JP2011078753A - Exercise learning support device and method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To effectively support exercise learning by selecting one or more exercise features from among many exercise features and presenting a state of selected exercise feature to a leaner. <P>SOLUTION: According to a correlation among a plurality of exercise feature quantities characterizing a target exercise, one or more exercise feature quantities is selected from among the plurality of exercise feature quantities, exercise of an exercise learner is measured, a measurement result is used to calculate one or more selected exercise feature quantities, a corresponding exercise feature evaluation function is used to evaluate the calculated exercise feature amount, and the evaluation result is presented in real time to the exercise learner in a recognizable way. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

The present invention relates to a technique for supporting motor learning.

In general, there are some exercises called “knacks”, and the person who performs the exercises grasps them sensibly and makes them their own by practicing repeatedly. Speaking of baseball, there are many “knacks” such as “get your waist down” and “watch the ball well”, and by practicing, you digest and absorb them yourself. It can be said that “improvement of exercise” means “to get the knack”.

It is important to grasp this knack in many sports, but in general, the knack is defined with a vague expression, and there are many things that are difficult to grasp, and even if you want to improve your exercise well, you will not be able to grasp the knack easily Many people have never experienced such things.

In the present invention, the tips of exercise are regarded as the exercise features that characterize the exercise, and such vague “knacks” are quantitatively handled. In the present specification, a numerical value of the motion feature is referred to as a motion feature amount.

In order to support motor learning, it is considered useful to present “knacks”, that is, motor features, to the subject, but since there are usually several motor features that characterize a certain motion, it is effective. In order to support motor learning, it is important how to present which motor features.

The conventional motor learning support technology includes a technology that supports motor learning by directly stimulating the sensory organs inherent in the muscles that characterize the motion, and increases the sensitivity (Patent Document 1). There are technologies that support motor learning by converting the structure and presenting it to the learner (Patent Literature 2, Patent Literature 3), etc., but the learner is guided so that a plurality of motion features simultaneously approach an ideal state. There was no method.

In Patent Document 4, a composite image including a learner feature point and a target feature point representing a learner's body shape and a target action and a learner's image is generated, and in the composite image, A technique for changing a figure representing a learner feature point according to a difference between the position and the position of the target feature point corresponding to the learner feature point has been disclosed. It is not something to present. The “feature point” in Patent Document 4 is a representative point (typically a joint) for defining the body shape, that is, the posture of the subject or the skeletal model, and is a concept different from the “motion feature” in the present invention. Please note that.

Special table 2005-524448 JP 2004-049608 A JP 2006-075398 JP-A-2005-198818

An object of the present invention is to effectively support motor learning by selecting one or a plurality of motor characteristics from a plurality of motor characteristics and presenting the state of the selected motor characteristics to a learner. is there.

The first technical means adopted by the present invention is:
A means of measuring the movement of a motor learner;
A plurality of motion features that characterize the motion are set, and in order to calculate each motion feature as a motion feature amount, a motion feature amount calculation function prepared for each motion feature,
Means for calculating a plurality of motion feature amounts from the motion measurement result and the motion feature amount calculation function;
In order to evaluate the calculated motion feature value, a motion feature evaluation function prepared for each motion feature value,
Means for evaluating the calculated motion feature amount using a corresponding motion feature evaluation function;
Means for presenting the evaluation results to the motor learner in a recognizable manner in real time;
Means for storing a correlation between the plurality of motion feature quantities;
With
The correlation between the plurality of movement feature quantities defines the presentation effect that the presentation of the evaluation result of one or more movement feature quantities has on the movement feature quantity on which the evaluation result is presented and other movement feature quantities And
According to the correlation between the plurality of movement feature quantities, one or a plurality of movement feature quantities are selected from the plurality of movement feature quantities, and the evaluation results of the selected movement feature quantities are presented to the exercise learner in real time. Motor learning support device, characterized by
It is.
Here, the “presentation effect” means that an exercise learner presents an evaluation result of a certain exercise feature amount to an exercise learner, so that the exercise feature amount and other exercise feature amount are targets, that is, an ideal state. , Means approaching.
For example, when there are motion feature amounts A, B, C, D, and E, when the evaluation result of the motion feature amount A is presented, the motion feature amounts A, B, C, D, and E are ideal. Whether or not the motion feature values A, B, C, D, and E approach the ideal state when the evaluation result of the motion feature value B is presented. Further, when the evaluation results of a plurality of motion feature quantities, for example, two motion feature quantities A and B, are presented, whether the motion feature quantities A, B, C, D, and E approach an ideal state by numerical values. You may prescribe.

In one aspect, the correlation between the plurality of movement feature quantities is stored corresponding to each learner,
According to the learner-specific correlation between the plurality of motor feature quantities, one or a plurality of motor feature quantities are selected from the plurality of motor feature quantities, and the evaluation results of the selected motor feature quantities are sent to the motor learner. Present in real time.
There are individual differences and individualities in learning ability, and the presentation of certain motor features does not always have the same effect on all learners.

In one aspect, the selected one or more exercise feature quantities are presented to the learner by providing the learner with an evaluation result of the one or more exercise feature quantities. In addition, it is selected so that the presentation effect is exhibited in one or a plurality of other motion feature quantities in which the evaluation result is not presented.
If there are multiple motion features to achieve the target motion, when one motion feature is brought closer to the target, another motion feature is moved away from the target, and the learner cannot be guided to the target motion as a whole The problem can be solved.
In one aspect, the learning support device stores a reference motion feature value used for comparing the calculated motion feature value with the calculated motion feature value when evaluating the calculated motion feature value according to the motion feature evaluation function. It has. There are different reference motion feature values depending on the motion features and the motion feature values. Target motion feature values such as near-ideal motion feature values generated by the learner and motor feature values of skilled workers, and motion features An example is a motion feature amount (threshold value) that is a criterion for evaluation judgment in the evaluation function.
In one aspect, the evaluation means determines whether the calculated motion feature value is “in an appropriate range” or “out of an appropriate range”.
The evaluation result is either “the calculated motion feature value is within an appropriate range” or “the calculated motion feature value is outside the appropriate range”.
Further, the evaluation result may be further subdivided and presented. For example, when “the calculated exercise feature value is outside the appropriate range”, the exercise feature value is discriminated into two types according to the degree of difference from the appropriate range, and the discrimination result is identified to the exercise learner. It may be presented as possible.

In one aspect, the learning support apparatus includes display means for displaying an evaluation result,
An object corresponding to each motion feature is prepared and displayed on the display means so as to be distinguishable from each other, and the comparison result is presented to the learner by a change of each object.
In one aspect, the display means displays and displays the measured movement data (posture / movement) of the movement learner in a human body model (including identification of a human body shape by a number of marker groups), Each object is displayed in association with the corresponding exercise data. For example, if the motion feature quantity is “waist height”, the object is displayed at the waist position of the human body model or at a position moved horizontally from there.
The visually distinguishable display mode is, for example, one selected from the color, shape, and size of the object, or any combination thereof. Then, by changing one or any combination selected from the color, shape, and size of the object, the learner can recognize whether each motion feature is close to or far from the target.

  Corresponding to each motion feature amount, sounds that can be distinguished from each other are prepared, and the evaluation result is presented by sound. Alternatively, a combination of visually identifiable presentation means and audibly identifiable presentation means may be used. For example, a certain motion feature is presented so as to be visually identifiable, and a certain motion feature is presented so as to be audibly distinguishable. Alternatively, the same motion feature may be presented by both visually identifiable means and audibly identifiable means.

The second technical means adopted by the present invention is:
A plurality of motion features that characterize the target motion are set, and in order to calculate each motion feature as a motion feature amount, a motion feature amount calculation function is prepared for each motion feature,
In order to evaluate the calculated motion feature value, a motion feature evaluation function is prepared for each motion feature value.
Correspondence between multiple motion feature values that specify the effect of presentation of the evaluation result of one or more motion feature values on the motion feature value for which the evaluation results are presented and other motion feature values is provided Has been
According to the correlation between a plurality of motion feature quantities characterizing the target motion, one or more motion feature quantities are selected from the plurality of motion feature quantities,
Measuring an exercise learner's exercise, calculating one or more selected exercise feature quantities using an exercise measurement result and an exercise feature quantity calculation function;
A motor learning support method that evaluates a calculated motor feature using a corresponding motor feature evaluation function and presents the evaluation result to a motor learner in a recognizable manner in real time.

In one aspect,
The correlation between the plurality of movement feature quantities is prepared for each learner,
One or more exercise feature values are selected from the plurality of exercise feature values in accordance with a learner-specific correlation between the plurality of exercise feature values.

  In one aspect, the selected one or more exercise feature quantities are presented to the learner by providing the learner with an evaluation result of the one or more exercise feature quantities. In addition, it is selected so that the presentation effect is exhibited in one or a plurality of other motion feature quantities in which the evaluation result is not presented.

  In one aspect, an object corresponding to each motion feature is prepared and displayed on the display means so as to be distinguishable from each other, and the comparison result is presented to the learner by a change of each object.

  In one embodiment, sounds that can be distinguished from each other are prepared corresponding to each motion feature quantity, and the comparison result is presented by sound.

The third technical means adopted by the present invention is:
A means of measuring the movement of a motor learner;
A plurality of motion features that characterize the motion are set, and in order to calculate each motion feature as a motion feature amount, a motion feature amount calculation function prepared for each motion feature,
Means for calculating a plurality of motion feature amounts from the motion measurement result and the motion feature amount calculation function;
In order to evaluate the calculated motion feature value, a motion feature evaluation function prepared for each motion feature value,
Means for evaluating the calculated motion feature amount using a corresponding motion feature evaluation function;
Means for presenting the evaluation results to the motor learner in a recognizable manner in real time;
Means for storing the evaluation results;
With
One or a plurality of movement feature quantities are selected from the plurality of movement feature quantities, and the evaluation results of the selected movement feature quantities are presented to the movement learner in real time, and the evaluation results of other movement feature quantities that are not presented The evaluation result of the selected motion feature amount is converted into the motion feature amount for which the evaluation result is presented and the other motion feature amount using the evaluation result that has been presented and the evaluation result that has not been presented. An apparatus for acquiring a correlation between motor feature quantities used for motor learning support, characterized by acquiring a correlation between a plurality of motor feature quantities that specify a presentation effect to be given peculiar to the motor and / or learner .
If the learner presents the motion feature value and the result of comparison with the reference motion feature value, the effect of bringing the presented motion feature value closer to the ideal state can be obtained than when the motion feature value is not presented. At this time, there may be a side effect that the other motion feature amount moves away from the ideal state. By using the present invention, it is possible to comprehensively investigate the influence of the presented motion feature value on other motion feature values. Therefore, by presenting, it is possible to identify an effective motion feature that can simultaneously bring a plurality of motion features close to the target, and the learner's motion is effective for the target motion as a whole. Can help you get closer.

  There are several modes for evaluating the presentation effect. One aspect is the evaluation of the presentation effect by the threshold, and the longer the feature amount is within the threshold, the more effective the presentation is, with the time when the feature amount is within the threshold (seconds within the threshold) as the evaluation material. And Another aspect is the evaluation of the presentation effect by continuous values. For the calculated motion feature value, the motion feature value is evaluated with an evaluation index that is closer to the target motion feature value as the value is larger. It is assumed that the larger the value, the better the presentation effect. It is understood by those skilled in the art that the presentation effect evaluation method is not limited to the above two, and other evaluation methods can be adopted.

The fourth technical means adopted by the present invention is:
A plurality of motion features that characterize the target motion are set, and in order to calculate each motion feature as a motion feature amount, a motion feature amount calculation function is prepared for each motion feature,
In order to evaluate the calculated motion feature value, a motion feature evaluation function is prepared for each motion feature value.
Measure the movement of the movement learner, calculate a plurality of movement feature quantities that characterize the movement using the movement measurement result and the movement feature amount calculation function,
Evaluate the calculated motion feature using a corresponding motion feature evaluation function,
One or a plurality of movement feature quantities are selected from the plurality of movement feature quantities, and the evaluation results of the selected movement feature quantities are presented to the movement learner in real time, and the evaluation results of other movement feature quantities that are not presented The evaluation result of the selected motion feature amount is converted into the motion feature amount for which the evaluation result is presented and the other motion feature amount using the evaluation result that has been presented and the evaluation result that has not been presented. A method for acquiring a correlation between motor feature values used for motor learning support, characterized by acquiring a correlation between a plurality of motor feature parameters that prescribes a presentation effect to be specific to a motor and / or a learner. .

In the present invention, one or a plurality of motion features are selected from the plurality of motion features according to the correlation between the plurality of motion feature quantities, and the selected motion feature state is presented to the learner effectively. Support motor learning.

In the present invention, by presenting a state of a different type of motion feature quantity from the target motion feature quantity, it is possible to identify a motion feature quantity that can change the target motion feature quantity as a result. By presenting the state of the identified motion feature quantity, it is possible to simultaneously lead a plurality of motion features to an ideal state. As a result, the learner's motion feature amount can be effectively guided in the target direction.

If the learner presents the motion feature value and the result of comparison with the reference motion feature value, the effect of bringing the presented motion feature value closer to the ideal state can be obtained than when the motion feature value is not presented. At this time, there may be a side effect that the other motion feature amount moves away from the ideal state. By using the present invention, it is possible to comprehensively investigate the influence of the presented motion feature value on other motion feature values. Therefore, by presenting, it is possible to identify an effective motion feature that can simultaneously bring a plurality of motion features close to the target, and the learner's motion is effective for the target motion as a whole. Can help you get closer.

1 is an overall view of a motor learning support system according to the present invention. It is a figure explaining the procedure which links the marker group with the three-dimensional coordinate obtained by motion measurement, and the body model on a computer. The upper left is wearing a suit with a marker on the whole body, the upper right is a group of markers with three-dimensional coordinates obtained by a motion measuring device, and the lower is a body model on a computer that is linked to the movement of the marker group . It is a figure which shows a mode that the body model interlock | cooperated with the marker group which has a three-dimensional coordinate output from a motion information calculation apparatus, and the motion feature output from a motion feature-value calculation apparatus were linked and shown. It is a figure which illustrates presentation of the state of a movement feature-value. The upper figure is when the waist height is within the threshold (in the example shown, the object corresponding to the motion feature is gray), and the lower figure is when the waist height is outside the threshold (in the example shown) The object corresponding to the movement feature amount is red). It is a figure which illustrates presentation of the state of a plurality of movement feature-values. It is a figure explaining designation | designated of the marker for a calculation. It is a figure which shows creation of the object corresponding to the gravity center position of the waist. It is a figure which shows the mode of a change of the height of a waist. The upper diagram shows a case where the threshold is not exceeded (in the illustrated example, the sphere object is gray), and the lower diagram shows the case where the threshold is exceeded (in the illustrated example, the sphere object is red). It is a figure which shows the marker position for calculating the space | interval of both feet. The presentation screen of the distance between both feet is shown. The upper diagram shows a case where the threshold value is not exceeded (in the illustrated example, the cylindrical object is green), and the lower diagram shows the case where the threshold value is exceeded (in the illustrated example, the cylindrical object is blue). It is a figure which shows the marker position for calculating the height of a hand. The hand height display screen is shown. The upper diagram shows a case where the threshold value is not exceeded (in the illustrated example, the cubic object is gray), and the lower diagram shows the case where the threshold value is exceeded (in the illustrated example, the cubic object is red). It is a figure which shows the marker position for calculating the direction of a body. It is a figure explaining calculation of the direction of a body. The body orientation screen is shown. In the example shown in the figure, when the body direction is leftward and exceeds the threshold, the cone-shaped object is green (A in the figure), and when the body direction is rightward and exceeds the threshold, the cone-shaped object The object turns red (D in the figure). It is a figure which shows the marker position for calculating the swing of a sword. It is a figure explaining calculation of a sword swing. In the example shown in the figure, the marker specified at the end of the sword should pass through the part surrounded by red diagonal lines, that is, the part satisfying [x <-40 and y>] and [x> 40 and y> 140] It is space. Shows the screen for sword swing. The upper left figure shows that the sword has passed through the space specified on the left side (in the example shown, the color of the cube object is green), and the upper right figure shows the space specified on the right side. (In the example shown, the color of the cubic object is blue), and the following figure shows the case where there is no sword marker in the specified space (in the example shown, the cubic object is gray) . It is a figure which shows the motor learning support system which presents the feature-value of an exercise | movement with an audio | voice in real time.

[A] Motor Learning Support System FIG. 1 shows a motor learning support system according to the present invention. The motor learning support system is configured with a motion measurement means (motion capture + motion information calculation device) that measures a learner's motion in real time and a plurality of motion features that characterize the motion (motion feature set). In order to calculate a feature as a motion feature amount, an arbitrary motion feature amount using a motion feature amount calculation function prepared for each motion feature and a motion feature amount calculation function preset from the measured motion data Motor feature quantity calculation means (motor feature quantity calculation device) for calculating the calculated exercise feature quantity storage means for storing the calculated learner's exercise feature quantity (calculated exercise feature quantity data), and the calculated exercise feature quantity Motion feature evaluation function prepared for each motion feature amount and the motion feature evaluation using the motion feature evaluation function corresponding to the calculated motion feature amount. And means (motion feature evaluation unit), the evaluation results comprises an evaluation result display means for recognizably presented in real-time motion learner (presentation unit), a. The presentation device is typically a display device, and displays the measured learner's exercise data (learner's posture / movement) in addition to the evaluation result. The exercise data is expressed by a human body model (including a case where a human body shape is specified by a large number of marker groups).

The motor learning support system further provides motor feature values that are the target, such as near-ideal motor feature values generated by the learner and motor feature values of the skilled person, and motor feature values that serve as the basis for evaluation judgments in the motor feature evaluation function Reference motion feature amount storage means for storing a reference motion feature amount (reference motion feature amount data) such as (threshold) is provided. The movement feature evaluation means compares the movement feature amount calculated during the exercise of the learner with the reference movement feature amount using a movement feature evaluation function, and presents the comparison result as an evaluation result.

Here, the evaluation result is a relationship between the reference motion feature amount and the calculated motion feature amount defined by the motion feature evaluation function. In one aspect, a certain motion feature amount at a certain time point is “ It is “within the proper range” or “out of the proper range”.

In one aspect, the motion feature evaluation function defines a center point of a certain target value, sets a value having a width of a size set by a certain threshold as a “target range”, Time is defined to satisfy the condition (within an appropriate range). For example, a case is defined in which the variance value of the waist height is set as the motion feature value and does not exceed a certain value set by the threshold value. In one aspect, the motion characteristic evaluation function is defined so as to satisfy a condition (within an appropriate range) when it exceeds or falls below a certain reference value (threshold value). For example, the waist height is defined as an exercise feature value, and a case where the width of a certain value does not exceed or falls below is defined.

The motor learning support system further includes an evaluation result storage unit that stores an evaluation result (motor feature evaluation result data), a motor feature correlation calculation device that calculates a correlation between a plurality of motor features using the evaluation result, and Correlation data storage means for storing the calculated correlation data between motion features is provided. The correlation data between motor features indicates that the presentation of the evaluation result of the selected motor feature value has a presentation effect on the motor feature value and other motor feature values of the evaluation result that are specific to the exercise and / or learner. Stipulate. The motor feature correlation calculation device selects one or a plurality of motor feature quantities from a plurality of motor feature quantities, presents the evaluation result of the selected motor feature quantity to the motor learner in real time, and is not presented The evaluation result of the movement feature quantity is acquired, and the presentation effect is evaluated using the presented evaluation result and the evaluation result that has not been presented, whereby the correlation between the movement feature quantity is quantified and acquired. For example, the correlation data between the movement feature amounts is composed of two values “1” when there is a presentation effect and “0” when there is no presentation effect.

When a motor learner starts using a motor learning support system for the first time, there is no correlation data between motor features unique to the motor learner. In one aspect, at first, the motor features are appropriately selected without correlation data, and the motor learner's motion is measured, the selected motor feature is calculated, the motor feature is evaluated, the evaluation result is presented, and the presentation effect is evaluated. If the system is operated once, correlation data between motion features is stored, and in the next trial, motion features can be selected with reference to the correlation data between motion features. In addition, correlation data between motor features as initial data corresponding to the target motion (for example, representative values of correlation data between motor features of a plurality of motor learners, specifically, average values and medians) It is also possible to acquire correlation data between motor features for a specific learner by correcting and updating the initial data by preparing a specific motor learner and repeating the trial. In addition, with respect to correlation data between motor features unique to a motor learner, the correlation data between motor features can be updated with the trial of the motor learner.

In the present embodiment, the motor learning support system can be composed of a motion capture system, one or a plurality of computers, and a display system. The computer includes a general-purpose computer such as a personal computer (specifically, an input device, an output device, a processing device such as a CPU, a storage device such as a ROM and a RAM, and a bus for connecting them). Can be configured. Further, when the evaluation result presenting means presents sound, the motor learning support system may include a speaker for outputting sound.

In order to calculate the exercise feature amount during the learner's exercise, first, it is necessary to prepare a plurality of exercise features that characterize the target exercise. There are several ways to prepare motion features. For example, select a plurality of exercise features to be noticed by listening to the subject's exercise proficiency, literature related to the exercise (motion analysis papers, books, magazines, information sources on the Internet, etc.) Set a set of motion features. The set of motion features differs for each target motion, and a set of motion features is prepared for each motion.

After determining the motion feature, the procedure for calculating the motion feature amount from the learner's motion data measured by the motion measuring means is set as a motion feature amount calculation function. Each motion feature value is calculated in real time using the motion data calculated by the motion measuring means by programming the calculation procedure of each motion feature value in advance.

In one aspect, an optical motion capture system is employed as the motion measurement means. The optical motion capture system includes a large number of optical markers (for example, infrared reflective markers) attached to a subject, a plurality of cameras that capture the motion of the subject wearing the optical markers, and two markers acquired by each camera. Reconstructing the 3D position, calculating the 3D position of the marker, and obtaining the 3D position (subject's posture) of each part of the body in real time from the 3D position of the optical marker and the 3D model of the body And a display unit that displays the results of the processing unit (operations acquired as time-series data of the posture of the subject). Various calculations in the motion capture system are executed by a computer.

The three-dimensional position and orientation of each part of the body is obtained from the three-dimensional position of the optical marker and the three-dimensional model of the body, which are obtained by the optical motion capture system and arranged at each part of the body. According to a preset calculation method of each motion feature, the motion feature is quantified by using the three-dimensional position / posture data of each part of the body acquired by the motion capture system, the marker position data, and the like, and the motion feature amount is obtained. The learner's motion feature value is stored in the measured motion feature value storage means. When a target exercise uses a tool, the tool motion data can be acquired by attaching a marker to the tool.

By using the motion capture system, it is possible to acquire the exercise data of the subject and the exercise data of the exercise equipment. By setting a marker to calculate a certain motion feature and pre-determining how to calculate the motion feature using that marker, all motion features can be numerically calculated using the measured motion data. Can be

In the present embodiment, a procedure for performing motion measurement using an optical motion capture system and calculating a motion feature amount by combining the motion information calculation device and the motion feature amount calculation device will be described below.

The motor learner wears a suit with a marker on the whole body (upper left in FIG. 2). The three-dimensional position coordinates of the marker group are obtained using the optical motion capture system (upper right of FIG. 2). The motion information calculation device obtains the position and orientation of the body model in conjunction with the position coordinates of the marker (bottom in FIG. 2). FIG. 3 shows a state in which a body model linked to a marker group having three-dimensional coordinates output from the motion information calculation device and a motion feature output from the motion feature amount calculation device are linked together and presented.

By connecting a projector to a computer constituting the motion information calculation device, motion feature amount calculation device, and motion feature evaluation device and presenting it on the screen, measured motion data, motion features (objects as symbols of motion features), It is possible to present real-time evaluation results for motor features. The learner can check his / her movement while exercising. FIG. 1 shows the state of the experiment when real-time presentation is performed. The subject is in front of the right side, and the real-time presentation is performed on the screen on the back left side.

A body model linked to a group of markers having three-dimensional coordinates, output from the motion information calculation device, and an object expressing the evaluation result of the motion feature obtained by the motion feature amount calculation device and the motion feature evaluation device, simultaneously The state of presentation is shown in FIG. 4A.

Various types of optical motion capture systems are commercially available, and the motor learning support system of the present invention can be configured using them and a computer.

In addition to the optical motion capture system, various methods are known for motion measurement, such as mechanical motion capture system, magnetic motion capture system, video analysis, electromyography, floor reaction force meter, etc. Since it is known what kind of motion data can be obtained by those motion measurement means, any kind of motion measurement means is used as long as the motion feature to be obtained can be calculated from the motion data. be able to. Further, motion data may be acquired using a plurality of types of motion measurement means, for example, a motion capture system and an electromyograph, a motion capture system and a force plate at the same time.

The reference movement feature quantity storage means stores a reference movement feature quantity used for comparison with the calculated movement feature quantity when the learner's movement feature quantity obtained by calculation is evaluated according to the movement feature evaluation function. Has been.

In one aspect, the reference movement feature amount is a target value or an ideal value (target movement feature amount) for the learner, and the target movement feature amount includes an exercise feature amount close to ideal generated by the learner in the past, The skilled person's movement feature-value can be used. These target motion feature quantities can actually measure the movements of the learner and the skilled person, and ideally set the ideal motion feature quantity state based on the measured movements. In addition, the target motion feature amount is not necessarily limited to the actually measured value, and depending on the nature of the motion feature amount, the motion feature amount may be arbitrarily set based on ideal values, statistical values, existing data, and the like. . For example, if the motion feature value is a variance value and it is desirable that the fluctuation with time is small, the variance value 0 may be set as the ideal motion feature value as an ideal value.

In an embodiment described later, the reference motion feature value is a threshold value, and the motion feature evaluation function compares the calculated motion feature value with a range of values set by the threshold value, and the calculated motion feature value is “ Defined to determine "within appropriate range" or "out of proper range". The motor learning support system includes a threshold value table that stores threshold values (reference motor feature values) corresponding to the motor feature values. The motion feature amount evaluation means is calculated by comparing the learner's motion feature amount stored in the calculated motion feature amount storage means with the threshold value stored in the threshold table using the motion feature evaluation function. It is determined whether the motion feature value is within a range of values set by the threshold value or outside the threshold value.

The method of setting the threshold differs depending on the target exercise and exercise feature, but those skilled in the art appropriately set the threshold based on the nature of the exercise feature, the learner's physical feature, the learner's past exercise data, etc. obtain. In setting the threshold value, for example, an exercise feature amount of a skilled person can be referred to. The motion feature quantity evaluation means does not directly compare the calculated motion feature quantity with a threshold value (reference motion feature quantity), but the calculated motion feature quantity and the reference motion stored in the reference motion feature quantity storage means. When comparing with feature quantities and determining whether the difference is within a preset threshold range, for example, when an ideal value is a reference motion feature quantity, If the average difference between the reference motion feature amount and d ₁ is d ₁ , the threshold th ₁ applied to the motor learner will be at least greater than d ₁ .

The reference motion feature value is stored in the reference motion feature value storage means in a format that can be compared with the learner's motion feature value obtained by calculation. The format that can be compared is not limited to the case where the comparison can be made directly, but includes the case where either or both of the reference movement feature amount and the learner's movement feature amount can be processed in real time by a predetermined calculation. . For example, the comparison may be performed by scaling or normalizing any feature amount.

The evaluation result obtained by the motor feature evaluation means shows the state of the learner's motor feature value (whether a certain motor feature value is “appropriate within range” or “out of the appropriate range” at a certain time). It is shown and presented in real time in a manner recognizable to the learner. Typically, the evaluation results are presented to the learner in a visually identifiable manner on the display system display.

How to display the evaluation result so as to be visually identifiable is determined in advance. In one aspect, an object corresponding to each movement feature is prepared, and each object is displayed on the display in a form that can be distinguished from each other. Examples of forms that can be distinguished from each other include shapes, sizes, colors, or any combination thereof. For example, the object is changed so as to be visually recognizable depending on whether or not the learner's motion feature value exceeds a threshold value that is a reference motion feature value. For example, the shape, size, color, or any combination of the objects on the display is changed depending on whether the learner's motion feature value is within the threshold value or outside the threshold value.

For example, when teaching a movement that is important to keep the waist height constant, calculate the three-dimensional position of the center of the waist and move the sphere (object) to the position translated from the actual waist while maintaining the height. Is presented (FIG. 4A). The body model of the user of the system is presented at a position in the actually measured space, and the color of the sphere is changed depending on whether or not a predetermined threshold range is exceeded. As a result, it is possible to present the learner in real time with information obtained by a conventional instructor who gives a visual instruction or analyzes the video measured by the video camera later.

Although one motor feature is presented in FIG. 4A, the motor learning support system according to the present invention is a system that can present a plurality of motor features. FIG. 4B shows an example of a presentation screen in the case of a slashing exercise. In FIG. 4B, it corresponds to five movement features consisting of waist height (A), width of both feet (B), hand height (C), hip orientation (D), and sword swing width (E). The object is being presented.

The evaluation result may be presented to the learner in an audible manner. For example, sounds corresponding to each of the movement feature quantities that can be distinguished from each other are prepared, and when the learner's movement feature quantity is outside the threshold, a sound corresponding to the movement feature quantity may be generated. .

The calculated exercise feature quantity storage means stores time series data of each calculated exercise feature quantity of the learner, and it is also possible to calculate the evaluation result at a certain time later. In one aspect, the motor learning support system includes an evaluation result storage unit that stores an evaluation result for each motor feature, and can analyze a learner's degree of learning from time series data of the evaluation result.

As a simple example, the calculation of the motion feature value will be described by taking a change in waist height during squat as an example. The motion characteristic in this case is “waist height”, and it can be said that the larger the fluctuation of “waist height”, the more effective squat.

When the motion feature is determined, the motion feature value is calculated and output. For this purpose, a marker for calculating an exercise feature is selected from an optical marker group arranged at each part of the body, which is calculated by the exercise information calculation device and displayed on the presentation device. A necessary marker for obtaining the waist height is designated from the marker group (FIG. 5A). In the motion feature amount calculation device, the waist height is selected here as the motion feature amount, and an object whose position coordinates change in conjunction with the waist height is presented to the learner.

As shown in FIGS. 5A and 5B, the center of gravity of the waist is obtained from the four markers of the waist, and the motion feature, that is, the height of the waist (the height of the center of gravity of the waist) is calculated using the center of gravity of the waist.

Each position vectors of the four markers of the hip, when _{V 1, V 2, V 3} , V 4, the centroid vector V _G waist,
It is expressed as

The position vector V _trans of the object linked to the center of gravity of the waist is
By doing so, an object linked to the waist height is created. Examples of the movement feature amount for determining whether or not an ideal squat is performed include the amount of fluctuation of the waist center of gravity and the height of the waist center when bent.

The coordinate position of this object can be output as time series data. The position coordinate data of the object, that is, the X component, the Y component, and the Z component can be output as text data at an appropriate frame rate, and the motion can be analyzed afterwards.

[B] Calculation of motor feature value, evaluation of motor feature value, presentation of evaluation result Hereinafter, with respect to the embodiment of the motor learning support system of the present invention, an example of a slashing motion that is one of the basic operations of the Kashima Shinryu Will be described in detail. About the setting of the movement feature, the feature of the movement that can be presented by this system was selected from the opinion of the expert and implemented. Specifically, as described in detail below, the movement characteristics include waist height (A), foot width (B), hand height (C), body orientation (D), sword swing (E) was set (see FIG. 4B). These motor features are just one example. Other motor features such as Mihoko Otake, Katsushi Yamane, Hitoshi Nakamura. Subjective evaluation based on somatosensory sensation and measurement of movement. Objective evaluation based on the 21st Annual Conference of the Robotics Society of Japan, 2J26, 2003. "The trajectory of the sword tip is straight and the trajectory of the sword tip does not go away like the letter" No "of Katakana (the trajectory of the trajectory Focusing on “straight-ahead performance as a motion feature”, “the inner product of the velocity vector of the sword tip and the velocity vector in the direction of the cutting edge” may be used as the motion feature amount.

(Waist height)
One of the basic points of the slashing movement is that the waist height is constant. Waist height is the motion feature and waist height variance is the feature. The lower the variance value, the more stable the waist. The waist height can be calculated as the position of the center of gravity of the waist. For a specific calculation method, the calculation of the position of the center of gravity of the waist in the squat operation can be used. A threshold is set in advance, and when the learner's waist height is outside the threshold, the color of the corresponding object is changed. The threshold value was varied from subject to subject. The threshold value can be determined as follows, for example. Analyze the expert's data, calculate the standard height from the obtained dispersion value / average value, and determine the standard threshold width. The subject takes the stance of slashing, and the threshold is determined for each subject from the ratio of the height of the waist in that state to the standard height of the expert.

Determining the center of gravity position vector V _G of the waist markers in four waist. The y component vGy of this vector represents the waist height. Assuming that the reference value of the subject's waist height is a and the deviation from the waist height reference value is d,
It becomes. Next, a threshold value is set. When the reference value for the waist height is 85 cm, the allowable width is 5 cm.
It becomes. When the deviation from the reference value exceeds this threshold, the color of the object is changed.
Function T (d, t)
Is a color discriminant. The vector V _color to change the color of the object,
As a result, the color of the object changes according to the waist height. V _color vector elements correspond to red, blue and green brightness values, respectively, (1,0,0) is red, (0,1,0) is green, (0,0,1) Represents blue. FIG. 7 shows a feature amount presentation. When the waist height is within the threshold, the sphere is gray (upper figure in FIG. 7). When the waist height is outside the threshold, the sphere is red (the lower diagram in FIG. 7).

(Between legs)
In the razor slashing movement, by opening both feet firmly, the whole body has a low center of gravity, making it a stable movement. For this reason, the distance between both feet was used as a feature value. The circled part in FIG. 8 is a marker for calculating the foot width. The center of two points in each circle was calculated, and the distance between the center points was defined as the distance between both feet.

If the left foot marker is V ₁ , V ₂ , the right foot marker is V ₃ , V ₄ , the center coordinate of the right foot tip is V _R , and the center coordinate of the left foot tip is V _L ,
It becomes. If the width of both feet is d,
It is expressed as The foot width threshold for each subject is set to _dt, and the corresponding object color is changed when the foot width exceeds the threshold. Function for color discrimination
The
It is determined. The vector V _color to change the color of the object,
Thus, an object whose color changes corresponding to the foot width is created.

Next, define the object to fit between both feet and stretch. Let V _{trans be} the position vector of the object. The center coordinates of the object are matched with the barycentric coordinates between both feet, and the position of the object is defined.
Synchronize the rotation of the foot with the rotation of the object. A vector V _dif obtained by subtracting the position vector of the left foot from the position vector of the right foot and a fixed vector V _f are determined as follows.
_If the angle between the vector V _dif and the vector V _f is θ,
Is calculated. Let V _rot be the vector that rotates the object,
By doing so, the rotation of the foot and the rotation of the object are synchronized. Finally, synchronize the foot width and the size of the object. Since the default width of the object is 20cm, you can synchronize by multiplying d by 20. A scale vector V _scale that changes the size of the object,
By doing so, the foot width and the size of the object are synchronized. Apply the four vectors V _color , V _trans , V _rot , and V _scale to the object and present it.

The presentation output by this is as shown in FIG. When the object is scaled to the distance between the foot markers and exceeds a preset threshold, V _color changes from (0,1,0) to (0,0,1) and the color is green (above) Changes from blue to blue (see below).

As a method for determining the threshold value, the slashing stance was taken, and the width of the foot when the test subject was in the correct posture was set as the threshold value.

(Hand height)
In the slashing movement, it is important to swing the wooden sword up firmly. Therefore, the height of the hand was set as the feature amount.

A calculation method will be described. The marker in the circled part in FIG. 10 is a marker to be measured. The position vector of the marker for calculating the hand height is V _h and its y component is v _hy . This v _hy is the height of the hand. The position vector V _{trans of the} object,
By doing so, the position of the hand marker and the object are synchronized. Since the default size of the object is small, the scale vector V _scale is
And expand.

Finally, the change in color due to the threshold is defined. The threshold of the target subject and h _t, when the height of the hand exceeds this threshold, set so that the color changes. The color discrimination function T (v _hy , h _t ) is
And A vector V _color that changes the color of the object
By doing so, an object whose color changes in accordance with the height of the hand is created.

The output presentation screen is as shown in FIG. The object (here, a cube) moves in the same way as the designated marker, and when the preset threshold is exceeded, the color changes to red.

As a method for determining the threshold value, the height of the designated marker when the sword was swung up and the sword was raised the highest was used as the threshold value. When this value is exceeded, the color of the object changes on the presentation screen.

(Direction of body)
When the sword is swung away in the slashing movement, it is desirable that the body is nearly parallel to the front direction, that is, the direction to be cut. Therefore, the angle between the front direction and the direction vector created by the marker near the hip bone was used as a feature point.

A calculation method will be described. Two markers surrounded by a circle in FIG. 12 were designated as markers near the hip bone. An angle formed by the vector between the markers and the vector in the front direction is defined as a feature amount. Here, the vector in the front direction is a vector having a component (0, 100) in the back direction from the origin on the absolute coordinates toward the screen. FIG. 13 shows an image of calculation.

Let V _w1 and V _w2 be the two markers on the waist.
V ₁ and V _{2 in} FIG.
It is expressed as If the angle between vector V ₁ and vector V ₂ is θ,
Is calculated. The color of the object is changed by this θ and the set threshold value. The angle from the front when the body is facing right is θ _R , and the angle from the front when the body is facing left is θ _L. When θ exceeds or falls below these two thresholds, the color of the object is set to change. Color discrimination functions T1 (θ, θ _R ), T2 (θ, θ _L )
And A vector V _color that changes the color of the object
By doing so, an object whose color changes corresponding to the direction of the body is created.

Object position vector V _trans , scale vector V _scale ,
And present it to the subject.

The output presentation screen is as shown in FIG. The value of the calculated angle (the unit is frequency) is set as the value of the x component, and the object moves on the X axis according to the value. The color changes when the set threshold is exceeded. When the body turns to the left and falls below the threshold, V _color becomes (0,1,0) and the object changes to green (FIG. 14A). When the body turns right and exceeds the threshold, V _color becomes (1,0,0), and the object changes to red (FIG. 14D).

A method for determining the threshold will be described. Based on the data of advanced users, when the calculated angle is 35 degrees or less or 145 degrees or more, the required condition is satisfied, that is, the threshold value is exceeded. When this value is exceeded, the color of the object changes on the presentation screen.

(Shake sword greatly)
In the slashing movement, it is important to shake the sword greatly. It is necessary to swing the sword in an image that swings down diagonally rather than slashing vertically. Therefore, the position information of the marker attached to the sword tip is acquired, and it is determined whether the marker passes through the designated space. Originally, it is necessary to swing only when the sword is shaken in the slashing movement, but in this presentation, it is also judged when the sword is raised for the convenience of the system.

First, a marker is specified for calculation. A marker surrounded by a circle in FIG. 15 is a marker at the sword, and is a designated marker in this calculation. Next, designation of the space through which the marker should pass will be described. A portion surrounded by diagonal lines in FIG. 16 is a designated space. That is, the portion satisfies [x <−40 and y> 140] and [x> 40 and y> 140]. At this time, the z direction, that is, the depth direction of the screen may take an arbitrary value.

The position vector of the sword tip marker is V _k = (v _kx , v _ky , v _kz ). Color discrimination functions T1 (v _kx , v _ky ), T2 (v _kx , v _ky )
And A vector V _color that changes the color of the object
By doing so, an object whose color changes corresponding to the swing of the sword is created.

The position of the object is fixed, and the position vector V _trans is
It was.

The output presentation screen is as shown in FIG. The left display screen of FIG. 17 indicates that the sword tip marker satisfies [x <−40 and y> 140], and the color of the object representing the state of the sword tip marker is (0,1,0 ), That is, green. The presentation screen on the right shows that the sword marker satisfies [x> 40 and y> 140], and the color of the object representing the sword marker state is (0,0,1), that is, blue Become. The lower presentation screen is in a state where none is satisfied.

The threshold value is determined by the method described in the explanation item of each motion feature. For the three items of waist height, foot width, and hand height, the values of the target feature points at the time of the first experiment and the target point when the wooden sword was raised were set as reference values. Regarding the waist height, the value corresponding to the subject's waist height (85 ± 5 cm when the waist height is 85 cm when holding a wooden sword) is within the threshold range, and the foot width and hand height are More than the defined reference value was set within the threshold range. For the two items of body orientation and sword swing, we analyzed the motion data of advanced sword-cutting motions taken in advance and determined approximate reference values. Regarding the body orientation, when the feature value satisfies x <35 or 145 <x, this is within the threshold range, and for the way of swinging the sword, the feature value is [x <−40 and y> 140. ] Or [x> 40 and y> 140] is satisfied within the threshold range. Table 1 shows the feature value within the threshold range for each subject.

[C] Correlation between motion features (eg, slashing motion)
Next, two evaluation experiments performed on the slashing exercise using the motor learning support system according to the present invention will be described.

[C-1] Comparison of movements based on presence / absence of real-time presentation The movement feature to be presented is set as one, and a comparison is made based on whether real-time presentation is present. The movement to be measured is slashing, which is one of the basic actions of the Kashima Shinryu. For the outside, dynamic characteristics that affect speed and force are required, and at the same time it is necessary to satisfy the conflicting conditions of maintaining the stability of the body, so we decided to use this system for evaluation. There are a plurality of motion characteristics corresponding to the conditions to be satisfied. Here, the waist height corresponding to the stability of the body is defined as the motion characteristics, and the variance is defined as the feature quantity. The experiment was conducted on 2 subjects. First, the subject showed a movie of a skilled person in order to learn the shape of the slashing movement. Next, after practicing for several minutes, three trials were performed when there was no presentation in the system, and three trials were conducted when there was presentation. Shake a wooden sword 10 times in one trial. The subject's waist height variance values decreased from 10.56 to 8.71 and 3.77 to 1.43, respectively, and were more stable when presented than when presented.

[C-2] Influence of motion features on different motion feature amounts The presentation of motion features evaluates the relative effect on each motion feature amount. Experiments were conducted when multiple motion features were set for the slashing motion. Waist height (A), width of both feet (B), hand height (C), waist orientation (D), and sword swing width (E) were characteristic of movement (FIG. 4B). A threshold value is set in advance for each motion feature, and each motion feature value changes the color of the corresponding object in accordance with the threshold value. The experiment was performed twice a week for a total of 6 subjects over 3 weeks for 3 subjects who learned basic motions. From the 1st to the 5th time, motor features were presented one by one, and all were presented on the 6th time. In each round, 7 sets of 20 slashers were performed. In the analysis, the relationship between motor features was examined by evaluating how much each feature changes depending on the types of motor features to be presented. Two comparative methods were analyzed below. One is that the time when the feature amount is within the threshold (seconds within the threshold) is used as an evaluation material, and the more the feature amount is within the threshold, the more the presentation effect is. The other is that for the calculated exercise feature value, the exercise feature value is evaluated with an evaluation index that becomes closer to the target exercise feature value as the value is larger, and the larger the evaluation index is, the more effective the presentation is.

(A) Correlation 1 between motion features
The relationship between the motion features is verified by evaluating how much each motion feature amount differs depending on the motion feature to be presented with a ratio of seconds within a threshold. For example, the ratio of the number of seconds within the threshold value of each waist height at the time of trial in which five motion features are presented is compared. Here, the ratio is determined with the ratio of the number of seconds within the threshold when presenting “waist height” as the denominator and the ratio of the number of seconds within the threshold when presenting other feature points as the numerator. When this ratio is smaller than 1, it can be said that it is more effective to present the “high waist” than to present other compared feature points. Conversely, it is effective to present other feature points. It can be said that there is a side effect of lowering. Also, if the ratio is greater than 1, it can be said that it is more effective to present other compared feature points than presenting “waist height” directly, in other words, presenting other feature points It can be said that the desired effect can be obtained.

The above calculation is performed for each motion feature and each subject, and the results are shown in Table 2. The column shows the motion feature and the row shows the motion feature to be compared. For example, in subject A's waist height row, the feature of interest of attention is waist height, and from left to right, waist height, width of both legs, hand height, body orientation, sword The ratio of the values when presenting how to swing is shown.

It has been found that the feature point that increases the effect of the motion feature value to be compared most is the same motion feature as the feature value, for example, it is the presentation of the waist height that increases the effect of the waist height. That is, in many cases, the movement feature amount is changed in the target direction when the state of the corresponding feature movement feature amount is presented. The part with a value lower than 1 in the table corresponds to that. On the other hand, when the value is larger than 1, for example, with respect to the “body direction” of the subject A, it is more effective to present “the waist height” and “the width of both legs” than the “body direction”. It means that

Relatively evaluate the effect status of each motion feature by presentation. The comparison method is shown in Table 3, with 1 being when it is above and 0 when it is below the average of each motion feature. In other words, the case of 1 is relatively effective, and the case of 0 is relatively ineffective. From this table, it can be seen that when all three subjects are presented, many feature values exceed the average, which is effective. In the case of subject B, all feature values are above the average value. In the case of subject C, all feature values except for the height of the hand are above the average value.

When presenting the target feature amount, the evaluation value of the target feature amount may be higher than when other feature amounts are presented, and the evaluation value may be higher than the average for other feature amounts. I knew that there was. In the case of subject A, when the waist height is presented, the waist height feature exceeds the average, but the waist height feature is also present when both legs are presented and all are presented. Can be seen to exceed the average. In the case of subject A, when the width of both legs is shown, all the features except for the swing of the sword exceed the average. The evaluation of many feature quantities is increasing by presenting an object with one motion feature quantity. Similarly, in the case of subject C, when the width of both feet is shown, all the feature values excluding the body direction exceed the average. That is, it can be seen that subjects A and C are also effective when the widths of both feet are presented. Conversely, for subject B, presentation of the width of both legs is not effective, and presentation of the hand height is relatively effective. Thus, the effective presentation method changes with subjects.

(B) Correlation 2 between motion features
Similar to the analysis by the ratio of the number of seconds within the threshold, the correlation of the movement feature amount between the movement feature presentations is evaluated. The value of the evaluation index of the feature amount corresponding to each motion feature presentation is set to 1, and the ratio of the evaluation indices in other feature point presentation is taken. In the case of the slashing movement, the evaluation index is either the reciprocal of the variance or the average of the maximum values depending on the target movement feature amount. This will be specifically described below.

(Waist height)
The lower the variance of the waist height, the better. Therefore, the evaluation index is the reciprocal of the variance value of waist height. The larger the evaluation index, the greater the effect. The value of the evaluation index when presenting the feature amount of the waist height is divided by the value of the evaluation index when presenting the other feature amount. When the value is 1 or more, the presentation effect of the other feature amount presentation is greater than the presentation of the waist height.

(Width of both feet)
The larger the average value of the maximum width of both feet, the better. Therefore, the evaluation index is an average value of the maximum values of the widths of both feet. The larger the evaluation index, the greater the effect. The value of the evaluation index at the time of presentation of the feature amount of the width of both legs is divided by the value of the evaluation index at the time of presentation of the feature amount of the other foot. When the value is 1 or more, the presentation effect is larger in the presentation of other feature amounts than in the presentation of the width of both feet.

(Hand height)
The larger the average value of the maximum hand height, the better. Therefore, the evaluation index is an average value of the maximum values of the hand height. The larger the evaluation index, the greater the effect. The value of the evaluation index at the time of presenting the feature quantity is divided by the value of the evaluation index at the time of presentation of the feature quantity of the hand height. When the value is 1 or more, the presentation effect of the other feature amount presentation is greater than the presentation of the hand height.

(Direction of body)
The larger the difference between the maximum and minimum body orientation during exercise, the better. Therefore, the evaluation index is an average value of the difference between the maximum value and the minimum value of the body direction during exercise. The larger the evaluation index, the greater the effect. The value of the evaluation index when presenting the feature amount of the body orientation is divided by the value of the evaluation index when presenting the other feature amount. When the value is 1 or more, the presentation effect of the other feature amount is greater than the presentation of the body orientation.

(Sword swing)
Taking the x-axis to the left and right of the athlete's body, the greater the swing width of the sword expressed by the difference between the maximum and minimum x-coordinates of the sword tip when swinging the sword down. Therefore, the evaluation index is an average value of the difference between the maximum value and the minimum value of the x coordinate of the sword tip. The larger the evaluation index, the greater the effect. The value of the evaluation index when presenting the feature amount of the body orientation is divided by the value of the evaluation index when presenting the other feature amount. When the value is 1 or more, the presentation effect of the other feature amount is greater than the presentation of the body orientation.

The results are shown in Table 4. When the value is larger than 1, the same or higher effect can be obtained by other motor feature presentation, and when it is smaller than 1, the effect diminishes by other motor feature presentation. It is less than 1 in most points, and it can be seen that the effect of the motion feature is improved when the object of the corresponding motion feature is presented. However, there are not a few cases that are larger than 1, for example, regarding the height of the hand of subject A, it can be seen that even if other movement characteristics such as the width of both feet and the body orientation are presented, subject B When all feature points are presented, the effect is improved with regard to hand height, body orientation, and sword height.

Relatively evaluate the effect status of each motion feature by presentation. As in the case of the number of seconds within the threshold, the comparison method evaluates whether it is above or below the average of the evaluation values in Table 4. Table 5 shows the case where the value is above 1, and the case where the value is below 0. When all motion feature objects were presented, many other motion feature values exceeded the average, and it was found that the effect could be seen by presenting multiple motion feature objects simultaneously. Also, it can be seen that the body orientation and sword swing presentation are effective for some subjects. As shown in the width of both feet in the subject A, it can be seen that even the presentation of an object having a single motion feature may be effective for a plurality of motion feature amounts, and such feature amounts differ depending on the subject.

Comparing Table 3 and Table 5, it can be seen that the same tendency is observed in the evaluation results of the presentation effect even when different evaluation methods for the presentation effect are used. The evaluation function defined by the number of seconds within the threshold described in (a) can easily obtain an evaluation value, but the evaluation value calculated directly from the motion feature value described in (a) depends on the threshold setting. It has the characteristic that it is not done. From the results of both, the following can be understood. First, it has been clarified that the presented motion feature value may affect the feature value that is not presented. When a state of a certain motion feature value is presented, in addition to the presented motion feature value approaching the target, a motion feature value that is not presented also changes simultaneously, that is, there is a correlation between the motion feature values. It has been clarified that the presented motion feature may affect the features that are not presented. It should be noted that even if the state of a single motion feature is presented, the state of another motion feature that is not presented may approach the target. Therefore, it is possible to effectively support a learner's motor learning by presenting a motor feature highly correlated with many other motor features. And even if it is the same exercise | movement, the correlation between exercise | movement feature-values changes with subjects.

By using this motor learning support system, it is possible to comprehensively examine the influence of the motor feature amount to be presented on other motor feature amounts. Therefore, by presenting, it is possible to identify an effective motion feature that can simultaneously bring a plurality of motion features close to the target, and the learner's motion is effective for the target motion as a whole. Can help you get closer. It is effective for each learner by evaluating the correlation of the exercise feature for each exercise and storing the correlation between multiple exercise features as a table corresponding to each learner. It is possible to present various motor features.

By using this motor learning support system, the motor characteristics of the target motor can be optimized. Prepare multiple motor feature candidates from experts' opinions, exercise analysis papers, books, magazines, information sources on the Internet, etc., and use each motor feature candidate to determine the degree of improvement of learner's motor learning and motor feature candidates It is also possible to evaluate the correlation between the two and optimize the motion feature candidate using the evaluation result to determine the motion feature.

[D] Motion and motion characteristics The motion characteristics have been discussed by taking the motion “slashing” as an example, but there are multiple motion features in every motion. Some examples of motion and motion characteristics are given below. In the following, (a) points of motion, (b) points of motion are expressed as feature quantities in motion measurement (motion capture), and (c) tips and points of movement are expressed by changes in feature quantities. Although a specific formula for calculating the motion feature is omitted, for example, when measuring motion using optical motion capture, which motion feature is given, which one is used to calculate the desired motion feature? A person skilled in the art can appropriately design whether to attach a marker to a site, how to calculate a motion feature amount from measured motion data, or the like.

[D-1] Basketball (3P shot)
(A) “The axis of the trunk does not shift”, “Draws an arch with a high ball trajectory”, “Backspin”, “Always apply a constant force”, etc.
(B) “The center of gravity of the waist is the center of gravity of the four markers on the waist”, “The elbow, wrist, shoulder and knee angles and angular velocities are measured with the three markers”.
(C) “The center of gravity of the waist is not shifted before and after the release”, “The wrist angle varies greatly between before and after the release”.

[D-2] Basketball (Jump Shoot)
(A) “Sturdy ball overhead form”, “Cannot open the elbow outside”, etc.
(B) “The center of the head marker and the center of the right hand marker are at the same height”, “An angle between the plane passing through the right arm marker and the plane passing through the four markers on the waist”.
(C) “The smaller the difference in height between the center of the head marker and the center of the right hand marker, the better.” “The angle between the plane passing through the right arm marker and the plane passing through the four markers on the waist is 90. “The closer it is, the better.”

[D-3] Baseball batting 1
(A) “Make the legs with the lower body”, “Fold your arms when swinging and hit the elbow against the ball”, “Do not lower the head”, etc.
(B) “The angle of the upper body is represented by a vector connecting the back of the chest and the front marker, the angle of the lower body is represented by the vector connecting the back of the waist and the front marker, and the angle formed by these vectors “Represents the torsional size of the upper body”, “Head speed at the tip of the bat”, “Relationship between the position of the head and the grip during the swing”, “Distance between the trunk and arm marker at the start of swing”, and so on.
(C) “The larger the angle formed by the vector representing the angle between the lower body and the upper body, the better.” “It is better if the head speed is faster”, “It is not better if the head is too lowered with respect to the grip”, “When swing starts The smaller the distance between the trunk and arm markers, the better. "

[D-4] Baseball batting 2
(A) “Batting by putting weight on the shaft foot”, “Centrifugal force is generated and the flight distance is longer when the bat head is used and rotated around”, etc.
(B) “Measure the positional relationship between the center of gravity and the shaft foot”, “Measure the trajectory of the bat”, etc.
(C) “The center of gravity should be on the foot”, “The radius of the locus of the bat should be large”, etc.

[D-5] Robot dance (A) “Fix joints”, “Never move parts that do not move”, etc.
(B) “Attach markers to each joint and measure position and movement”.
(C) “The joint marker that you decided not to move does not move”, “When you rotate around a joint, the relative position of the joint marker ahead of it is maintained (ex. When moving the arm around the center, the triangle shape of the shoulder, elbow, and wrist is maintained. "

[D-6] Tennis (Forehand stroke 1)
(A) “Shoulder rotation (put right shoulder forward, but do not brace body axis)”.
(B) “A marker is placed on the right and left shoulders and on the right and left of the waist, and the axis is a straight line perpendicular to the ground that passes through the center of gravity of the four points.”
(C) “Make sure the axis does not shake while swapping the right and left shoulders”.

[D-7] Tennis (Forehand stroke 2)
(A) “Strike with the twist of the waist”, “Strike the stance widely and hit the upper body so that it does not protrude”, etc.
(B) “Set the center of gravity of the waist to the center of gravity of the four markers on the waist”, “Create a rectangle with the four points of the marker on the feet”, etc.
(C) “It is good that the four hip markers rotate according to takeback → impact → follow-through”, “the center of gravity of the hips should be within the rectangle of the foot marker”, etc.

[D-8] Tennis (stroke)
(A) “Push back as many as four spheres are lined up”, “Turn off force”, “Has a sense of a single pillar from head to back”.
(B) “Markers are placed in a straight line from the top of the head, neck, back muscles, and waist”, “markers are attached to the tip and grip of the racket, and the two marks move in parallel”.
(C) “A straight line from the head to the back is maintained, so that it does not move sideways”.

[D-9] Tennis (serve)
(A) “Movement of the center of gravity”, “Raise the elbow firmly”, “Impact when the elbow is fully lifted and extended”, “Follow-through like throwing a racket”, etc.
(B) “Movement of the center of gravity”, “Elbow height”, “Elbow angle at impact”, “Racquet trajectory and speed”, etc.
(C) “Is the center of gravity moving in a timely manner”, “The elbow angle is closer to 180 degrees at the time of impact,” etc.

[D-10] Tennis (smash)
(A) “Landing with your right foot and holding the racket up”, “Stretching your back and hitting it as high as possible”, etc.
(B) “Outer product of shoulder-elbow vector and shoulder-elbow vector and shoulder-to-hand vector with markers on shoulder, elbow and hand” "Outer product of vectors on shoulders".
(C) “The closer the cross product value is to 0, the better”.

[D-11] Model walking (A) “Pull the chin lightly”, “Do not bend the knee of the person who is touching the ground”, “Do not put weight on the heel”, “Always rub the inside of both feet. "Walk while crossing the front and rear", "Shoulder the ribs inward", "Wave your hand greatly according to the stride".
(B) "Head direction vector and eye-to-gaze vector angle", "Foot-to-knee vector and knee-to-ankle vector angle", "Heel and toe on the ground `` Difference in height '', `` Minimum distance when moving the bone of the grounding foot and the bone of the reverse foot '', `` Distance between the nipple and both shoulders (compared to normal) '', `` Moving distance of the toe and fingertip in one-step exercise '' Compare distances ".
(C) “The smaller the difference between the direction of the gaze point and the head, the better”, “The smaller the difference between the two vectors of the foot, the better”, “The difference in height is 0 or more, the closer it is to a certain value”, “Shortest “The distance between the bones is better as the distance between the bones is better”, “The distance between the nipples and both shoulders should be a little wider than usual”, “The movement distance of the fingertips changes according to the movement distance of the toes ”

In addition to the above-mentioned motion characteristics based on books and oral traditions, motion characteristics can be extracted by using various motion analysis papers and motion analysis services. If a motion feature can be extracted, it will be understood that a person skilled in the art can appropriately design a calculation method for digitizing the motion capture data or the like.

[E] Presentation of Evaluation Result of Motor Feature Using Voice In the above embodiment, the motor learning support system for visually presenting the evaluation result of motor feature has been described. And may be presented audibly.

An embodiment in which the evaluation result of the motion feature is presented by using a voice will be described by taking a sword as an example. According to the master, if you can start swinging, if you swing it down with a breath afterwards, it will be used as a result, and as a result, it will naturally fit into the end position. If you do not use your full power, you will have to stop using your own power because the power will remain and will not stop even if you are at the end of the swing. It is also said that if you keep conscious of your posture during exercise, your movement tends to become stiff or slow. Therefore, in the chopping, a motion feature amount that characterizes the “swing start posture” (since the swing start posture is a part of the chopping motion, this motion feature amount is a motion feature that characterizes the chopping motion) and “the end of swinging” A motion feature that characterizes the “posture” (because the swing end posture is a part of the cut-off motion, this motion feature is a motion feature that characterizes the cut-off), which affects other motion feature values in the cut-off (That is, there is a correlation between the movement features). Therefore, the slashing is performed after practicing in advance so that the motion feature values in the “swing start posture” and “swing end posture” have the motion feature values of the master.

First, a plurality of motion features that characterize the “swing start posture” are prepared, and the motion feature values in the “swing start posture” of the master are acquired. The motion feature amount in the “swing start posture” of the motion learner is acquired in real time, and the motion feature amount in the “swing start posture” of the motion learner is evaluated using a motion feature evaluation function prepared in advance.

When practicing in advance, only when the motion feature value that characterizes the “swing start posture” of the motor learner approaches the motion feature value of the master “swing start posture” (within an appropriate range), It is assumed that the corresponding sound is presented, and the “swing start posture” is ready when the sounds corresponding to all the movement features are presented. For example, if the center of gravity position below the waist that is the feature amount is within the appropriate range, a sound of a de will be emitted, and if the center of gravity position of the whole body that is the feature amount is within the appropriate range, a sound of Mi will be emitted, When the center of gravity position above the waist, which is the feature amount, falls within an appropriate range, a sound of a saw is produced, and when a chord of the dormitory is heard, the swinging posture is ready.

Similarly, the sound corresponding to each motion feature is only when the motion feature that characterizes the “swing end posture” of the motor learner approaches the motion feature amount of the “swing end posture” of the master (within an appropriate range). It is assumed that when the sound corresponding to all the movement features is presented, the “swinging end posture” is ready. For example, if the center of gravity position below the waist, which is the feature amount, is within the range, a sound is heard, and if the center of gravity position of the whole body, which is the feature amount, is within the range, the sound of the fa When the center of gravity position above the waist is within the range, the sound of a la is heard, and when the referrer's chord is heard, the swing end posture is in place.

After the above exercises have been sufficiently performed and the body has learned the “starting swing posture” and “swinging end posture”, a series of slashing movements will be performed on the spot, and the sound of the original learning will be produced at the start and end of swinging. Practice repeatedly while checking whether or not. In the above example, the overall feature value of the center of gravity position is presented, resulting in other feature values (for example, waist height, width of both feet, hand height, waist orientation, sword swing width) Etc.) are also optimized.

In this embodiment, the evaluation result of only the motion feature in a certain important posture (in the above example, “swing start posture”, “swing end posture”) is presented by voice, and in the middle of “swing start” to “swing end” Evaluation of the movement characteristics of the posture is not performed (even if it is evaluated, it is not presented). Such an evaluation result can be presented by detecting the posture of the motion learner in real time and determining whether the posture of the motion learner is in an important posture.

The above-described method focusing on the “starting posture” is not limited to scissors cutting but can be applied to other exercises. For example, in tennis, “If you can move to a position where you can strike in the best position wherever the ball can fly and create a position in that position, you can hit it back freely.” Common. In addition, other sports and martial arts such as baseball batting and kendo beating should be applied to exercises that require quick exercises, especially those that involve quick exercises, especially during the exercises. It is thought that you can.

Using audio to present information to an exercise learner is as follows, compared to visually presenting information in real time during exercise (exercise feature quantity judgment results and exercise feature quantities). There are advantages.
● When visually presenting information (results of determination of motion feature values or motion feature values), it is necessary to direct a line of sight to a monitor or the like, which may hinder motion. By using sound, it is possible to present information that does not depend on the line of sight, and it can also be used for exercises that face down, such as golf.
● There is too much information to be obtained visually, and it is often difficult to grasp the judgment results of motion feature quantities. As described above, an exercise learner who is exercising can recognize a chord as long as three sounds are played at the same time, but it is difficult to recognize three pieces of visual information at the same time. In addition, it is easy to grasp the determination result of the movement feature value and the change of the movement feature value by using the sound.

By using voice presentation, it is possible to convert the motion feature value into voice and present it in real time. A motor learning support system that presents motion feature quantities in real time by voice will be described. The configuration of the motor learning support system is shown in FIG. A motion capture system is used to measure movement. The feature quantity of the exercise is calculated from the acquired data in real time, and the calculated exercise feature quantity is converted into a frequency and presented by voice. It is also possible to play a sound from the recorded motion data. XAudio2 was used for sound playback and frequency change. When analyzing data, using the created system, the feature amount is output as text data from the recorded data and analyzed.

In the motor learning support system, the frequency can be calculated from the height of the center of gravity of the body, the height of each part of the body, and the rotation angle of the waist to present the sound. Here, the feature amount changes with various movements, and it is easy to grasp the rhythm of the exercise, and the body centroid at which this system can be used for many exercises is used as the presentation feature amount. Calculate the body centroid from the mass ratio of each part of the body (analysis of body centroid of Mitsuo Nagasawa standing long jump Akita University Faculty of Education) (34), 58-68, 1984-02) Converted to frequency. A note is presented by truncating the fractional part of the exponent (Twelve Equal Temperament).
f is the output frequency [Hz], f ′ is the frequency [Hz] at the reference height, a is the degree of change in sound, h is the reference height [inch], and x is the height of the body center of gravity [inch]. To do.

The motor learning support system can be used as follows. First, the user learns the sound produced by the exercise of the advanced person (in the above example, the change in the height of the center of gravity presented by converting it into a sound) together with the visual sense. If you have imitated a similar sound (similar until you heard a similar sound), you have acquired an exercise with the characteristics of an expert. Memorize the sound when you are exercising well, and remember the feedback from muscles etc. while leaving the sound image in your head. Next, reproduce the movement while imagining the sound. The user can practice exercise while confirming the deviation from the advanced user by voice.

In order to investigate the usefulness and problems of the voice learning support system, the motor learning support system was used for learning the truncation exercise (the body's center of gravity changes greatly with exercise).

The experimental method is as follows. Compare the slashing movements of three beginners using the motor learning support system and three beginners not using it. The voice was presented using a speaker of a notebook computer, and a and f ′ in equation (1) were a = 200, f ′ = 440, and h was the height of the center of gravity of the body at the start of the experiment. The experiment consists of 3 trials. The subjects were told that the point of the slashing movement was to make the sword trajectory slash the sword and only raise the hand when raising the sword.
[Trial 1]
After explaining the tips for exercise that is not related to waist height, show the exercise of an advanced person and perform minimal practice. Then measure.
[Trial 2]
When presenting speech: Instruct to learn speech from advanced exercise, practice while presenting speech in real time, and then measure again with speech.
When not presenting speech: Practice and measurement after showing the exercise of an advanced person again.
[Trial 3]
Measure again without making a voice, and finally conduct a questionnaire.

We measured the height of the waist, the distance between both feet, the speed of the hand, and the swing period, which are the differences between the advanced and beginners in the slashing movement. As a result of the experiment, it was found that by adjusting the whole body motion with only one global feature quantity called the center of gravity of the whole body, the individual feature quantity changed as a result. When the voice was presented at the center of gravity of the body, the feature amount approached the value of the advanced player, but it was found that the feature amount approaching greatly differs depending on the subject. For example, at the lowest waist height during swing, the change was significant in subjects with sound presentation. In the questionnaire, the opinion that the difference with the exercise of the advanced person was recognized more strongly by voice.

The present invention can be used for exercise learning support in a fitness club or the like.

Claims (16)

A means of measuring the movement of a motor learner;
A plurality of motion features that characterize the motion are set, and in order to calculate each motion feature as a motion feature amount, a motion feature amount calculation function prepared for each motion feature,
Means for calculating a plurality of motion feature amounts from the motion measurement result and the motion feature amount calculation function;
In order to evaluate the calculated motion feature value, a motion feature evaluation function prepared for each motion feature value,
Means for evaluating the calculated motion feature amount using a corresponding motion feature evaluation function;
Means for presenting the evaluation results to the motor learner in a recognizable manner in real time;
Means for storing a correlation between the plurality of motion feature quantities;
With
The correlation between the plurality of movement feature quantities defines the presentation effect that the presentation of the evaluation result of one or more movement feature quantities has on the movement feature quantity on which the evaluation result is presented and other movement feature quantities And
According to the correlation between the plurality of movement feature quantities, one or a plurality of movement feature quantities are selected from the plurality of movement feature quantities, and the evaluation results of the selected movement feature quantities are presented to the exercise learner in real time. A motor learning support device, characterized by: Correlations between the plurality of motion feature quantities are stored corresponding to each learner,
According to the learner-specific correlation between the plurality of motor feature quantities, one or a plurality of motor feature quantities are selected from the plurality of motor feature quantities, and the evaluation results of the selected motor feature quantities are sent to the motor learner. The motor learning support device according to claim 1, wherein the motor learning support device is presented in real time.   The selected one or more exercise feature values are obtained by presenting the evaluation result of the one or more exercise feature values to the learner, in addition to the exercise feature value in which the evaluation result is presented, The motor learning support device according to claim 1, wherein the motor learning support device is selected so as to exhibit a presentation effect in one or a plurality of other motor feature quantities that are not presented.   The learning support device includes means for storing a reference motion feature used for comparing the calculated motion feature with the calculated motion feature when evaluating the calculated motion feature according to the motion feature evaluation function. The motor learning support device according to claim 1. The evaluation means determines whether the calculated motion feature value is “within an appropriate range” or “out of an appropriate range”;
The evaluation result is either “the calculated motion feature value is within an appropriate range” or “the calculated motion feature value is outside the appropriate range”.
The motor learning support device according to claim 1. The learning support device includes display means for displaying an evaluation result,
6. An object corresponding to each motion feature is prepared, displayed on the display means so as to be distinguishable from each other, and the evaluation result is presented to a learner by a change of each object. The motor learning support apparatus described.   The motor learning support device according to claim 1, wherein sounds that can be distinguished from each other are prepared in correspondence with each motor feature, and the evaluation result is presented by sound. A plurality of motion features that characterize the target motion are set, and in order to calculate each motion feature as a motion feature amount, a motion feature amount calculation function is prepared for each motion feature,
In order to evaluate the calculated motion feature value, a motion feature evaluation function is prepared for each motion feature value.
Correspondence between multiple motion feature values that specify the effect of presentation of the evaluation result of one or more motion feature values on the motion feature value for which the evaluation results are presented and other motion feature values is provided Has been
According to the correlation between a plurality of motion feature quantities characterizing the target motion, one or more motion feature quantities are selected from the plurality of motion feature quantities,
Measuring an exercise learner's exercise, calculating one or more selected exercise feature quantities using an exercise measurement result and an exercise feature quantity calculation function;
A motor learning support method that evaluates a calculated motor feature using a corresponding motor feature evaluation function and presents the evaluation result to a motor learner in a recognizable manner in real time. The correlation between the plurality of movement feature quantities is prepared for each learner,
9. The motor learning support method according to claim 8, wherein one or a plurality of motor feature quantities are selected from the plurality of motor feature quantities according to a learner-specific correlation between the plurality of motor feature quantities. .   The selected one or more exercise feature values are obtained by presenting the evaluation result of the one or more exercise feature values to the learner, in addition to the exercise feature value in which the evaluation result is presented, 10. The motor learning support method according to claim 8, wherein selection is made so as to exhibit a presentation effect in one or a plurality of other motor feature quantities that are not presented.   The object corresponding to each movement feature amount is prepared and displayed on the display means so as to be distinguishable from each other, and the evaluation result is presented to the learner by a change of each object. The motor learning support method described.   11. The motor learning support method according to claim 8, wherein sounds that can be distinguished from each other are prepared corresponding to each motor feature quantity, and the evaluation result is presented by sound. A means of measuring the movement of a motor learner;
A plurality of motion features that characterize the motion are set, and in order to calculate each motion feature as a motion feature amount, a motion feature amount calculation function prepared for each motion feature,
Means for calculating a plurality of motion feature amounts from the motion measurement result and the motion feature amount calculation function;
In order to evaluate the calculated motion feature value, a motion feature evaluation function prepared for each motion feature value,
Means for evaluating the calculated motion feature amount using a corresponding motion feature evaluation function;
Means for presenting the evaluation results to the motor learner in a recognizable manner in real time;
Means for storing the evaluation results;
With
One or a plurality of movement feature quantities are selected from the plurality of movement feature quantities, and the evaluation results of the selected movement feature quantities are presented to the movement learner in real time, and the evaluation results of other movement feature quantities that are not presented The evaluation result of the selected motion feature amount is converted into the motion feature amount for which the evaluation result is presented and the other motion feature amount using the evaluation result that has been presented and the evaluation result that has not been presented. An apparatus for acquiring a correlation between motor feature quantities used for motor learning support, characterized by acquiring a correlation between a plurality of motor feature quantities that define a presentation effect to be given peculiar to a motor and / or a learner.   The learning support device includes means for storing a reference motion feature used for comparing the calculated motion feature with the calculated motion feature when evaluating the calculated motion feature according to the motion feature evaluation function. The apparatus for acquiring correlation of motion feature values according to claim 13. The evaluation means determines whether the calculated motion feature value is “within an appropriate range” or “out of an appropriate range”;
The evaluation result is either “the calculated motion feature value is within an appropriate range” or “the calculated motion feature value is outside the appropriate range”.
15. The apparatus for acquiring correlation between motion feature values according to claim 13. A plurality of motion features that characterize the target motion are set, and in order to calculate each motion feature as a motion feature amount, a motion feature amount calculation function is prepared for each motion feature,
In order to evaluate the calculated motion feature value, a motion feature evaluation function is prepared for each motion feature value.
Measure the movement of the movement learner, calculate a plurality of movement feature quantities that characterize the movement using the movement measurement result and the movement feature amount calculation function,
Evaluate the calculated motion feature using a corresponding motion feature evaluation function,
One or a plurality of movement feature quantities are selected from the plurality of movement feature quantities, and the evaluation results of the selected movement feature quantities are presented to the movement learner in real time, and the evaluation results of other movement feature quantities that are not presented The evaluation result of the selected motion feature amount is converted into the motion feature amount for which the evaluation result is presented and the other motion feature amount using the evaluation result that has been presented and the evaluation result that has not been presented. A method for acquiring a correlation between motor feature quantities used for motor learning support, which includes acquiring a correlation between a plurality of motor feature quantities that define a presentation effect to be given peculiar to a motor and / or a learner.