depends solely on the physical therapist. Physical therapists are required to respond to emergency situations such as unexpected robot abnormal behavior and patient emergency during robot rehabilitation training using the conventional lower extremity rehabilitation robot, and at the same time, it is necessary for the patient to Rehabilitation training should be successfully led through interaction. For example, in the case of a lower extremity rehabilitation robot that provides repetitive gait motions to the patient for gait correction, continuous monitoring is required to ensure that the patient is riding the rehabilitation robot while properly following the gait trajectory intended by the rehabilitation robot. In addition, the burden on the physical therapist has increased, such as to maximize the effect of training using the robot by inducing active and voluntary participation of the patient riding the rehabilitation robot. Therefore, there is a need for a coach assist system that can relieve the burden on the physical therapist and assist with motion correction.

Korean Patent No. 10-1859549 (Registration Date 2018.05.14)

Therefore, the technical problem to be solved by the present invention is to provide a rehabilitation training method and system that provides a virtual coaching function capable of cooperating with a physical therapist.

Rehabilitation training system according to the present invention for solving the above technical problem is a sensing unit that detects bio-signal data and motion data of a patient who is training using a rehabilitation robot device, and bio-signal data and motion detected by the sensing unit It includes a classification unit that receives data and classifies the patient's rehabilitation training state and motion using a pre-trained classification algorithm, and a coaching unit that provides coaching based on the result of classifying the patient's rehabilitation training status and motion .

The classification algorithm may be trained using learning data including trainer biosignal data, trainee motion data, and coaching information of a therapist for the trainee collected in a rehabilitation training process using the rehabilitation robot device.

The training data may be labeled with coaching information of a therapist in correspondence to the trainee bio-signal data and the trainee motion data.

The learning data may be collected during a predetermined clinical trial.

The coaching unit receives a value set according to a therapist's tendency as a result of classifying the patient's rehabilitation training state and motion, and provides coaching according to a result of comparing the coaching reference value calculated according to a predetermined equation and the current coaching probability You can decide whether to output information.

The coaching unit receives a robot parameter for controlling the operation of the rehabilitation robot device, and uses the input robot parameter and the result of classifying the patient's rehabilitation training state and motion to recommend for changing the operation of the rehabilitation robot device You can provide robot parameters.

The coaching unit may provide coaching for the classification result when a result of dividing the patient's biosignal data and motion data by a predetermined window size is equal to or greater than a predetermined threshold.

The classification unit may include a plurality of classification algorithms trained for each type of coaching.

Data necessary for each of the plurality of classification algorithms may be selected and inputted from the biosignal data and motion data sensed by the sensing unit.

The coaching type may include a plurality of motion correction coaching classified for each rehabilitation training motion and training stop coaching for stopping rehabilitation training when an emergency occurs.

The sensing unit may include at least one of a surface electromyography sensor, a current skin response sensor, and an electrocardiogram sensor for measuring the patient's biosignal data.

It may include at least one of an inertial measurement sensor, an electronic goniometer, and a vision sensor for measuring the motion data of the patient.

Rehabilitation training method according to the present invention for solving the above technical problem, the step of detecting the bio-signal data and motion data of a patient undergoing rehabilitation training using a rehabilitation robot device, the detected bio-signal data and the motion data receiving the input, classifying the patient's rehabilitation training state and motion using a pre-trained classification algorithm, and providing coaching based on a result of classifying the patient's rehabilitation training status and motion.

According to the present invention, there is an advantage that safe rehabilitation training is possible because the patient's bio-signals and motion data are monitored. In addition, there is an advantage in that it is possible to provide robot parameters suitable for the patient based on the patient's bio-signals and motion data, and the robot parameters. In addition, there is an advantage that the patient's posture can be corrected based on the patient's bio-signals and motion data and robot parameters. In addition, it has the advantage of reducing the burden on the therapist by providing robot parameters, correcting the patient's posture, and monitoring the patient.

1 is a block diagram of a rehabilitation training system using a rehabilitation robot according to an embodiment of the present invention.
2 is a diagram illustrating a position of a sensor attached to a trainee in a rehabilitation training system using a rehabilitation robot according to an embodiment of the present invention.
3 is an operation flowchart of a rehabilitation training system using a rehabilitation robot according to an embodiment of the present invention.
4 is a view for explaining the operation of the coaching unit according to an embodiment of the present invention.

Then, with reference to the accompanying drawings, the embodiments of the present invention will be described in detail so that those of ordinary skill in the art can easily carry out the present invention.

1 is a block diagram of a rehabilitation training system using a rehabilitation robot according to an embodiment of the present invention.

Referring to FIG. 1 , the rehabilitation training system according to the present invention may include a virtual coaching system 100 and a rehabilitation robot apparatus 200 .

The rehabilitation robot device 200 is a robot device that can help the patient 20 to perform postures or motions required for rehabilitation training such as lower extremity rehabilitation training, upper extremity rehabilitation training, balance training, and gait rehabilitation training. The rehabilitation robot apparatus 200 may be implemented in various forms according to the purpose of rehabilitation training.

The patient 20 may board the rehabilitation robot apparatus 200 or may perform rehabilitation training while wearing the rehabilitation robot apparatus 200 . The patient 20 may perform rehabilitation training while interacting with the rehabilitation robot apparatus 200 . The therapist 10 may coach the patient 10 and control the operation of the rehabilitation robot apparatus 200 using the coaching information and recommended robot parameters provided by the virtual coaching system 100 .

According to an embodiment, the role of the therapist 10 may also be performed by an automated device rather than a human. For example, if the therapist 10 is a device equipped with a visual information output means (such as a monitor) and an audio information output means (such as a speaker), coaching information may be output through the corresponding means. Meanwhile, the therapist 10 may directly apply the recommended robot parameters through communication with the rehabilitation robot apparatus 200 . Of course, if the therapist 10 is provided with a physical means (eg, a robot arm) capable of changing the robot parameters of the rehabilitation robot apparatus 200, the robot parameters of the rehabilitation robot apparatus 200 may be changed through the physical means.

The rehabilitation robot apparatus 200 may perform an operation corresponding to the set robot parameter. The therapist 10 may control the rehabilitation robot apparatus 200 to properly assist the patient 20 during rehabilitation training by adjusting the robot parameters. For example, if the rehabilitation robot apparatus 200 is a lower extremity rehabilitation training robot apparatus, it may receive robot parameters such as walking speed and stride length and operate accordingly.

The virtual coaching system 100 may include a sensing unit 110 , a classification unit 130 , and a coaching unit 150 .

The sensing unit 110 includes a surface electromyography (sEMG) sensor, a Galvanic Skin Response (GSR) sensor, and an electrocardiogram (ECG) sensor for measuring the biosignal data of the patient 20 . can do. In addition, the sensing unit 110 may include at least one of an inertial measurement sensor (IMU), an electrogoniometer, and a vision sensor for measuring motion data of the patient 20 undergoing rehabilitation training. The vision sensor may include an RGB camera, an infrared camera, and the like.

2 is a diagram illustrating a position of a sensor attached to a trainee in a rehabilitation training system using a rehabilitation robot according to an embodiment of the present invention.

The sensing unit 110 may detect biosignal data and motion data of the patient 20 who is training using the rehabilitation robot apparatus 200 . To this end, the sensing unit 110 may attach various sensors to various positions of the patient 20 as illustrated in FIG. 2 . Referring to FIG. 2 , the inertial measurement sensor (Head IMU) is the back part of the patient's head, the inertial measurement sensor (Sacrum IMU) is the sacrum part at the waist, and the inertial measurement sensor (Left/Right Thigh IMU) is the left/right calf part , shows an example in which inertial measurement sensors (Left/Right Shank IMU) are attached to the left/right shins. Four IMUs attached to the left/right calf area and left/right shin area can measure the joint angle. In addition to inertial measurement sensors, EMG, ECG, and plantar contact sensors are attached to detect biosignal data and motion data.

Of course, it is also possible to attach the sensor to a position different from that illustrated in FIG. 2 or to use a different type of sensor.

The classification unit 130 may receive the biosignal data and motion data sensed by the sensing unit 110, and output a result of classifying the patient's rehabilitation training state and motion using a pre-trained classification algorithm.

The classification algorithm may be trained using learning data including trainer biosignal data, trainee motion data, and coaching information of a therapist for the trainee collected in a rehabilitation training process using the rehabilitation robot apparatus 200 .

For example, the classification algorithm may be trained using learning data collected while performing rehabilitation training using the rehabilitation robot apparatus 200 for a plurality of trainees during a predetermined clinical trial period.

And in the process of collecting the learning data, the trainee may be a patient who needs rehabilitation training, but the learning data may not be enough for the patient alone. Therefore, data collected in the process of virtual rehabilitative training for a healthy person who does not need rehabilitation training may also be included in the learning data.

The classification algorithm may use a neural network suitable for processing time series data such as a Long Short Term Memory Network (LSTM) or a Gated Recurrent Unit (GRU).

As the learning data, coaching information made by a therapist in the rehabilitation training process may be labeled with the trainee bio-signal data and trainee motion data collected during rehabilitation training. And, considering that the training data is generally recommended by a professional therapist after observing 2-3 cadences (about 3 to 5 seconds), the basic unit time of the training data is set to 5 seconds and the training status ( normal or abnormal), etc. Of course, the basic unit time of the training data may vary according to embodiments.

For example, during rehabilitation training for a stroke patient, the therapist may coach the patient to correct motion. When coaching related to motion correction such as 1) straightening the head 2) straightening the chest 3) centering the body 4) bending the knee 5) straightening the knee during rehabilitation training, the coaching information can be labeled on the corresponding trainee's motion data .

Meanwhile, when coaching information such as stopping rehabilitation training due to occurrence of an emergency situation of a trainee during rehabilitation training is input, coaching information may be labeled on the trainee's biosignal data corresponding thereto.

Of course, it is also possible to label all of the coaching information without distinguishing between the trainee's biosignal data and the trainee's motion data.

Meanwhile, according to an embodiment, data obtained by adjusting the robot parameters of the rehabilitation robot apparatus 200 while the therapist is performing rehabilitation training may be labeled and included in the learning data.

Since the therapist adjusts the robot parameters when it is determined that it is necessary to adjust the rehabilitation training intensity during the rehabilitation training process, it is also desirable to match the trainee's biosignal data and the trainee's motion data so that they can learn together.

According to an embodiment, the classification unit 130 may include a plurality of classification algorithms each separately learned for each type of coaching. The type of coaching can be broadly divided into motion correction correction for patient rehabilitation training motion correction and training suspension coaching to stop rehabilitation training in case of emergency. Also, motion correction coaching can be divided into detailed movements such as 1) straightening the head, 2) straightening the chest, 3) aligning the center of the body, 4) bending the knee, and 5) straightening the knee. Of course, training suspension coaching may also be classified by type of emergency.

In addition, the classification unit 130 may extract and selectively input features required for each classification algorithm from the trainee bio-signal data and the trainee motion data input in real time. The classification unit 130 can process faster and more accurately in real time because classification algorithms are provided in parallel for each coaching type to determine whether coaching is required.

For example, 1) for head straightening coaching, head inertia measurement sensor data (Head IMU data), head absolute and joint angle, etc. are input to the classification algorithm as features, 4) for knee bending coaching Left/right hip and knee absolute and joint angles, etc. may be input to a corresponding classification algorithm as features to determine whether the trainee's corresponding motion is normal and output.

The coaching unit 150 may provide coaching based on a result of classifying the rehabilitation training state and motion of the patient 20 by the classification unit 130 . Here, the rehabilitation training state may correspond to a result of classifying whether the corresponding motion is a normal motion or a pathological motion.

For example, if it is a result of classifying the rehabilitation training state and motion of the patient 20 and the patient needs to straighten the head, the coaching unit 150 may provide 'head straightening' for motion coaching for the patient.

In addition, if an emergency situation occurs as a result of classifying the patient's rehabilitation training state and motion, the coaching unit 150 may provide coaching to stop the operation of the rehabilitation robot apparatus 200 .

The coaching unit 150 may receive a currently set robot parameter from the rehabilitation robot apparatus 200 . In addition, when the patient's training intensity adjustment is required as a result of classifying the rehabilitation training state and motion of the patient 20 , the coaching unit 150 may provide a recommended robot parameter for changing the motion of the rehabilitation robot apparatus 200 . For example, when it is necessary to increase or decrease the stride length or the walking speed according to the currently set robot parameter, the recommended robot parameter may be provided.

On the other hand, since the patient's biosignal data and motion data are time series data, it may be difficult to clearly distinguish between normal motion and abnormal motion. It can be implemented that a certain amount of window masking is applied to the LSTM prediction data, and the result is finally determined by checking whether a certain threshold is exceeded. That is, since the LSTM neural network divides time-series data in units of windows and processes it, if the result classified as an abnormal state for a certain period of time exceeds a threshold, it can be determined that coaching for the abnormal behavior is necessary.

3 is an operation flowchart of a rehabilitation training system using a rehabilitation robot according to an embodiment of the present invention.

Referring to FIGS. 1 and 3 , first, training data including bio-signal data of a trainee collected in a rehabilitation training process using the rehabilitation robot device 200 such as a clinical test, movement data of a trainee, and coaching information of a therapist for the trainee are used. to train the classification algorithm (S310). In step S310, the training data may be labeled with coaching information made by a therapist in the rehabilitation training process in the trainee bio-signal data and trainee motion data collected during rehabilitation training. In addition, when coaching information such as suspension of rehabilitation training due to occurrence of an emergency situation of a trainee during rehabilitation training is input, coaching information may be labeled on the trainee bio-signal data corresponding thereto. Of course, it is also possible to label all of the coaching information without distinguishing between the trainee's biosignal data and the trainee's motion data. Meanwhile, according to an embodiment, data obtained by adjusting the robot parameters of the rehabilitation robot apparatus 200 while the therapist performs rehabilitation training may also be labeled and included in the learning data. On the other hand, it is also possible to train a classification algorithm for each motion that requires coaching.

Thereafter, the classification unit 130 receives the biosignal data and motion data sensed by the sensing unit 110 while the patient 20 is actually performing rehabilitation training using the rehabilitation robot apparatus 200, and performs pre-trained classification. The result of classifying the rehabilitation training state and operation of the patient 20 may be output using the algorithm (S320).

Next, the coaching unit 150 may provide coaching based on a result of classifying the rehabilitation training state and motion of the patient 20 by the classification unit 130 ( S330 ). In step S330 , the coaching unit 150 receives the currently set robot parameters from the rehabilitation robot device 200 , and as a result of classifying the rehabilitation training state and motion of the patient 20 , the rehabilitation training intensity of the patient 20 is adjusted. If necessary, a recommended robot parameter for changing the operation of the rehabilitation robot apparatus 200 may be provided.

According to an embodiment, the coaching unit 150 may operate as illustrated in FIG. 4 to output coaching information according to the tendency of the therapist 10 .

4 is a view for explaining the operation of the coaching unit according to an embodiment of the present invention.

)와 치료사 성향에 따라 설정되는 값(

,

,

)을 입력으로 코칭 정보를 출력할 수 있다.Referring to FIG. 4 , the coaching unit 150 displays the results of the rehabilitation training state and motion classification output from the classification unit 130 (

) and the value set according to the therapist’s disposition (

,

,

) can be input to output coaching information.

)(the level of activity of physical therapists)은 치료사가 코칭을 얼마나 자주하는 지를 나타내며, 치료사의 성향에 따라 설정될 수 있다. 코칭 확률 마진(

)(coaching probability margin)은 코칭 판단이 유보되는 정도를 나타내는 것으로, 치료사의 성향에 따라 설정될 수 있으며, 직전의 코칭 결과에 비례하여 기준값이 높아지도록 설계될 수 있다. 코칭 유보 시간(

)(and coaching hold time)은 코칭을 위한 최소 간격, 코칭 출력을 내보내고 나서 최소한으로 환자의 동작을 관찰하는 시간을 의미하며, 이 역시 치료사의 성향에 설정될 수 있다.The therapist's level of activity (

) (the level of activity of physical therapists) indicates how often the therapist coaches, and can be set according to the therapist's disposition. Coaching odds margin (

) (coaching probability margin) indicates the degree to which coaching judgment is withheld, and can be set according to the therapist's disposition, and can be designed to increase the reference value in proportion to the previous coaching result. Coaching Reservation Time (

) (and coaching hold time) means the minimum interval for coaching, the minimum amount of time to observe the patient's motion after outputting the coaching output, and this can also be set according to the therapist's disposition.

In the following, i is an index indicating a coaching type selected by the classification unit 130, and k indicates a current time index.

)은 아래 수학식에 의해 현재 코칭 확률(

)과 코칭 기준값(

)을 비교하여 결정된다.Final Coaching Recommendation (

) is the current coaching probability (

) and the coaching reference value (

) is determined by comparing

)은 아래 수학식 1에 의해 계산된다.Current coaching odds (

) is calculated by Equation 1 below.

[Equation 1]

는 분류부(130)의 출력, 즉 재활 훈련 상태 및 동작 분류에 대한 결과를 나타내는 것이다. Nw는 코칭 유보 시간(

)을 고려하여 설정되는 값으로, 예컨대 훈련자의 동작 데이터가 1초에 50번(즉 50Hz)로 샘플링 되고, 치료사가 5초 동안 훈련자의 동작 상태를 살펴보는 것으로 고려하면, 250(50Hz*5sec)로 설정될 수 있다.here

is the output of the classification unit 130, that is, indicates the result of the rehabilitation training state and motion classification. Nw is the coaching retention time (

), for example, when the trainee's motion data is sampled 50 times per second (that is, 50Hz) and the therapist examines the trainee's motion state for 5 seconds, 250 (50Hz*5sec) can be set to

)은 아래 수학식 2에 의해 도출된다.Coaching standards (

) is derived by Equation 2 below.

[Equation 2]

는 치료사의 활동 수준으로 설정된 값을 나타낸 것이다. 여기서

는

로서, 치료사의 활동 수준에 비례하여 코칭 기준값(

)을 낮추는데 사용되는 함수이다.here,

is the value set as the therapist's activity level. here

Is

As , the coaching reference value (

) is a function used to lower

)은 아래 수학식 3에 의해 결정된다.Final Coaching Recommendation (

) is determined by Equation 3 below.

[Equation 3]

)이 현재 코칭 확률(

) 이상이면 코칭 출력을 하지 않고, 코칭 기준값(

)이 현재 코칭 확률(

)보다 작으면 코칭 출력을 한다.where m is the index used to reserve the final coaching recommendation for the coaching reservation time, f _s is the sampling period of trainee motion data, and τ _i is the coaching reservation time. That is, the coaching reference value (

) is the current coaching probability (

) or higher, coaching output is not performed, and the coaching standard value (

) is the current coaching probability (

) is smaller than the coaching output.

An embodiment of the present invention includes a computer-readable medium including program instructions for performing various computer-implemented operations. This medium records a program for executing the method described above. The medium may contain program instructions, data files, data structures, etc. alone or in combination. Examples of such media include hard disks, magnetic media such as floppy disks and magnetic tapes, optical recording media such as CDs and DVDs, optical disks and magneto-optical media, and program instructions such as ROM, RAM, flash memory, etc. hardware devices configured to store and perform Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

Although preferred embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concept of the present invention as defined in the following claims are also provided. is within the scope of the

Claims (21)

A sensing unit that detects bio-signal data and motion data of a patient undergoing rehabilitation training using a rehabilitation robot device;
A classification unit that receives the biosignal data and motion data sensed by the sensing unit, and classifies the patient's rehabilitation training state and motion using a pre-trained classification algorithm; and
A coaching unit that provides coaching based on a result of classifying the patient's rehabilitation training state and motion
Rehabilitation training system using a rehabilitation robot comprising a. In claim 1,
The classification algorithm is
Rehabilitation training system using a rehabilitation robot that is trained using the training data including the trainer bio-signal data, the trainee motion data, and the coaching information of the therapist for the trainee collected in the rehabilitation training process using the rehabilitation robot device. In claim 2,
The learning data is
A rehabilitation training system using a rehabilitation robot in which coaching information of a therapist is labeled in response to the trainer's bio-signal data and the trainee's motion data. In claim 3,
The learning data is a rehabilitation training system using a rehabilitation robot that is collected during a predetermined clinical trial. In claim 2,
The coaching department
As a result of classifying the patient's rehabilitation training state and motion, a value set according to the therapist's propensity is received, and coaching information is output according to the result of comparing the coaching reference value calculated according to a predetermined equation with the current coaching probability. Rehabilitation training system using a rehabilitation robot to determine. In claim 2,
The coaching department
A robot parameter for controlling the operation of the rehabilitation robot apparatus is received, and a recommended robot parameter for changing the operation of the rehabilitation robot apparatus is provided using the input robot parameter and the result of classifying the patient's rehabilitation training state and operation Rehabilitation training system using rehabilitation robot. In claim 6,
The coaching department
A rehabilitation training system using a rehabilitation robot that provides coaching for the classification result when a result of dividing the patient's biosignal data and motion data into a predetermined window size is greater than or equal to a predetermined threshold. In claim 1,
The classification unit,
Includes a plurality of classification algorithms trained for each type of coaching,
A rehabilitation training system using a rehabilitation robot for selecting and inputting necessary data for each of the plurality of classification algorithms from the biosignal data and motion data sensed by the sensing unit. In claim 8,
The type of coaching is
Rehabilitation training system using a rehabilitation robot, including a plurality of motion correction coaching separated by each rehabilitation training motion and training stop coaching to stop the rehabilitation training in case of an emergency. In claim 1,
The sensing unit,
At least one of a surface electromyography sensor, a current skin response sensor, and an electrocardiogram sensor for measuring the patient's biosignal data,
Rehabilitation training system using a rehabilitation robot comprising at least one of an inertial measurement sensor, an electronic goniometer, and a vision sensor for measuring the motion data of the patient. Detecting bio-signal data and motion data of a patient undergoing rehabilitation training using a rehabilitation robot device;
receiving the detected biosignal data and motion data, and classifying the patient's rehabilitation training state and motion using a pre-trained classification algorithm; and
providing coaching based on a result of classifying the patient's rehabilitation training state and motion
Rehabilitation training method using a rehabilitation robot comprising a. In claim 11,
The classification algorithm is
Rehabilitation training method using a rehabilitation robot that is trained using the training data including the trainer bio-signal data, the trainee motion data, and the coaching information of the therapist for the trainee collected in the rehabilitation training process using the rehabilitation robot device. In claim 12,
The learning data is
A rehabilitation training method using a rehabilitation robot in which coaching information of a therapist is labeled in response to the trainer's bio-signal data and the trainee's motion data. In claim 13,
The learning data is a rehabilitation training method using a rehabilitation robot that is collected during a predetermined clinical trial. In claim 12,
As a result of classifying the patient's rehabilitation training state and motion, a value set according to the therapist's propensity is received, and coaching information is output according to the result of comparing the coaching reference value calculated according to a predetermined equation with the current coaching probability. Rehabilitation training system using a rehabilitation robot to determine. In claim 12,
A robot parameter for controlling the operation of the rehabilitation robot apparatus is received, and a recommended robot parameter for changing the operation of the rehabilitation robot apparatus is provided using the input robot parameter and the result of classifying the patient's rehabilitation training state and operation Rehabilitation training method using a rehabilitation robot. 17. In claim 16,
The coaching department
Rehabilitation training method using a rehabilitation robot that provides coaching for the classification result when a result of dividing the patient's biosignal data and motion data into a predetermined window size is greater than or equal to a predetermined threshold. In claim 11,
Includes a plurality of classification algorithms trained for each type of coaching,
A rehabilitation training method using a rehabilitation robot for selecting and inputting necessary data for each of the plurality of classification algorithms from the biosignal data and motion data sensed by the sensing unit. In claim 18,
The type of coaching is
A rehabilitation training method using a rehabilitation robot, comprising a plurality of motion correction coaching classified for each rehabilitation training motion and training stop coaching to stop the rehabilitation training in case of an emergency. In claim 11,
The patient's biosignal data is sensed by at least one of a surface electromyography sensor, a current skin response sensor, and an electrocardiogram sensor,
The patient's motion data is a rehabilitation training method using a rehabilitation robot that is sensed by at least one of an inertial measurement sensor, an electronic goniometer, and a vision sensor. A computer-readable recording medium in which a program for executing any one of the methods according to any one of claims 11 to 20 is recorded on a computer.

Images