KR20200144996A - Motor performance coashing system based on muscle synergy library and calculation method for muscle synergy characteristic based on muscle synergy library - Google Patents

Abstract

The present invention relates to an exercise coaching system using a muscle synergy library and a calculation method of muscle synergy characteristics using a muscle synergy library which calculate muscle synergy from a motion of a user to feed back proper exercise states or postures from characteristics of muscle synergy activated for each user motion. According to the present invention, the exercise coaching system comprises: a first analysis unit to recognize a main motion of a user, and compare the main motion of the user with a first library including previously stored motion information to calculate a plurality of detailed motions included in the main motion; a second analysis unit to measure electromyogram data in accordance with the main motion of the user, and calculate muscle synergy which is a muscle group activated during the main motion of the user based on the electromyogram data; and a characteristic extraction unit to extract characteristics of the muscle synergy of the user for each detailed motion based on the detailed motions calculated by the first analysis unit and the muscle synergy calculated by the second analysis unit. According to the present invention, the calculation method of muscle synergy characteristics comprises: a first calculation step of recognizing a main motion of a user, and comparing the main motion of the user with a first library including previously stored motion information to calculate a plurality of detailed motions included in the main motion; a second calculation step of measuring electromyogram data in accordance with the main motion of the user, and calculating muscle synergy which is a muscle group activated during the main motion of the user for each detailed motion based on the electromyogram data; and a characteristic extraction step of extracting characteristics of the muscle synergy by comparing a second library storing muscle synergy of a plurality of users and the muscle synergy for each detailed motion.

Description

Exercise coaching system using muscle synergy library and method of calculating muscle synergy characteristics using muscle synergy library {MOTOR PERFORMANCE COASHING SYSTEM BASED ON MUSCLE SYNERGY LIBRARY AND CALCULATION METHOD FOR MUSCLE SYNERGY CHARACTERISTIC BASED ON MUSCLE SYNERGY LIBRARY}

The present invention relates to an exercise coaching system using a muscle synergy library and a method of calculating muscle synergy characteristics using a muscle synergy library, and more particularly, to recognize a user's motion and classify it into detailed motions, and extract EMG data according to the user's motion. By calculating the muscle synergy consisting of the muscle groups that are activated based on the muscle synergy characteristics and extracting the characteristics of the muscle synergy, the movement of the user can be recognized through the muscle synergy characteristics or the user can check the correct muscle activation pattern according to the movement, so exercise and rehabilitation , An exercise coaching system using a muscle synergy library applicable to the purpose of treatment, and a method of calculating muscle synergy characteristics using a muscle synergy library.

An EMG (Electromyography) signal or an EMG signal means an electrical signal that occurs along a muscle fiber from a muscle surface according to a movement, and is measured using an EMG sensor that responds to the displacement of a potential accompanying muscle activity. The muscle activity measured through the EMG sensor is applied to various application cases such as medical diagnosis, muscle rehabilitation, exercise and training, prosthetic control, and electronic devices.

In recent years, a pattern recognition algorithm is used to analyze electrical data in order to identify a physical gesture from an action performed by a user by applying the above EMG signal.

The pattern recognition algorithm can use various statistical models, among which HMM (Hidden Markov Model) is a tool frequently used in the field of gesture recognition as an advantage for gesture recognition.

Meanwhile, an acceleration sensor, a geomagnetic sensor, a gyro sensor, etc. may be used as a sensor for measuring the movement of an object, and an inertial measurement unit (IMU) may be used.

(Document 0001) KR 10-2017-0091963 A (Document 0002) KR 10-1829356 B

It is an object of the present invention to classify and store muscle synergies that are activated in performing a specific motion for each user, and to calculate muscle synergy characteristics based on this, an exercise coaching system using a muscle synergy library and a muscle synergy using a muscle synergy library. It is to provide a method of calculating characteristics.

In addition, as the muscle synergy characteristics are calculated, the motion coaching system using a muscle synergy library that can be used for the purpose of rehabilitation, exercise, treatment, etc. in reverse by recognizing the user's motion or comparing it with the muscle synergy of multiple users, and using a muscle synergy library It is to provide a method of calculating muscle synergy characteristics.

The exercise coaching system using the muscle synergy library according to the present invention for achieving the above object recognizes the user's main motion and calculates a plurality of detailed motions included in the main motion by comparing it with the first library containing motion information stored in advance. A first analysis unit that measures EMG data according to the user's main motion, and calculates muscle synergy, which is a muscle group activated during the user's main motion, based on the EMG data. It includes a characteristic extracting unit for extracting characteristics of the user's muscle synergy for each detailed movement based on the detailed movement and the muscle synergy calculated by the second analysis unit.

In addition, a display unit for visually displaying the muscle synergy characteristics of the user extracted from the characteristic extracting unit may be further included.

In addition, it may further include a setting unit for obtaining prior information on the user's body and main motion.

In addition, the first analysis unit may recognize the user's main motion by analyzing an image or an image acquired by photographing a user's motion, or analyzing inertia change information acquired through an inertial sensor.

In addition, the second analysis unit includes a sensing module that acquires EMG data, a calculation module that calculates muscle synergy from a group pattern of EMG data activated when the user moves, and a similar among a number of muscle synergies calculated by the calculation module using clustering. It may include a clustering module for grouping muscle synergy.

In addition, the second analysis unit may further include a filtering module that removes dynamic noise included in the EMG signal based on the detailed operation calculated by the first analysis unit.

In addition, the characteristic extracting unit may extract muscle synergy characteristics by comparing the muscle synergy calculated by the second analysis unit with the second library in which the muscle synergies of multiple users for each detailed motion are stored.

The method for calculating muscle synergy characteristics using a muscle synergy library according to the present invention is a first method of recognizing a user's main motion and calculating a plurality of detailed motions included in the main motion by comparing it with a first library including pre-stored motion information. In the calculation step, the EMG data according to the user's main motion is measured, and based on the EMG data, the muscle synergy, which is a muscle group activated during the user's main motion, is calculated for each detailed motion, and the muscle synergy of multiple users is stored. And a characteristic extraction step of extracting characteristics of muscle synergy by comparing the muscle synergy for each detailed motion with the second library.

In addition, it may further include a setting step of obtaining prior information on the user's body or main motion before performing the first calculation step.

In addition, a display step of visually displaying the characteristic of the muscle synergy extracted in the characteristic extraction step may be further included.

According to an embodiment of the present invention, a user's exercise state or posture may be fed back by analyzing the user's motion and calculating a muscle synergy, which is a combination of activated muscle groups, and conversely, the user's motion is recognized through muscle synergy. Compared with the muscle synergy of many users, it can be used for rehabilitation, exercise, and treatment purposes.

1 is a block diagram showing an exercise coaching system using a muscle synergy library according to an embodiment of the present invention.
2 is a conceptual diagram showing an exercise coaching system using a muscle synergy library according to an embodiment of the present invention.
3 is a flow chart showing a method for calculating muscle synergy characteristics using a muscle synergy library according to an embodiment of the present invention.

In describing the present invention, if it is determined that a detailed description of a related known function or configuration may unnecessarily obscure the subject matter of the present invention, a detailed description thereof will be omitted.

Since the embodiments according to the concept of the present invention can apply various changes and have various forms, specific embodiments will be illustrated in the drawings and described in detail in the present specification or application. However, this is not intended to limit the embodiments according to the concept of the present invention to a specific form of disclosure, and it should be understood that all changes, equivalents, and substitutes included in the spirit and scope of the present invention are included.

The terms used in the present specification are only used to describe specific embodiments, and are not intended to limit the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In the present specification, terms such as "comprise" or "have" are intended to designate the existence of a specified feature, number, step, action, component, part, or combination thereof, but one or more other features or numbers. It is to be understood that the possibility of addition or presence of, steps, actions, components, parts, or combinations thereof is not preliminarily excluded.

Hereinafter, the present invention will be described in detail.

The exercise coaching system using the muscle synergy library of the present invention may have a structure of various types of devices that can be worn on the user's body.

1 is a block diagram showing an exercise coaching system using a muscle synergy library according to an embodiment of the present invention, and FIG. 2 is a conceptual diagram showing an exercise coaching system using a muscle synergy library according to an embodiment of the present invention.

1 and 2, an exercise coaching system according to an embodiment of the present invention includes a first analysis unit 200, a second analysis unit 400, and a characteristic extraction unit 600.

The first analysis unit 200 recognizes the user's main motion and compares it with the first library 300 including pre-stored motion information to calculate a plurality of detailed motions included in the main motion.

The first analysis unit 200 may perform AI analysis to recognize the user's main motion. AI analysis may be performed by analyzing an image or image acquired by photographing the user's motion, or by analyzing inertial change information acquired through an inertial sensor.

In one embodiment, AI analysis is performed by acquiring an image or an image generated by photographing a main operation performed by a user through a camera or a photographing device, and analyzing each pixel in the photographed image or a characteristic embedded in the photographed image. Can be recognized and patterned. In addition, deep learning technology can be applied to AI analysis.

In another embodiment, the AI analysis may analyze inertia change information according to a user's main motion acquired through an inertial sensor such as an Inertia Measurement Unit (IMU). In this case, the user wears an inertial sensor on the body, and the inertial sensor provides inertial change information according to the user's motion to the first analysis unit 200. The inertial sensor may be plural, and it is preferable that the inertial sensor is worn at a predetermined specific position on the user's body for analysis of inertial change information.

The first analysis unit 200 may compare the main motion of the user recognized through the above AI analysis with motion information stored in the first library 300 to classify the main motion into detailed motions.

The first library 300 stores a preset exercise type and motion information on a detailed motion corresponding to the exercise type, and the first analysis unit 200 allows the user to perform a main motion, for example, the first library 300 When the user performs an operation corresponding to any one of the types of exercise stored in, the main motion recognized through the first analysis unit 200 is compared with the motion information for the detailed motion previously stored in the first library 300. Classify detailed actions included in the user's main action.

For example, when the user performs a squat operation, the first analysis unit 200 compares the detailed motion of the squat stored in the first library 300 with the main motion recognized through AI analysis. The detailed motion includes a first motion in which the user sits from a standing state to an intermediate point, a second motion in which the user sits from the middle point to a lower point, a third motion in which the user stands up to the middle point, and stands at the middle point It can be divided into the fourth operation.

As for the detailed motions, the classification criteria and the number of detailed motions may be changed according to an exercise type or a set criterion. In addition, the detailed operation may mean a plurality of sections constituting the main operation performed by the user. In this case, the plurality of sections are sequentially connected to constitute the main operation.

In an embodiment of the present invention, the first analysis unit 200 may compare the detailed motion with the first library 300 to determine the exercise level for each of the detailed motions.

More specifically, the first library 300 stores a reference value of the angle or speed of the motion for each of the stored motion types, and the first analysis unit 200 stores each detailed motion constituting the main motion performed by the user. Compared with the reference value, if the shaking of the body or the change in the angle in the detailed motion exceeds the reference value, the exercise level is determined as high performance, and if it is less than the reference value, the exercise level is determined as low performance. Level information can be calculated. The level information may be different for each detailed operation, and may be stored in the first library 300. The first library 300 may store the level information for each detailed motion of the user as one data and use it as future motion information.

In an embodiment of the present invention, it may further include a setting unit 100 for acquiring information on the type of exercise. The type of exercise can be directly input by the user through a separate operation means. In addition, the exercise type information may be determined based on the user's motion analyzed through the AI analysis module. It is preferable that the setting unit 100 stores information on a plurality of types of exercises that can be compared with the types of exercises performed by the user. In addition, the setting unit 100 may transmit information on the acquired type of exercise to the first library 300, and the setting unit 100 obtains not only the type of exercise but also information on the user's body. It is desirable.

In addition, user authentication can be performed through the setting unit 100, wherein the user generates an ID in advance, and the setting unit 100 stores the generated ID. ID can be authenticated.

The second analysis unit 400 measures EMG data according to the user's main motion, and calculates muscle synergy, which is a muscle group activated during the user's main motion, based on the EMG data. Here, the muscle synergy refers to a muscle group that is activated according to the user's body movement, and there may be a plurality of muscle groups that are activated in one motion.

More specifically, muscle synergy is defined as a vector that determines muscle contraction patterns, and the activity of each muscle synergy is regulated by individual neural command signals. Each muscle synergy includes a number of muscles, and the ratio of muscles to be contracted simultaneously can be determined. Each of the different muscle synergies may contain the same muscle group at the same time. The characteristic of muscle synergy is that it is easier to control muscle groups than to control individual muscles, and can reduce the burden on the nervous system.

In an embodiment of the present invention, the second analysis unit 400 includes a sensing module 410 that acquires the EMG data, and a calculation module that calculates the muscle synergy from a group pattern of EMG data activated when the user operates. 420) and a clustering module 430 for grouping similar muscle synergies among a plurality of muscle synergies calculated by the calculation module 420 by using clustering.

The sensing module 410 measures an EMG signal activated according to a main operation performed by the user through an EMG sensor wearable on the user's body. The EMG sensor may contact the user's body and may include a plurality of channels. It is preferable that there are at least two or more of the plurality of channels. The channel may have an anode, a cathode, and a ground.

The sensing module 410 may acquire EMG data as much as the number of channels during the main operation of the user, and the plurality of channels measure different EMG data signals for each individual, and the sensing module 410 Acquires EMG data based on the signal. Also, the EMG signals measured in a plurality of channels may have different strengths.

The calculation module 420 calculates muscle synergy from a group pattern in which the EMG data are grouped. There may be a plurality of EMG data, and muscle synergy is calculated by applying an algorithm set from the group pattern of the EMG data.

More specifically, muscle synergy is calculated based on EMG data that is activated when the user performs the main action, and the calculation of EMG data or muscle synergy is performed using nonnegative matrix decomposition and clustering technology according to a preset algorithm. I can. In addition, a machine learning algorithm may be applied to grouping the activated EMG data or calculating muscle synergy.

In an embodiment, the calculation module 420 may include an input layer into which a plurality of EMG data is input and an output layer through which the muscle synergy is output. There may be a plurality of input layers, and a set algorithm is applied to each of the plurality of input layers. Assuming that the plurality of input layers are composed of first to third input layers, the first input layer receives EMG data obtained from the sensing module 410. There may be a plurality of first input layers, and the plurality of first input layers respectively correspond to channels of the sensing module 410. The plurality of first input layers transfer the EMG data to the second input layer. The second input layer is grouped based on the EMG data and transmitted to the third input layer. There may be a plurality of second input layers, and grouping is performed by grouping EMG data activated at the same time. The third input layer calculates muscle synergy by applying an algorithm set based on the group pattern of the EMG data. There may be a plurality of third input layers. It is preferable that the third input layer calculates only muscle synergy corresponding to a specific operation performed by the user, and filters other data signals.

The muscle synergy is output through the output layer, the output layer may be a plurality corresponding to the number of muscle synergy, and the muscle synergy output from each output layer may be different from each other. When a user performs a specific motion multiple times, muscle synergy may be calculated for each number of motions.

In addition, the calculation module 420 may remove dynamic noise included in the EMG signal based on the detailed operation calculated by the first analysis unit 200, and EMG data or EMG data not corresponding to the detailed operation By removing the group of, only muscle synergies corresponding to the detailed motion can be calculated.

Assuming that the user performs the squat motion multiple times among the types of exercise, the calculation module 420 groups a plurality of EMG data obtained according to the squat motion, and then calculates muscle synergy from the muscle group that contracts simultaneously. When there are 4 muscle synergies calculated per squad movement, 80 muscle synergies may be calculated by performing 20 squat movements. The muscle synergy may be stored in the second library 500.

The muscle synergy calculated by the calculation module 420 may be classified by grouping similar muscle synergies through the clustering module 430. In this case, a plurality of muscle synergies can be classified into the same cluster according to the set classification criteria. In the above example, 80 muscle synergies may be classified into 6 muscle synergy clusters. The muscle synergy clusters classified through the clustering module 430 include, for example, a first cluster that is a muscle group that is activated simultaneously in a standing state, a second cluster that is a muscle group that is simultaneously activated in a sitting motion from a standing state to an intermediate point, The third cluster, which is a muscle group that is activated simultaneously in the sitting motion from the middle point to the lower point, the fourth cluster, which is a muscle group that is activated simultaneously while sitting at the lower point, and the muscles that are activated simultaneously in the motion of standing from the lower point to the middle point. The group can be classified into a fifth cluster and a sixth cluster, which is a muscle group that is activated when a person stands in a standing state at an intermediate point.

The second library 500 stores the muscle synergy and muscle synergy cluster, and the muscle synergy or muscle synergy cluster may be classified and stored according to the detailed operation calculated by the first analysis unit 200.

Further, the second library 500 may store muscle synergy or muscle synergy clusters according to each of a plurality of users performing a specific operation. In this case, it is preferable that the second library 500 stores the muscle synergies or muscle synergy clusters by classifying them to correspond to muscle synergies of a plurality of users stored in advance.

The characteristic extracting unit 600 of the present invention compares the muscle synergy calculated by the second analysis unit 400 with the second library 500 in which the muscle synergy of a plurality of users for each detailed motion is stored, and the muscle synergy of the user is Extract features. In this case, the characteristic extracting unit 600 may obtain information on a detailed motion from the first analysis unit 200 or the first library 300 and use it for muscle synergy comparison.

For example, when the muscle synergy of multiple users stored in the second library 500 has a first cluster in the first detailed motion, the characteristic extracting unit 600 may have the first cluster in the first detailed motion of the specific user. You can calculate the probability. For example, squat among the types of exercise, the user can correct the exercise posture and method through the probability calculated by the characteristic extraction unit 600 when it is assumed that it is desirable to activate the muscle synergy of the first cluster in the detailed sitting motion. .

In addition, the characteristic extracting unit 600 calculates the occupancy at the high-performance exercise level of each muscle synergy cluster and the occupancy at the low-performance exercise level in response to the level information. I can. For example, muscle synergies included in the first cluster and the second cluster may have a share of 90% at high strength, and the share of muscle synergies included in the third cluster at low strength may be calculated as 80%.

The characteristic extracting unit 600 that calculates the share of the muscle synergy cluster according to the above method calculates the share of the muscle synergy that is activated according to the detailed motion of the user, so that the user can check the wrong muscle activity pattern at the high or low intensity exercise level. It is possible to provide feedback on which muscle pattern should be used at the high or low intensity exercise level in order to exhibit improved athletic performance.

In an embodiment of the present invention, a display unit 700 for visually displaying the muscle synergy characteristics of the user extracted by the characteristic extraction unit 600 may be further included.

It is preferable that the muscle synergy characteristics displayed through the display unit 700 include the probability of the muscle synergy cluster in the detailed motion described above or the share of the muscle synergy cluster in the exercise level.

In addition, the display unit 700 displays the type of exercise or motion information stored in the first library 300, and muscle synergy of multiple users stored in the second library 500 so that the user can visually compare and check their own muscle pattern. It is desirable to make it possible. The muscle synergy characteristic displayed through the display unit 700 may be displayed after a preset content is displayed on the display so that the user can easily recognize it, and the muscle synergy may be processed to conform to the standard of the content and then displayed.

The user can recognize his/her own exercise motion and exercise state by comparing the pattern of the muscle group activated according to his/her own exercise motion, i.e., muscle synergy and muscle synergy according to the exercise motion of others through the display unit 700. You can give feedback.

3 is a flow chart showing a method of calculating muscle synergy characteristics using a muscle synergy library according to an embodiment of the present invention.

Referring to FIG. 3, the method for calculating muscle synergy characteristics using a muscle synergy library according to an embodiment of the present invention recognizes the user's main motion and compares it with the first library 300 including pre-stored motion information. A first calculation step (S200) of calculating a plurality of detailed motions included in the main motion, measuring EMG data according to the user's main motion, and being a muscle group activated during the main motion of the user based on the EMG data A second calculation step (S300) of calculating muscle synergy for each of the detailed movements and a characteristic extraction step of extracting the characteristics of the muscle synergy by comparing the muscle synergy for each detailed movement with the second library 500 in which the muscle synergy of multiple users is stored (S400) is included.

The first calculation step (S200) may be performed using the first analysis unit 200 and the first library 300 described above, and the recognition of the user's main motion is an image or image acquired by photographing the user's motion. Can be analyzed or an inertial sensor that can contact the user's body can be used.

The second calculation step (S300) can be performed using the second analysis unit 400 and the second library 500 described above, and muscle synergy corresponding to each of the detailed motions calculated in the first calculation step (S200) Yields The muscle synergy may be calculated from a group of activated EMG data.

In the characteristic extraction step (S400), the above-described characteristic extraction unit 600 may be used, and the characteristic of the muscle synergy includes the probability of the muscle synergy cluster in the detailed motion or the share of the muscle synergy cluster at the exercise level. It is desirable to do.

In an embodiment of the present invention, a setting step (S100) of acquiring prior information on the user's body or main motion may be further included. In the setting step (S100), the above-described setting unit 100 may be used, and the user directly inputs dictionary information before the first calculation step (S200) is performed, or after the first calculation step (S200) is performed. Prior information can be obtained through recognition of the user's main motion.

In an embodiment of the present invention, a display step (S500) of visually displaying the characteristic of the muscle synergy extracted in the characteristic extraction step (S400) may be further included. In the display step S500, the above-described display unit 700 may be used, and the user may recognize and feed back an exercise motion and state by checking the muscle synergy characteristics.

100: setting unit
200: first analysis unit
300: first library
400: second analysis unit
410: sensing module
420: calculation module
430: clustering module
500: second library
600: characteristic extraction unit
700: display
S100: Setting step
S200: First calculation step
S300: second calculation step
S400: Feature extraction step
S500: Display stage

Claims (10)

A first analysis unit that recognizes a user's main motion and compares it with a first library including pre-stored motion information to calculate a plurality of detailed motions included in the main motion;
A second analysis unit that measures EMG data according to the user's main motion, and calculates muscle synergy, which is a muscle group activated during the user's main motion, based on the EMG data; And
A muscle characterized by comprising: a characteristic extracting unit for extracting characteristics of the user's muscle synergy for each of the detailed movements based on the detailed motion calculated by the first analysis unit and the muscle synergy calculated by the second analysis unit. Exercise coaching system using synergy library.
The method of claim 1,
An exercise coaching system using a muscle synergy library, further comprising: a display unit for visually displaying the muscle synergy characteristics of the user extracted from the characteristic extraction unit through a display.
The method of claim 1,
An exercise coaching system using a muscle synergy library characterized by further comprising a; setting unit for obtaining prior information on the user's body and main motion.
The method of claim 1,
The first analysis unit,
Exercise using a muscle synergy library characterized by recognizing the user's main motion by analyzing the image or image acquired by photographing the user's motion or analyzing the inertia change information acquired through the inertial sensor Coaching system.
The method of claim 1,
The second analysis unit,
A sensing module that acquires the EMG data;
A calculation module for calculating the muscle synergy from a group pattern of EMG data activated when the user operates; And
An exercise coaching system using a muscle synergy library, comprising: a clustering module for grouping similar muscle synergies among a plurality of muscle synergies calculated by the calculation module using clustering.
The method of claim 5,
The second analysis unit,
An exercise coaching system using a muscle synergy library characterized by removing dynamic noise included in the EMG signal based on the detailed motion calculated by the first analysis unit.
The method of claim 1,
The feature extraction unit,
An exercise coaching system using a muscle synergy library characterized by extracting the muscle synergy characteristics by comparing the muscle synergy calculated by the second analysis unit with a second library in which the muscle synergies of multiple users for each detailed motion are stored.
A first calculation step of recognizing a user's main motion and comparing it with a first library including pre-stored motion information to calculate a plurality of detailed motions included in the main motion;
A second calculation step of measuring EMG data according to the user's main motion, and calculating muscle synergy, which is a muscle group activated during the user's main motion, for each of the detailed motions based on the EMG data; And
A method for calculating muscle synergy characteristics using a muscle synergy library comprising a; characteristic extraction step of extracting characteristics of muscle synergy by comparing the muscle synergy for each detailed motion with a second library in which muscle synergies of multiple users are stored.
The method of claim 8,
A method for calculating muscle synergy characteristics using a muscle synergy library, further comprising: a setting step of obtaining prior information on a user's body or main motion before performing the first calculation step.
The method of claim 8,
A display step of visually displaying the characteristics of the muscle synergy extracted in the characteristic extraction step through a display; a method for calculating muscle synergy characteristics using a muscle synergy library as a characteristic.

Images