WO2022265153A1

WO2022265153A1 - Contactless coaching system using artificial intelligence

Info

Publication number: WO2022265153A1
Application number: PCT/KR2021/010927
Authority: WO
Inventors: 유제광
Original assignee: 동국대학교 산학협력단
Priority date: 2021-06-15
Filing date: 2021-08-18
Publication date: 2022-12-22
Also published as: KR102638918B1; KR20220168242A

Abstract

Disclosed is a contactless coaching system using artificial intelligence. An aspect of the present invention may provide a contactless coaching system using artificial intelligence, the system comprising: a motion capture unit for capturing a trainee's motions so as to collect motion information; a biometric signal collecting unit for sensing the trainee's biometric signals so as to collect biometric signal information; an analysis unit for determining motion errors and psychological conditions by means of a machine learning algorithm on the basis of the motion information and the biometric signal information so as to provide one of a technical solution and a psychological solution, the machine learning algorithm conducting reinforced learning by using expert evaluations as rewards; and a feedback unit for delivering the technical or psychological solution provided by the analysis unit to the trainee. The analysis unit comprises: a motion error detecting unit for detecting motion errors on the basis of the motion information; a first determination unit for determining a psychological condition on the basis of the biometric signal information if the motion error detection unit detects a motion error; a second determination unit for determining whether the cause of the motion error detected by the motion error detecting unit is a technical problem or a psychological problem on the basis of the psychological condition determined by the first determination unit; and a solution providing unit for providing a technical solution when the second determination unit has determined that the cause of the motion error is a technical problem, and providing a psychological solution when the second determination unit has determined that the cause of the motion error is a psychological problem.

Description

Non-face-to-face coaching system using artificial intelligence

The present invention relates to a non-face-to-face coaching system using artificial intelligence.

In general, it is recognized that motion evaluation and correction of sports players are possible only through direct face-to-face meetings with coaches. However, as face-to-face coaching has become virtually difficult due to the spread of the COVID-19 virus and motion capture technology has been commercialized, the need for and technical skills to develop non-face-to-face coaching technology has been met. In addition, since motions of sports players are made in almost constant patterns, motion evaluation and correction may be performed through machine learning.

Meanwhile, the cause of the operation error may be due to a psychological problem as well as a technical problem.

However, most of the conventional non-face-to-face coaching systems only provide technical solutions for finding and correcting motion errors, and could not confirm whether the cause of the motion error was due to a psychological problem. There was a problem that couldn't provide a solution. These problems can be applied to rehabilitation training for the disabled as well.

An embodiment of the present invention provides a non-face-to-face coaching system using artificial intelligence that detects a trainee's motion error, determines the cause of the motion error, and selects and provides a suitable solution.

According to one aspect of the present invention, a motion capture unit for capturing motion of a trainee and collecting motion information; a bio-signal collection unit that senses the bio-signal of the trainee and collects bio-signal information; Based on the motion information and the biosignal information, a motion error and a psychological state are determined by a machine learning algorithm to provide one of a technical solution and a psychological solution, wherein the machine learning algorithm proceeds with reinforcement learning as a reward for expert evaluation analysis unit; and a feedback unit for delivering a technical solution or psychological solution provided by the analysis unit to the trainee, wherein the analysis unit includes: a motion error detection unit for detecting a motion error based on the motion information; a first determination unit for determining a psychological state based on the biosignal information when a motion error is detected by the motion error detection unit; a second determination unit for determining whether the cause of the motion error detected by the motion error detection unit is a technical problem or a psychological problem based on the psychological state determined by the first determination unit; and a solution providing unit providing a technical solution if the second determination unit determines that the cause of the operation error is a technical problem, and providing a psychological solution if the second determination unit determines that the cause of the operation error is a psychological problem. A non-face-to-face coaching system using artificial intelligence may be provided.

The bio-signal collection unit may sense the trainee's heart rate, respiration, brain waves, eye movement, eye blinking, and electromyogram.

The first determination unit determines the psychological state for a set time before the motion error is detected based on the biosignal information collected during the set time before the motion error is detected, and the second determination unit determines the psychological state before the motion error is detected. The cause of the motion error may be determined based on the mental state for the previously set time period.

If the first determination unit determines that the trainee's psychological state is neither overexcited nor anxious, the second determination unit may determine that the cause of the motion error is a technical problem.

If the first determination unit determines that the trainee's psychological state is neither excessively excited nor anxious, the solution providing unit may provide a technical solution for correcting the motion error.

If the first determination unit determines that the trainee's psychological state is overexcited, the second determination unit may determine that the cause of the motion error is a psychological problem.

If the first determination unit determines that the trainee's psychological state is in an overexcited state, the solution provider performs biofeedback, gradual relaxation training, autogenous training, systematic desensitization, breathing control, cognitive restructuring, and thought suspension as psychological solutions. Information on at least one can be provided.

If the first determination unit determines that the trainee's psychological state is in an anxious state, the second determination unit may determine that the cause of the motion error is a psychological problem.

If the first determination unit determines that the trainee's psychological state is in an anxious state, the solution providing unit provides psychological solutions such as gradual relaxation training, short-term relaxation training, spontaneous training, desensitization training, transcendental meditation, biorepatriation, and breath control. It can provide information on at least one of , cognitive reconstruction, and thought suspension.

The feedback unit may also provide a motion image captured by the motion capture unit when providing a technical solution or a psychological solution.

When the feedback unit provides a technical solution or a psychological solution, the analysis unit may provide information on the motion error and psychological state determined by the analysis unit.

According to the embodiment of the present invention, it is possible to select and provide either a technical solution or a psychological solution by determining the cause of the motion error rather than just detecting the trainee's motion error, thereby managing the trainee's emotional aspect as well. can do.

In addition, since motion evaluation and correction of trainees are performed through machine learning of a reinforcement learning method, feedback efficiency can be improved, especially when applied to the field of sports in which motions of trainees have almost constant patterns.

1 is a diagram of a non-face-to-face coaching system using artificial intelligence according to an embodiment of the present invention;

FIG. 2 is a diagram illustrating a motion capture image by the motion capture unit of FIG. 1 .

Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

Terms used in the embodiments of the present invention may be interpreted as meanings that can be generally understood by those of ordinary skill in the art to which the present invention belongs, unless clearly defined otherwise, and only specific The examples are intended to be illustrative and are not intended to limit the present invention.

In this specification, the singular form will be considered to include the plural form as well, unless otherwise specified.

In addition, when a part is described as "including" a certain component, it means that the corresponding part may further include other components.

In addition, when a component is described as "on", it means above or below the corresponding component, and does not necessarily mean that it is located on the upper side relative to the direction of gravity.

Also, when a component is described as being “connected” or “coupled” to another component, that component is not only directly connected to or coupled to another component, but also indirectly through another component. It may also include the case of being connected or combined with.

In addition, although terms such as first and second may be used to describe a certain component, these terms are only used to distinguish the corresponding component from other components, and the essence or sequence of the corresponding component is determined by the term. or order, etc., is not intended to be limiting.

1 is a diagram of a non-face-to-face coaching system using artificial intelligence according to an embodiment of the present invention, and FIG. 2 is a diagram showing a motion captured image by the motion capture unit of FIG. 1 .

1 and 2, the non-face-to-face coaching system 10 using artificial intelligence according to an embodiment of the present invention includes a motion capture unit 100, a biosignal collection unit 200, an analysis unit 300, and A feedback unit 400 may be included.

The motion capture unit 100 may collect information about the trainee's motion by capturing the trainee's motion, for example, a sports motion of a sports player or a rehabilitation training motion of a disabled person.

For example, the motion capture unit 100 may use a marker-based motion capture technology, and for this purpose may include a marker attached to the trainee's body and an infrared camera for capturing the trainee's motion. there is.

As another example, the motion capture unit 100 may use a markerless motion capture technology, and may include a depth camera for capturing a motion of a trainee for this purpose.

The motion information collected by the motion capture unit 100 may include the position, direction, angle, speed, acceleration, angular velocity, and angular acceleration of each part of the trainee's body.

The bio-signal collection unit 200 may collect bio-signal information by sensing bio-signals of trainees.

For example, the bio-signal information collected by the bio-signal collection unit 200 may include information about a trainee's heart rate, respiration, brainwave, eyeball movement, blinking, and electromyogram. To this end, the biosignal collection unit 200 includes an electrocardiogram signal sensor (ECG signal), an brain wave signal sensor (EEG signal), an eye movement signal sensor (EOG signal), an electromyogram signal sensor (EMG signal), an eye tracking sensor, and the like. can do.

The analysis unit 300 detects a motion error by a machine learning algorithm based on the motion information collected by the motion capture unit 100 and the bio signal information collected by the bio signal collection unit 200, while determining the psychological state Either a technical solution or a psychological solution is provided, but the machine learning algorithm can proceed with reinforcement learning as a reward for expert evaluation.

For example, a machine learning algorithm can initially learn based on expert judgment results, and after learning has progressed to a certain extent, reinforcement learning can be performed as a reward for expert evaluation of the machine learning algorithm’s judgment and solution provision results. there is.

Specifically, the analysis unit 300 may include an operation error detection unit 310, a first determination unit 320, a second determination unit 330 and a solution providing unit 340, and may further include an input unit 350. may also include

The motion error detection unit 310 may detect an error in the motion of the trainee, that is, the motion, by a machine learning algorithm based on the motion information collected by the motion capture unit 100 . For example, when a trainee kicks a ball, if the leg angle, speed, hitting point, etc. are out of the normal range, the trainee's motion may be detected as having a motion error, and the degree of error may be qualitatively calculated. Here, the normal range may be set by a machine learning algorithm. For example, a machine learning algorithm collects a lot of motion information (eg, leg angle, speed, hitting point, etc.) when a trainee performs the corresponding motion normally, and calculates the average (A) and standard deviation (SD). Later, a range that does not deviate from the average by a predetermined multiple of a standard deviation, for example, a range from A-2SD to A+2SD may be set as a normal range. This principle can also be applied to setting the normal range of bio-signal information, which is a criterion for determining a psychological state.

When a motion error is detected by the motion error detection unit 310, the first determination unit 320 can determine the trainee's psychological state by a machine learning algorithm based on the bio signal information collected by the bio signal collection unit 200. there is.

For example, the first determination unit 320 may determine that the trainee's psychological state is in an overexcited state when the trainee's heart rate, respiration, and EMG are out of normal ranges.

As another example, the first determination unit 320 may determine that the trainee's psychological state is in an anxious state when the trainee's heart rate, respiration, brain wave, eye movement, eye blinking, and EMG are out of normal ranges.

The second determination unit 330 may determine which one of a technical problem and a psychological problem is the cause of the motion error detected by the motion error detection unit 310 according to the psychological state determined by the first determination unit 320 .

In particular, the first determination unit 320 determines the bio-signal information collected by the bio-signal collection unit 200 for a set period of time before a motion error is detected by the motion error detection unit 310, that is, before the trainee performs an erroneous motion. Based on the above, the mental state for a set time before the motion error is detected, that is, the mental state for a set time before the trainee performs the erroneous motion, can be determined, and the second determination unit 330 makes the first judgment. Based on the psychological state for a set time before the motion error determined by the unit 320 is detected, it is possible to determine which one of the technical problem and the psychological problem is the cause of the motion error detected by the motion error detection unit 310. Here, the time may be set to several seconds, for example, 0 to 10 seconds, because a time difference between an unstable mental state and an operation error may occur.

Therefore, a psychological problem may follow after an operation error due to a technical problem. In this case, the possibility of providing a psychological solution rather than a technical solution by misdiagnosing the cause of the operation error as a psychological problem can be blocked.

Such psychological problems can be naturally solved by solving technical problems that cause motion errors.

For example, the second determination unit 330 may determine that the cause of the motion error is a psychological problem when the first determination unit 320 determines that the trainee's psychological state is an excessive excitement state or an anxiety state.

As another example, the second determination unit 330 may determine that the cause of the motion error is a technical problem when the first determination unit 320 determines that the trainee's psychological state is neither excessively excited nor anxious.

The solution provider 340 provides a technical solution if the second determination unit 330 determines that the cause of the operation error is a technical problem, but if the second determination unit 330 determines that the cause of the operation error is a psychological problem, the solution providing unit 340 provides a technical solution. can provide a solution.

For example, if the solution provider 340 determines that the trainee's psychological state is overexcited in the first determination unit 320 and determines that the cause of the motion error is a psychological problem in the second determination unit 330, the trainee is Information on at least one of biofeedback, gradual relaxation training, autogenous training, systematic desensitization, breathing control, cognitive restructuring, and thought suspension can be provided to help prevent overexcitement in the corresponding motion. .

As another example, the solution providing unit 340 determines that the trainee's psychological state is in an anxious state in the first determination unit 320 and determines that the cause of the motion error is a psychological problem in the second determination unit 330, the trainee is corresponding Psychological solutions that help not to fall into a state of anxiety in motion, such as gradual relaxation training, short-term relaxation training, autogenous training, numbing training, transcendental meditation, biorepatriation, breathing control, cognitive restructuring, and thinking stop for at least one can provide information.

As another example, the solution provider 340 determines in the first determination unit 320 that the trainee's psychological state is neither excessively excited nor anxious, and determines in the second determination unit 330 that the cause of the operation error is technical If it is determined that the problem is a trainee's lack of proficiency, etc., a technical solution can be provided to correct the trainee's motion error.

Meanwhile, a technical solution or a psychological solution may be provided in various modalities such as text, voice, sound (music), video, and image.

On the other hand, biofeedback or biological repatriation may mean a technique or training that induces a trainee to consciously control a body function by providing a biosignal to a trainee in real time.

Gradual relaxation training or short-term relaxation training may refer to training in which trainees completely relax their body within a short period of time by repeating tension and relaxation for each body part.

Autogenous training is similar to progressive relaxation training, but can refer to training that focuses on how to feel a body part rather than how to relax it.

Systematic desensitization or desensitization training may refer to a method of training to become persistently insensitive to anxiety or stress by showing a relaxation response instead of an anxiety response to a stimulus that causes anxiety or stress.

Breathing control is a training that focuses all attention on breathing and slowly and deliberately repeats the process of inhaling and exhaling. Examples include diaphragmatic breathing, loud exhalation, rhythmic breathing, 1:2 ratio breathing, and mindful breathing.

Cognitive restructuring can refer to training in replacing negative thoughts with positive ones.

Thought stopping can refer to the practice of stopping negative thoughts from progressing by consciously telling yourself to stop when they arise.

Transcendental meditation can free the village by concentrating thoughts and emotions inside the body, and at this time, it can mean training to concentrate the mind using specific words or images.

The solution provider 340 may include a database in which technical solutions and psychological solutions are stored. In this case, technical solutions and psychological solutions may be stored in the database by being classified according to the type of motion error and psychological state, and may be supplemented, for example corrected, deleted, or added by a machine learning algorithm of a reinforcement learning method. there is.

The input unit 350 may receive an expert evaluation of the result of the decision of the first determination unit 320 or the second determination unit 330 and the solution providing result of the solution providing unit 340 .

On the other hand, each component of the analysis unit 300, for example, the operation error detection unit 310, the first determination unit 320, the second determination unit 330, the solution providing unit 340, etc. are controlled by the artificial intelligence server. can be implemented

The feedback unit 400 may deliver the technical solution or psychological solution provided by the analysis unit 300 to the trainee.

For example, the feedback unit 400 may include a display device that displays a motion image captured by the motion capture unit 100 together with a technical solution or a psychological solution.

At this time, a video call may be made between the expert and the trainee through the display screen of the display device.

In addition, the feedback unit 400 may also provide information on the motion error determined by the analysis unit 300 and the psychological state when providing a technical solution or a psychological solution.

Although the above has been described with a focus on preferred embodiments of the present invention, this is only an example and does not limit the present invention. Those skilled in the art to which the present invention pertains can modify and change the embodiments in various ways by adding, changing, deleting, or adding components within the scope that does not deviate from the spirit of the present invention described in the claims. It will be possible, and this will also be said to be included within the scope of the present invention.

Claims

a motion capture unit that collects motion information by capturing a motion of a trainee;

a bio-signal collection unit that senses the bio-signal of the trainee and collects bio-signal information;

Based on the motion information and the biosignal information, a motion error and a psychological state are determined by a machine learning algorithm to provide one of a technical solution and a psychological solution, wherein the machine learning algorithm proceeds with reinforcement learning as a reward for expert evaluation analysis unit; and

A feedback unit for delivering the technical solution or psychological solution provided by the analysis unit to the trainee;

The analysis unit,

an operation error detection unit for detecting an operation error based on the operation information;

a first determination unit for determining a psychological state based on the biosignal information when a motion error is detected by the motion error detection unit;

a second determination unit for determining whether the cause of the motion error detected by the motion error detection unit is a technical problem or a psychological problem based on the psychological state determined by the first determination unit; and

and a solution provider providing a technical solution if the second determination unit determines that the cause of the operation error is a technical problem, and providing a psychological solution if the second determination unit determines that the cause of the operation error is a psychological problem. A non-face-to-face coaching system using artificial intelligence.
According to claim 1,

The biosignal collection unit is a non-face-to-face coaching system using artificial intelligence, characterized in that for sensing the trainee's heart rate, breathing, brain waves, eye movements, eye blinks and electromyography.
According to claim 1,

The first determination unit determines a psychological state for a set time before a motion error is detected based on biosignal information collected during a set time before a motion error is detected;

The second determination unit non-face-to-face coaching system using artificial intelligence, characterized in that for determining the cause of the motion error based on the psychological state for a set time before the motion error is detected.
According to claim 3,

If the first determination unit determines that the trainee's psychological state is neither excessively excited nor anxious, the second determination unit determines that the cause of the motion error is a technical problem. Characterized in that non-face-to-face coaching using artificial intelligence system.
According to claim 4,

If the first determination unit determines that the trainee's psychological state is neither excessive excitement nor anxiety, the solution providing unit provides a technical solution for correcting motion errors. coaching system.
According to claim 4,

If the first determination unit determines that the trainee's psychological state is in an overexcited state, the second determination unit determines that the cause of the motion error is a psychological problem.
According to claim 6,

If the first determination unit determines that the trainee's psychological state is in an overexcited state, the solution provider performs biofeedback, gradual relaxation training, autogenous training, systematic desensitization, breathing control, cognitive restructuring, and thought suspension as psychological solutions. Non-face-to-face coaching system using artificial intelligence, characterized in that for providing information on at least one.
According to claim 4,

If the first determination unit determines that the trainee's psychological state is in an anxious state, the second determination unit determines that the cause of the motion error is a psychological problem.
According to claim 8,

If the first determination unit determines that the trainee's psychological state is in an anxious state, the solution providing unit provides psychological solutions such as gradual relaxation training, short-term relaxation training, spontaneous training, desensitization training, transcendental meditation, biorepatriation, and breath control. , cognitive reconstruction, non-face-to-face coaching system using artificial intelligence, characterized in that for providing information on at least one of the suspension of thinking.
According to claim 1,

The feedback unit non-face-to-face coaching system using artificial intelligence, characterized in that for providing together with the motion image captured by the motion capture unit when providing a technical solution or psychological solution.
According to claim 10,

The feedback unit non-face-to-face coaching system using artificial intelligence, characterized in that for providing information on the operational error and psychological state determined by the analysis unit when providing a technical solution or psychological solution.