WO2020241738A1 - Training assisting method and device - Google Patents

Training assisting method and device Download PDF

Info

Publication number
WO2020241738A1
WO2020241738A1 PCT/JP2020/021085 JP2020021085W WO2020241738A1 WO 2020241738 A1 WO2020241738 A1 WO 2020241738A1 JP 2020021085 W JP2020021085 W JP 2020021085W WO 2020241738 A1 WO2020241738 A1 WO 2020241738A1
Authority
WO
WIPO (PCT)
Prior art keywords
difference
sequence
learner
motion
acoustic information
Prior art date
Application number
PCT/JP2020/021085
Other languages
French (fr)
Japanese (ja)
Inventor
中村 仁彦
江 山本
洋介 池上
玲至 田中
彬光 櫻井
文香 山田
Original Assignee
国立大学法人東京大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 国立大学法人東京大学 filed Critical 国立大学法人東京大学
Publication of WO2020241738A1 publication Critical patent/WO2020241738A1/en

Links

Images

Classifications

    • AHUMAN NECESSITIES
    • A63SPORTS; GAMES; AMUSEMENTS
    • A63BAPPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
    • A63B69/00Training appliances or apparatus for special sports
    • A63B69/36Training appliances or apparatus for special sports for golf
    • AHUMAN NECESSITIES
    • A63SPORTS; GAMES; AMUSEMENTS
    • A63BAPPARATUS FOR PHYSICAL TRAINING, GYMNASTICS, SWIMMING, CLIMBING, OR FENCING; BALL GAMES; TRAINING EQUIPMENT
    • A63B71/00Games or sports accessories not covered in groups A63B1/00 - A63B69/00
    • A63B71/06Indicating or scoring devices for games or players, or for other sports activities
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00Image analysis
    • G06T7/20Analysis of motion

Definitions

  • the present invention relates to training support methods and devices.
  • Non-Patent Document 1 attempted to measure the position and angle of joints in a swing and to quantify the norms and skills expressed in the golf instruction book (Non-Patent Document 2).
  • Non-Patent Document 3 modeled the arm and golf club during a golf swing as a double pendulum
  • Inoue et al. investigated the relationship between the joint torque of the arm and wrist and the swing. The mode analysis of the double pendulum model was performed for the purpose of. Tinmark et al.
  • Non-Patent Document 5 focused on three parts, the hand, shoulder, and waist, and investigated the relationship between the order in which the rotation speed of each part was maximized and the speed of the head.
  • Vena et al. investigated the relationship between the timing at which the rotational speeds of the left arm, shoulder, and hip are maximized and the skill level.
  • a golf swing is a movement that is completed while a plurality of body parts interact with each other in a short time, and it is difficult to visually grasp the chain of movements of the body parts even if the swing movie is played.
  • Patent Documents 1 to 3 It has been conventionally known to use sound for golf practice, and it is described in, for example, Patent Documents 1 to 3. The methods described in these documents allow the learner to swing in time with the sound, and do not evaluate the learner's swing.
  • the present inventors have tackled the problem of how to calculate, compare, and convey this to the learner, especially in physical exercise in which the sequence of movement is important.
  • the technical means adopted by the present invention is Prepare the first movement sequence of physical exercise as a reference, Obtain the second motion sequence from the exercise data of the learner's physical exercise, Presenting the difference (difference) between the first motion sequence and the second motion sequence to the learner as acoustic information.
  • Training support method One embodiment comprises obtaining the difference between the first and second motion sequences. The acquired difference is presented to the learner as acoustic information.
  • the motion sequence is defined by a plurality of events that occur at different times.
  • One embodiment comprises obtaining the difference between the first and second motion sequences.
  • the difference includes the difference in the order in which the events occur and / or the timing in which the events occur.
  • each of the events is assigned a sound.
  • the first operation sequence and the second operation sequence can be converted into acoustic information (sound time series information).
  • the sound of the event in the wrong order is changed when the second acoustic information is presented.
  • the learner is provided with a means for converting the difference into acoustic information, and the difference converted into acoustic information is presented to the learner.
  • the difference between the first motion sequence and the second motion sequence is presented to the learner in a visually recognizable manner. For example, a diagram comparing the first operation sequence and the second operation sequence (for example, the difference may be emphasized) may be displayed. Alternatively, the analysis result of the difference between the first operation sequence and the second operation sequence may be illustrated. The frame image corresponding to the event with a large difference may be displayed. Further, the swing image may be presented at the same time as the presentation of the sound sequence.
  • the training support device adopted by the present invention is A means of storing the first motion sequence of physical exercise as a reference, A means of acquiring a second motion sequence from the learner's physical motion data, It is provided with a means for presenting the difference (difference) between the first operation sequence and the second operation sequence to the learner as acoustic information.
  • the means for obtaining the difference between the first operation sequence and the second operation sequence is included.
  • the presenting means presents the acquired difference to the learner as acoustic information.
  • the sequence of actions is defined by multiple events that occur at different times. It is equipped with a means for acquiring the event occurrence time.
  • the difference includes a difference in the order in which the events of the first operation sequence and the second operation sequence occur, and / or the timing of occurrence of the events. Includes means for obtaining the event occurrence order and / or the difference in event occurrence timing.
  • each of the events is assigned a sound.
  • a means for converting the first operation sequence and the second operation sequence into acoustic information (time-series information of sound) is provided.
  • the learner is provided with a means for converting the difference into acoustic information, and the difference converted into acoustic information is presented to the learner.
  • the present invention can also be provided as a computer program that causes a computer to function as each of the above means.
  • the physical exercise is a golf swing.
  • the present invention is a method for quantitatively evaluating a learner's movement, and can be used as a learner's self-training support means and a training guidance support means.
  • the learner can perform his / her own movement by presenting the difference between the reference movement sequence and the learner's movement sequence acquired from the movement data to the learner as acoustic information in real time or near real time. It provides an intuitive way to be aware and make corrections.
  • the accuracy and validity of one's own physical exercise (for example, golf swing) can be fed back to the learner in real time or near real time.
  • the present invention can be used to support training of a golf swing, but the present invention is not limited to the golf swing, and the training effect can be enhanced in other exercise training, dance training, rehabilitation and the like.
  • FIG. 1 It is a figure which shows the axis (k, l) of a body. It is a figure which shows the shoulder line segment, the waist line segment, both arms, and a golf club. The definitions of the left thigh and left lower limb are shown. It is a figure which shows the definition of the left forearm vector and the club shaft vector. It is a figure which shows a series of movements of a golf swing. The figure on the left shows the first event (start of left knee extension: time when the angle of the left knee starts to increase), and the figure on the right shows the second event (release of left wrist fixation: time when the left wrist is returned).
  • the figure on the left shows the third event (waist brake: time when the rotation angular velocity of the waist reaches the maximum), and the figure on the right shows the fourth event (maximum hand speed: time when the center position of the wrist reaches the maximum speed). is there.
  • the figure on the left shows the fifth event (maximum head speed: time when the maximum speed of the club head), and the figure on the right shows the sixth event (shoulder brake: time when the rotation angular velocity of the shoulder becomes maximum). It shows the conversion of the time series of events into a time series of sounds, and the audio intervals are evenly spaced.
  • Example 1 Shows the operation sequence converted into acoustic information
  • the model is the reference operation data
  • Example 1 is when only the occurrence timing of some events is shifted (no order error)
  • Example 2 is the order of some events It corresponds to the case where is replaced.
  • It is a figure exemplifying the deviation information obtained from the deviation of the order, and the position of the wrong event, the phase difference of each event, and the event of the maximum phase difference are obtained.
  • It is a figure which shows the presentation of acoustic information paying attention to the event of the maximum phase difference.
  • Kinematic Sequence Kinematic sequence
  • the training support system according to the present embodiment includes a moving image acquisition unit, a storage unit for storing various data, a processing unit for executing various calculations, an image output unit (display), and an audio output unit.
  • the video acquisition unit is composed of a video camera
  • the storage unit, processing unit, image output unit, and audio output unit include one or more computers (input unit, arithmetic unit consisting of CPU, RAM, ROM). It is composed of a storage unit and an output unit).
  • motion capture The means for acquiring time-series data of physical exercise is so-called motion capture, and in one aspect, motion capture is a motion acquisition unit (video camera) that acquires moving image data of a target motion. ) And an action acquisition unit that acquires time-series data representing the target action based on the moving image data.
  • the moving image data acquired by the camera and the time-series data of the posture of the target acquired by motion capture are stored in the storage unit.
  • the movement of the target is defined by the time series data of the posture of the target acquired by motion capture.
  • the posture of the target is specified by a plurality of feature points (typically joints) on the target body, and by acquiring three-dimensional coordinate values of the plurality of feature points in each frame, three of the plurality of feature points are obtained.
  • the target operation is defined from the time-series data of the dimensional coordinate values.
  • the type of motion capture used in the present invention is not limited, and an optical motion capture using an optical marker that identifies a feature point, a so-called inertial sensor such as an acceleration sensor, a gyroscope, or a geomagnetic sensor is attached to the target body.
  • a method of acquiring the target motion data a so-called markerless motion capture in which an optical marker or a sensor is not attached, and the like can be exemplified.
  • motion measurement is performed by optical motion capture using an optical marker, but markerless motion capture is advantageous from the viewpoint of not interfering with the natural motion of the target.
  • markerless motion capture motion capture using a system equipped with a camera and a depth sensor (represented by Kinect) or deep learning is used to analyze RGB images from one or more viewpoints and motion data. A motion capture to acquire can be exemplified.
  • the motion sequence of physical exercise is an event in whole body exercise in which a predetermined body part or a plurality of linked body parts express as a change in the relative positional relationship with other body parts.
  • the group is described as having an order relationship and a time relationship (timing). That is, the operation sequence is defined by a plurality of events that occur at different times.
  • the event is defined by the velocity (maximum velocity, maximum angular velocity, minimum angular velocity, etc.), position (angle, etc.) of one or more body parts.
  • the operation sequence acquisition means includes means for acquiring the occurrence time of each event.
  • the event occurrence time acquisition means acquires the occurrence time of each event that defines the learner's operation sequence, and stores the occurrence time of each event and the order of the events obtained from the occurrence time in the storage unit.
  • the reference motion sequences are E 1 , E 2 , E 3 , E 4 , E 5 and so on. It can be specified by the order of occurrence and the timing of occurrence.
  • the order of the events can be known from the occurrence time.
  • the occurrence timing (relative timing) of each event can be acquired from the occurrence time of each event.
  • the time between the occurrence of events in the physical exercise is relative because the time from the start to the end varies from person to person. Therefore, it is possible to compare the occurrence timings of each event by acquiring the start time point and the end time point of the physical exercise, sharing the entire time of the physical exercise, and relativizing the occurrence timing.
  • Event Occurrence Time Acquisition Means In the event occurrence time acquisition means, what kind of information is used and how the event occurrence time is detected may differ depending on the event.
  • the event occurrence time acquisition means uses time-series data of the target motion, and uses a predetermined angle, angular velocity, angular acceleration, velocity, rotation speed, acceleration, and position (relative to other parts) of one or more parts that define the event. Acquire the occurrence time of a predetermined event using the relative position, relative position with respect to an object other than the body, etc.), posture, center of gravity, etc. as conditions.
  • Sound source data is stored in the sound source data storage unit.
  • Sound source data when sound is assigned to each event that defines the operation sequence (reference operation sequence), or when sound is assigned to the difference between the reference operation sequence and the target person's operation sequence.
  • a sound stored as sound source data is assigned to each event in advance, and the operation sequence is converted into time-series information of the sound by the acoustic information conversion means of the operation sequence and output from the audio output unit.
  • each sound assigned to each event is a sound that can be distinguished from each other and that the sequence of sounds can be easily identified.
  • a familiar and known sound source may be used as a sound sequence (sound time series information) corresponding to the reference operation sequence.
  • the sound assigned to each event is independent of the sound assigned to each event (including the case where no sound is assigned to each event), and the sound is assigned to the difference, or the sound assigned to each event is assigned.
  • An example is the case where the sound is presented in a changed manner (for example, the pitch is raised or lowered by a semitone).
  • the reference operation sequence storage unit stores the time and order of occurrence of each event that defines the reference operation sequence.
  • the reference motion sequence is obtained from ideal motion data that serves as a model / model.
  • the physical exercise of the target to be a model is videotaped to acquire the exercise data, and the time-series data of the exercise data to be the model is acquired.
  • the occurrence times of a plurality of events that define the movement sequence of physical exercise are acquired, and the occurrence time and the occurrence order of each event are stored in the storage unit as a reference movement sequence.
  • Video of physical exercise and time-series data of posture are also stored in the storage unit.
  • a plurality of different reference operation data (event set that defines the operation sequence and the occurrence time and order of events), and select an operation sequence to be used as a reference from a plurality of candidates. Since the model physical exercise and the ideal physical exercise are not necessarily one, a plurality of reference motion sequences may be acquired and stored in the storage unit. In addition, the set of events that define a certain motion sequence is not always one set, and multiple types of motion sequences (at least some events are different) are acquired from the same physical exercise data as a reference and stored in the storage unit. You may keep it.
  • [A-7] Conversion of operation sequence to music information The step of converting an operation sequence into acoustic information is illustrated in FIG. First, a moving image of the movement when the target performs physical exercise is acquired, and time-series data of the movement of the target is acquired from the moving image data. Based on the time-series data of the target motion, the occurrence time and the occurrence order of each event that defines the motion sequence of the physical exercise are acquired. By assigning a sound to each event, the operation sequence is converted into acoustic information (sound time series information). Immediately after the subject performs physical exercise, the motion sequence can be converted into acoustic information and the acoustic information can be presented to the subject.
  • acoustic information sound time series information
  • the same interval as the actual event may be used, or the relative interval may be maintained and the time may be lengthened.
  • the time width of each event that defines the swing sequence and the time width of the presented sound sequence may be different (not swinging according to the sound).
  • predetermined sounds may be presented in a predetermined order at equal intervals.
  • the learner listens and compares the acoustic information corresponding to the reference motion sequence and the acoustic information corresponding to his / her own motion sequence.
  • a reference motion sequence for example, a golf swing
  • the learner's physical exercise for example, golf swing
  • a motion sequence is acquired from the learner's motion
  • the acquired motion sequence is converted into acoustic information and presented to the learner.
  • the pitch of the sound corresponding to the event having a different order from the reference motion sequence may be changed (FIG. 17 “Example 2”).
  • the learner is made to recognize the difference by comparing the acoustic information corresponding to the reference motion and the acoustic information corresponding to the own motion to the learner.
  • the timing and number of times that the acoustic information corresponding to the reference motion and the acoustic information corresponding to the own motion are presented to the learner are not limited. For example, by comparing the acoustic information obtained by converting the learner's motion sequence with the reference acoustic information for each swing, it is intuitive to know which swing is closer to or far from the reference swing depending on the degree of difference in hearing. Can be judged.
  • the action sequence comparison means compares the occurrence time and order of each event that defines the learner's action sequence with the occurrence time and order of each event that defines the reference action sequence. , The difference (difference) in the operation sequence is acquired, and the information on the difference in the operation sequence is stored in the storage unit.
  • the difference in the operation sequence includes, but is not limited to, the difference in the event occurrence order and the difference in the event occurrence timing.
  • the difference in the occurrence timing (occurrence interval) can be obtained, for example, by comparing the ratio of each occurrence interval to the total time of physical exercise.
  • items other than the motion sequence may be compared between the reference motion and the learner's motion.
  • Various items can be compared by using the time-series data of the posture that defines the movement as a reference and the time-series data of the posture that defines the movement of the learner. For example, the speeds of reference movements and learner movements (eg, head speed in a golf swing) may be compared.
  • the reference motion sequence and the learner's motion sequence are compared to obtain the difference, the difference is converted into output information (for example, acoustic information), and the output information (for example, acoustic information) is presented.
  • the difference in the operation sequence is, for example, a difference in the event occurrence order and a difference in the event occurrence timing.
  • the information on the difference in the order of occurrence includes, but is not limited to, the position of the wrong event, the phase difference of each event, and the event of the maximum phase difference (see FIG. 18).
  • the difference between the learner's motion sequence and the reference motion sequence is converted into acoustic information by, for example, an acoustic information conversion means of the difference in motion sequence, and is output as voice. It can be output from the unit.
  • the order in which the sound sequence is presented after the learner's action is the order corresponding to the actual learner's action sequence, and the sounds corresponding to the events whose order is different from the reference action sequence are changed and presented. More specifically, in listening to the sound corresponding to the reference action sequence and listening to the sound corresponding to the own action sequence, the sequence of events and the self action that define the reference action sequence are specified. When there is a difference in the order of the events that define the sequence, the order of the events is the order of the learner's action sequence (wrong order), and the sound corresponding to the wrong event is presented in a different manner.
  • the change in sound can be exemplified by raising or lowering a semitone by typically changing the pitch (Fig. 17, Example 2).
  • a dissonant sound when multiple sounds corresponding to the own action sequence are played at the same time.
  • a dissonant sound By changing the pitch), the sound corresponding to each event is played at the same time after the self-motion is completed, so that the presence or absence of a difference between the reference motion sequence and the self-motion sequence is immediately recognized. Can be done.
  • the order in which the sounds are presented after the learner's actions is the order corresponding to the reference action sequence, and the event that occurs in a different order from the reference action sequence (for example, the event that occurs first among a plurality of events in different orders). ) May be changed (for example, pitch bend) and presented.
  • the pitch of the event with the maximum phase difference may be changed and presented by using the order corresponding to the reference operation sequence (see FIG. 19). That is, the event of the maximum phase difference and the phase difference thereof can be used as the difference information.
  • the position of the maximum phase difference in the acoustic information is used, and the value of the phase difference is added as information to the sounds in the same place as the ideal order. If there are multiple events with the maximum phase difference, the one that is the first in the obtained order is adopted. By always giving priority to the one in front of you, you can fix what you are conscious of. Further, the pitch may be changed according to the magnitude of the phase difference.
  • the sound sequence corresponding to the reference operation sequence may be converted into equal intervals and presented to the learner (see FIG. 16).
  • the sound sequence corresponding to the reference action sequence is generated. It may be changed with respect to the timing (Fig. 17, Example 1).
  • the sound sequence of the model data is sounded at equal intervals (0.2 seconds), while the sound sequence of the evaluation target data is delayed at the timing of a certain event, for example, at equal intervals. It may be delayed by 0.1 seconds, and if it is early, it may be sounded 0.1 seconds earlier than the equal interval. If there is a difference in the event occurrence order, it may be set not to evaluate the difference in the event occurrence interval, or even if there is a difference in the event occurrence order, it occurs. The difference in intervals may be evaluated.
  • the conversion of the difference in the motion sequence into acoustic information is not limited to the above example, and for example, during the learner's movement (for example, swing), the event of the motion sequence is referred to as the event of the motion sequence.
  • the sound may be played in real time at a time different from the order of.
  • the name for specifying the event and the corresponding body part may be presented by voice, and the event may be presented.
  • An image of the frame at the time of occurrence may be presented.
  • the frame image corresponding to each event may be displayed, or the frame image corresponding to each event may be displayed in comparison with the reference frame image and the learner's frame image.
  • the image corresponding to each event may be displayed alone or in comparison with the reference as an image having skeletal information and musculoskeletal information.
  • Akihiko Murai, Kosuke Kurosaki, Katsu Yamane and Yoshihiko Nakamura. “Musculoskeletal-see-throuhg mirror: Computational modeling and algorithm for whole-body muscle activity visualization in physiology in real time” 310-317, 2010) may be used to present a series of skeletal movements in a swing. In this way, by visualizing the movement of the skeleton of the whole body in a series of swing movements, it is possible to encourage the player to understand his / her swing.
  • an image of the frame at that time may be presented to encourage the user to pay close attention to the motion of the body part.
  • the contrast or difference between the learner's motion sequence and the reference motion sequence may be visually presented to the learner using a diagram or a graph.
  • the learner's physical exercise motion analysis results (which may include comparison with reference exercise data) may be displayed using figures and graphs.
  • Target physical exercise The target physical exercise is not limited, but a sequence of golf swing, baseball bat swing, baseball pitching motion, karate kata, dance, rehabilitation, and other motions is established. It is possible to exemplify exercises that are being performed and exercises that have a fixed routine. For example, golf is one of the sports for which a sequence of movements has been established. In dance practice and rehabilitation, training starts from the point of acquiring a sequence of movements.
  • the operation sequences are (1) start of left knee extension, (2) release of left wrist fixation, (3) waist brake, and (4) hand. It consists of 6 events: maximum speed of (5) maximum speed of head, (6) brake of shoulder, and each uses the time of angle change of a predetermined part and the time of maximum speed of a predetermined part. Get the time of the event.
  • the time of each event is calculated from the time series data of the three-dimensional position information of each marker measured by the optical motion capture, using the time of the angle change and the time of the maximum speed.
  • the correct answer for a set of events of interest is not necessarily one, and the above six events are examples.
  • the reference operation sequence may be composed of some of these events, or the reference operation sequence may be composed of other events. It will be appreciated by those skilled in the art that a set of other events that identify the kinematic sequence may be selected. Further, as long as the time series data of the posture of the target is obtained, it is possible to specify the occurrence time of any event.
  • the body movement during the swing is measured by optical motion capture, and the swing movement data is based on the time change of the marker position attached to the whole body and the golf club.
  • the feature points of interest in the whole body and the golf club and the golf swing process are defined as follows. The selection and definition of these feature points are examples and do not limit the contents of the present invention.
  • the left and right positions of the pelvis are defined as the left hip position x l, hi and the right hip position x r, hi, respectively, and the line segment connecting these two points is defined as the waist line segment (Fig. 9).
  • the upper body axis is the line segment connecting the two points of the midpoint x u, ax of the shoulder line segment and the midpoint x l, ax of the waist line segment (Fig. 9).
  • the positions of the outer side x lo, wr and the inner side x li, wr of the left wrist are averaged, and the position is defined as the center position x l, wr of the left wrist.
  • the line segment connecting the two points of the center of the left elbow x l, el and the center of the left wrist x l, wr is the left forearm (Fig. 11).
  • Address time t add The time to determine the initial posture and start swinging the golf club backwards. That is, it is defined as the time when the face center position x fa starts to move from the stationary state in the opposite direction to the hit ball.
  • Finish time t fin The end time of a series of swing movements. It is defined as the time when the movement speed of the face center position x fa after impact is the smallest.
  • the golf ball was actually hit and the marker position of the whole body in the movement was measured. By attaching a marker to the driver to be used, the movement of the club was also measured at the same time. In addition, the number of shots measured was 12, and the number of hit balls was also measured using a ballistic measuring device for each shot. After each shot, the shot feeling and swing posture were self-evaluated, and the one with a high self-evaluation score was regarded as a successful shot, and the one with a low self-evaluation score was regarded as a failed shot. The trial with the highest score was used as a model shot. In golf instruction, we focused on the body parts that are most noticeable when visually evaluating shots, and performed motion analysis. After that, the success swing and the failure swing are compared, and the factors that strongly affect the swing performance are adopted.
  • the collapse of the upper body axis in the opposite direction to the hitting ball is a change brought about by the braking and the movement of the waist to the rotational movement of the whole body, and it also has a meaning to prevent the center of gravity of the whole body from excessive movement in the hitting direction by the rotational movement. it seems to do.
  • the change in the angle between the hand and the golf club also suggests the implication of increasing the speed of the golf club head.
  • Waist rotation For the waist line segment defined in the previous section (for example, the line segment connecting the markers attached to the left side of the pelvis and the right side of the pelvis), at an arbitrary time based on the data (line segment) at the time of addressing.
  • Each of the waist line segment and the reference line segment is projected onto the horizontal plane, the angle formed by the two line segments at each time is calculated, and the value is defined as the rotation angle of the waist.
  • the time when the rotational angular velocity of the waist becomes maximum is set as the attention event.
  • the line segment of the shoulder defined in the previous section (the average of the front position of the right shoulder and the rear position of the right shoulder is the center position of the right shoulder, and the front position of the left shoulder and the rear position of the left shoulder The average is the center position of the left shoulder.
  • the shoulder is formed from the angle formed by the reference (line segment at the address time) and the projection of the line segment on the horizontal plane at any time. Define the rotation angle of. In the present embodiment, the time when the rotational angular velocity of the shoulder becomes maximum is set as the event of interest.
  • the angle between the left thigh and the left lower leg is defined as the angle of the left knee.
  • the state in which the knee is bent most immediately before kicking up the floor, that is, the time when the angle of the left knee is minimized is set as a noteworthy event.
  • the timing of the event of interest is the start of extension of the left knee, the release of the left wrist, the brake of the waist, the maximum speed of the hand, the maximum speed of the golf club head, the impact, and the brake of the shoulder.
  • the moment of impact almost coincided with the time when the speed of the golf club head was maximum.
  • the time elapsed from the start of extension of the left knee to the return of the shoulder was almost the same, but the timing of releasing the fixation of the left wrist was uneven.
  • the event occurred at almost the same timing except for the timing of releasing the fixation of the left wrist, while in the unsuccessful shot, the interval from the start of extension of the left knee to the return of the hip was shortened. It was. From this result, it was confirmed that the investigation of the timing of the extracted events is effective for the evaluation of the golf swing.
  • M is a matrix represented by the following.
  • each event is associated with a different frequency f ref i sound.
  • h ref i expresses the pitch difference from the sound of A (frequency 440 Hz) by the number of semitones.
  • a chord called C11 is set. Table 4 shows the assignment of sounds to each event.
  • the sound is associated with each event.
  • the pitch is shifted up by a semitone, and if it is late, the pitch is shifted down by a semitone.
  • the timing evaluation of the swing of the model is presented by playing the sound of the frequency f ref i in order for the time t, and after the time t ⁇ i from the end time of the presentation of the timing evaluation of the model. Make the sound of frequency f i play.
  • a "model” is a time series of events in a reference motion sequence.
  • “Example 1” is a case where the left wrist is delayed and the shoulder brake is early, and there is no order error. The second note is delayed and the sixth note is earlier.
  • “Example 2” is a case where the order of the wrist and the head is reversed, and the event intervals are set to be the same. The 4th note is a semitone lower and the 5th note is a semitone higher, and the order is reversed (all intervals remain 0.2 seconds).
  • the characteristic body parts and their movements in the movement are confirmed, and the evaluation items for evaluating the golf skill are selected.
  • Ten items of posture evaluation and six items of timing evaluation of the turning point in the movement of a specific part were adopted.
  • the evaluation results for improving golf skills were presented in comparison with the model using auditory information.
  • the difference from the ideal kinematic sequence is calculated based on the analysis result (time-series information of the whole body skeletal movement of all movements), and it is presented as acoustic information immediately after the movement.
  • the present invention calculates the difference between its own motion sequence and the ideal motion sequence based on human motion measurement and motion analysis in real time or near real time, converts it into acoustic information, and immediately conveys it. It is intended to intuitively grasp the points to be corrected in the movement of the learner and enhance the training effect of the learner, and can be used for exercise training, golf training, dance training, rehabilitation, exercise support, and the like.

Abstract

The problem is how to calculate and compare motion sequences, and communicate the motion sequences to a learner with respect to a physical activity in which the motion sequences are considered important in particular. A training assisting method comprises: preparing a first motion sequence of a physical activity as a reference; acquiring a second motion sequence from activity data of the physical activity of a learner; and presenting a difference between the first motion sequence and the second motion sequence to the learner as acoustic information.

Description

トレーニング支援方法及び装置Training support methods and equipment
本発明は、トレーニング支援方法及び装置に関するものである。 The present invention relates to training support methods and devices.
スポーツの指導においては、身体運動について計測や撮影を行い、その結果に基づいて指導者が改善点を提案するという手法が多く行われている。ゴルフの指導を例にとると、弾道測定器によって打球の測定を行い、その結果とビデオカメラで撮影したスイング時の姿勢に基づいて指導者が改善点を提案する手法がよく採られる。 In sports instruction, there are many methods in which physical exercise is measured and photographed, and the instructor proposes improvement points based on the results. Taking golf instruction as an example, a method is often adopted in which a ballistic measuring instrument is used to measure a hit ball, and the instructor proposes points for improvement based on the result and the posture during a swing taken by a video camera.
しかし、提案内容の多くは、指導者の経験や勘に依存している場合も多く、指導者と学習者との感覚の差により、学習者に改善点がうまく伝わらないこともあり、さらには、指導者の意図に反した伝わり方をしてしまう可能性もある。 However, many of the proposals often depend on the experience and intuition of the instructor, and due to the difference in feelings between the instructor and the learner, the improvement points may not be well communicated to the learner, and even more. , There is a possibility that the communication will be contrary to the intention of the leader.
ゴルフは動作のシーケンスが確立されているスポーツの1つであり、ゴルフスイングにおける身体運動の研究が行われている。穂苅ら(非特許文献1)は、スイングにおける関節の位置や角度の測定を行うとともに、ゴルフ教則本(非特許文献2)において表現された規範及び技能を定量化することを試みた。また、Jorgensen(非特許文献3)は、ゴルフスイング時の腕とゴルフクラブを2重振子としてモデル化し、井上ら(非特許文献4)は腕、手首の関節トルクとスイングとの関係の調査を目的として、2重振子モデルのモード解析を行った。Tinmarkら(非特許文献5)は、手、肩、腰の3箇所に着目し、各部位の回転速度が最大になる順番とヘッドの速度との関係を調査した。Venaら(非特許文献6)は、左腕、肩、腰の3箇所の回転速度が最大になるタイミングと技能レベルについて関係性を調査した。 Golf is one of the sports for which a sequence of movements has been established, and research on physical exercise in golf swings is being conducted. Hogari et al. (Non-Patent Document 1) attempted to measure the position and angle of joints in a swing and to quantify the norms and skills expressed in the golf instruction book (Non-Patent Document 2). In addition, Jorgensen (Non-Patent Document 3) modeled the arm and golf club during a golf swing as a double pendulum, and Inoue et al. (Non-Patent Document 4) investigated the relationship between the joint torque of the arm and wrist and the swing. The mode analysis of the double pendulum model was performed for the purpose of. Tinmark et al. (Non-Patent Document 5) focused on three parts, the hand, shoulder, and waist, and investigated the relationship between the order in which the rotation speed of each part was maximized and the speed of the head. Vena et al. (Non-Patent Document 6) investigated the relationship between the timing at which the rotational speeds of the left arm, shoulder, and hip are maximized and the skill level.
ゴルフスイングにおいて、複数の体の部位における動作の順序やタイミングの差はKinematic Sequence(キネマティック・シーケンス)と呼ばれ、効率よくヘッドスピードを増加させるために重要な要素として認識されている(非特許文献7)。図20に示すように、スイング中の各部位の回転速度の最大値が、腰(骨盤)、胸(上半身)、腕、ゴルフクラブの順で生起する。すなわち、ダウンスイングにおいては、下半身(脚)、腰、肩、腕、クラブの順に動作が連鎖することを示している。 In a golf swing, the difference in the order and timing of movements in multiple body parts is called Kinematic Sequence, and is recognized as an important factor for efficiently increasing head speed (non-patented). Document 7). As shown in FIG. 20, the maximum value of the rotational speed of each part during the swing occurs in the order of the waist (pelvis), chest (upper body), arms, and golf club. That is, in the downswing, it is shown that the movements are chained in the order of the lower body (legs), the waist, the shoulders, the arms, and the club.
ゴルフスイングは短時間で複数の身体部位が相互作用しながら完了する動作であり、スイング動画を再生しても、身体部位の動作の連鎖を視覚的に把握することは難しい。また、各身体部位の動作のタイミングが望ましい形で行われているかを認識でき、自覚して修正できるのはすでに上級者である。これらのことを踏まえ、Kinematic Sequence(キネマティック・シーケンス)を主とした動作シーケンスの評価を学習者にわかりやすく伝えることが重要である。 A golf swing is a movement that is completed while a plurality of body parts interact with each other in a short time, and it is difficult to visually grasp the chain of movements of the body parts even if the swing movie is played. In addition, it is already advanced people who can recognize whether the timing of movement of each body part is performed in a desirable manner and can consciously correct it. Based on these facts, it is important to convey to the learner the evaluation of the motion sequence, mainly the Kinematic Sequence, in an easy-to-understand manner.
従来からゴルフの練習に音を用いることは知られており、例えば、特許文献1~3に記載されている。これらの文献に記載された手法は、学習者が音に合わせてスイングを行うものであり、また、学習者のスイングについて評価するものではない。 It has been conventionally known to use sound for golf practice, and it is described in, for example, Patent Documents 1 to 3. The methods described in these documents allow the learner to swing in time with the sound, and do not evaluate the learner's swing.
特開2001-299976JP 2001-299966 特開2013-39357JP 2013-39357 米国特許第5743807号U.S. Pat. No. 5,743,807
本発明者らは、特に動作のシーケンスが重要視される身体運動において、これをどのように計算して比較し、学習者に伝えるかという課題に取り組んだ。 The present inventors have tackled the problem of how to calculate, compare, and convey this to the learner, especially in physical exercise in which the sequence of movement is important.
 本発明が採用した技術手段は、
 身体運動の第1動作シーケンスを参照として用意し、
 学習者の身体運動の運動データから第2動作シーケンスを取得し、
 第1動作シーケンスと第2動作シーケンスの差(違い)を音響情報として学習者に提示する、
 トレーニング支援方法、である。
 1つの態様では、第1動作シーケンスと第2動作シーケンスの差を取得することを含み、
 取得した差を音響情報として学習者に提示する。
The technical means adopted by the present invention is
Prepare the first movement sequence of physical exercise as a reference,
Obtain the second motion sequence from the exercise data of the learner's physical exercise,
Presenting the difference (difference) between the first motion sequence and the second motion sequence to the learner as acoustic information.
Training support method.
One embodiment comprises obtaining the difference between the first and second motion sequences.
The acquired difference is presented to the learner as acoustic information.
 1つの態様では、動作シーケンスは、異なる時刻で発生する複数のイベントによって規定される。
 1つの態様では、第1動作シーケンスと第2動作シーケンスの差を取得することを含み、
 前記差には、イベントの発生順序、および/あるいは、イベントの発生タイミングの差が含まれる。
In one aspect, the motion sequence is defined by a plurality of events that occur at different times.
One embodiment comprises obtaining the difference between the first and second motion sequences.
The difference includes the difference in the order in which the events occur and / or the timing in which the events occur.
 1つの態様では、前記各イベントには、音が割り当てられており、
 前記第1動作シーケンス、前記第2動作シーケンスを、音響情報(音の時系列情報)に変換可能である。
In one aspect, each of the events is assigned a sound.
The first operation sequence and the second operation sequence can be converted into acoustic information (sound time series information).
 1つの態様では、第1動作シーケンスと第2動作シーケンスの差を音響情報で学習者に提示することは、
 第1動作シーケンスに対応する第1音響情報を提示すること、
 第2動作シーケンスに対応する第2音響情報を提示すること、を含み、
 第1音響情報と第2音響情報を学習者が対比することで差を認識させることを含む。
 1つの態様では、第1動作シーケンスのイベントの順序と第2動作シーケンスのイベントの順序が異なる場合に、前記第2音響情報を提示する時に、順序が誤ったイベントの音を変化させる。
In one embodiment, presenting the learner with acoustic information the difference between the first and second motion sequences
Presenting the first acoustic information corresponding to the first motion sequence,
Including presenting the second acoustic information corresponding to the second motion sequence,
This includes making the learner recognize the difference by comparing the first acoustic information and the second acoustic information.
In one aspect, when the order of the events in the first action sequence and the order of the events in the second action sequence are different, the sound of the event in the wrong order is changed when the second acoustic information is presented.
 1つの態様では、前記差を音響情報に変換する手段を備えており、音響情報に変換された差が学習者に提示される。 In one aspect, the learner is provided with a means for converting the difference into acoustic information, and the difference converted into acoustic information is presented to the learner.
 1つの態様では、第1動作シーケンスと第2動作シーケンスの差を視覚的に認識可能に学習者に提示することを含む。
 例えば、第1動作シーケンスと第2動作シーケンスを対比した図(例えば、差を強調して示す)を表示してもよい。
 あるいは、第1動作シーケンスと第2動作シーケンスの差の分析結果を図示してもよい。
 差が大きかったイベントに対応するフレーム画像を表示してもよい。
 また、音のシーケンスの提示と同時にスイング画像を提示してもよい。
In one embodiment, the difference between the first motion sequence and the second motion sequence is presented to the learner in a visually recognizable manner.
For example, a diagram comparing the first operation sequence and the second operation sequence (for example, the difference may be emphasized) may be displayed.
Alternatively, the analysis result of the difference between the first operation sequence and the second operation sequence may be illustrated.
The frame image corresponding to the event with a large difference may be displayed.
Further, the swing image may be presented at the same time as the presentation of the sound sequence.
 本発明が採用したトレーニング支援装置は、
 身体運動の第1動作シーケンスを参照として記憶する手段、
 学習者の身体運動の運動データから第2動作シーケンスを取得する手段と、
 第1動作シーケンスと第2動作シーケンスの差(違い)を音響情報として学習者に提示する手段と、を備えている。
 1つの態様では、第1動作シーケンスと第2動作シーケンスの差を取得する手段を含み、
 前記提示手段は、取得した差を音響情報として学習者に提示する。
The training support device adopted by the present invention is
A means of storing the first motion sequence of physical exercise as a reference,
A means of acquiring a second motion sequence from the learner's physical motion data,
It is provided with a means for presenting the difference (difference) between the first operation sequence and the second operation sequence to the learner as acoustic information.
In one embodiment, the means for obtaining the difference between the first operation sequence and the second operation sequence is included.
The presenting means presents the acquired difference to the learner as acoustic information.
 1つの態様では、動作シーケンスは、異なる時刻で発生する複数のイベントによって規定され、
 イベント発生時刻取得手段を備えている。
 1つの態様では、前記差には、第1動作シーケンスと第2動作シーケンスのイベントの発生順序、および/あるいは、イベントの発生タイミングの差が含まれ、
 イベントの発生順序、および/あるいは、イベントの発生タイミングの差を取得する手段を含む。
In one aspect, the sequence of actions is defined by multiple events that occur at different times.
It is equipped with a means for acquiring the event occurrence time.
In one embodiment, the difference includes a difference in the order in which the events of the first operation sequence and the second operation sequence occur, and / or the timing of occurrence of the events.
Includes means for obtaining the event occurrence order and / or the difference in event occurrence timing.
 1つの態様では、前記各イベントには、音が割り当てられており、
 前記第1動作シーケンス、前記第2動作シーケンスを、音響情報(音の時系列情報)に変換する手段を備えている。
In one aspect, each of the events is assigned a sound.
A means for converting the first operation sequence and the second operation sequence into acoustic information (time-series information of sound) is provided.
 1つの態様では、前記差を音響情報に変換する手段を備えており、音響情報に変換された差が学習者に提示される。 In one aspect, the learner is provided with a means for converting the difference into acoustic information, and the difference converted into acoustic information is presented to the learner.
 本発明は、コンピュータを、上記各手段として機能させるコンピュータプログラムとしても提供され得る。 The present invention can also be provided as a computer program that causes a computer to function as each of the above means.
 1つの態様では、身体運動は、ゴルフスイングである。 In one aspect, the physical exercise is a golf swing.
 本発明は、学習者の運動を定量的に評価する手法であり、学習者の自己トレーニングの支援手段、及び、トレーニング指導の支援手段として用いることができる。
 本発明では、参照となる動作シーケンスと、運動データから取得した学習者の動作シーケンスとの差を、音響情報として学習者にリアルタイムないしニア・リアルタイムで提示することで、学習者が自らの動作を自覚し修正を行うことができる直感的な方法を提供する。
 本発明によって、自己の身体運動(例えば、ゴルフスイング)の正確性や妥当性をリアルタイムないしニア・リアルタイムで学習者にフィードバックすることができる。
 本発明は、1つの例では、ゴルフスイングのトレーニングの支援に用い得るが、ゴルフスイングに限定されるものではなく、その他の運動トレーニング、ダンストレーニング、リハビリテーションなどにおいてトレーニング効果を高めることができる。
The present invention is a method for quantitatively evaluating a learner's movement, and can be used as a learner's self-training support means and a training guidance support means.
In the present invention, the learner can perform his / her own movement by presenting the difference between the reference movement sequence and the learner's movement sequence acquired from the movement data to the learner as acoustic information in real time or near real time. It provides an intuitive way to be aware and make corrections.
According to the present invention, the accuracy and validity of one's own physical exercise (for example, golf swing) can be fed back to the learner in real time or near real time.
In one example, the present invention can be used to support training of a golf swing, but the present invention is not limited to the golf swing, and the training effect can be enhanced in other exercise training, dance training, rehabilitation and the like.
本実施形態に係るトレーニング支援システムの全体図である。It is an overall view of the training support system which concerns on this embodiment. 本実施形態に係るトレーニング支援方法の概要図である。It is a schematic diagram of the training support method which concerns on this embodiment. 運動データを音の時系列情報として出力するステップを示す図である。It is a figure which shows the step which outputs the motion data as the time series information of a sound. 学習者に、自己の動作と参照となる動作を音響情報として比較させるステップを示す図である。It is a figure which shows the step which a learner is made to compare a self-movement and a reference movement as acoustic information. 学習者に、自己の動作と参照となる動作の差を音響情報として提示するステップを示す図である。It is a figure which shows the step which presents the difference between a self-movement and a reference movement to a learner as acoustic information. 身体の特徴点の部位及び番号(1~20)を示す図である。It is a figure which shows the part and the number (1 to 20) of a feature point of a body. 各身体部位の中心(a~j)を示す図である。It is a figure which shows the center (a to j) of each body part. 身体の軸(k,l)を示す図である。It is a figure which shows the axis (k, l) of a body. 肩の線分、腰の線分、両腕、ゴルフクラブを示す図である。It is a figure which shows the shoulder line segment, the waist line segment, both arms, and a golf club. 左大腿、左下肢の定義を示す。The definitions of the left thigh and left lower limb are shown. 左前腕ベクトル及びクラブシャフトベクトルの定義を示す図である。It is a figure which shows the definition of the left forearm vector and the club shaft vector. ゴルフスイングの一連の動作を示す図である。It is a figure which shows a series of movements of a golf swing. 左図は第1イベント(左膝伸展開始:左膝の角度が大きくなり始める時刻)、右図は第2イベント(左手首固定解除:左手首を返す時刻)を示す図である。The figure on the left shows the first event (start of left knee extension: time when the angle of the left knee starts to increase), and the figure on the right shows the second event (release of left wrist fixation: time when the left wrist is returned). 左図は第3イベント(腰のブレーキ:腰の回転角速度が最大になる時刻)、右図は第4イベント(手の速さ最大:手首の中心位置の速さ最大の時刻)を示す図である。The figure on the left shows the third event (waist brake: time when the rotation angular velocity of the waist reaches the maximum), and the figure on the right shows the fourth event (maximum hand speed: time when the center position of the wrist reaches the maximum speed). is there. 左図は第5イベント(ヘッドの速さ最大:クラブヘッドの速さ最大の時刻)、右図は第6イベント(肩のブレーキ:肩の回転角速度が最大となる時刻)を示す図である。The figure on the left shows the fifth event (maximum head speed: time when the maximum speed of the club head), and the figure on the right shows the sixth event (shoulder brake: time when the rotation angular velocity of the shoulder becomes maximum). イベントの時系列の音の時系列への変換を示し、音声間隔は等間隔となっている。It shows the conversion of the time series of events into a time series of sounds, and the audio intervals are evenly spaced. 音響情報に変換された動作シーケンスを示し、手本は参照となる動作データ、例1は一部のイベントの発生タイミングのみがずれた場合(順序誤差なし)、例2は一部のイベントの順番が入れ替わった場合に対応している。Shows the operation sequence converted into acoustic information, the model is the reference operation data, Example 1 is when only the occurrence timing of some events is shifted (no order error), Example 2 is the order of some events It corresponds to the case where is replaced. 順序のズレから得られるズレ情報を例示する図であり、間違えたイベントの位置、各イベントの位相差、最大位相差のイベントが得られている。It is a figure exemplifying the deviation information obtained from the deviation of the order, and the position of the wrong event, the phase difference of each event, and the event of the maximum phase difference are obtained. 最大位相差のイベントに着目した音響情報の提示を示す図である。It is a figure which shows the presentation of acoustic information paying attention to the event of the maximum phase difference. ゴルフスイングにおけるKinematic Sequence(キネマティック・シーケンス)を示す図である(非特許文献7参照)。It is a figure which shows the Kinematic Sequence (Kinematic sequence) in a golf swing (see Non-Patent Document 7).
[A]トレーニング支援システムの概要
[A-1]システムのハードウェア構成
図1、図2を参照しつつ、本実施形態に係るトレーニング支援システムについて説明する。本実施形態に係るトレーニング支援システムは動画取得部と、各種データを記憶する記憶部と、各種計算を実行する処理部と、画像出力部(ディスプレイ)と、音声出力部と、を備えている。典型的には、動画取得部はビデオカメラから構成され、記憶部、処理部、画像出力部、音声出力部は、1つあるいは複数のコンピュータ(入力部、CPUからなる演算部、RAM、ROMを備えた記憶部、出力部を備えている)から構成される。
[A] Outline of the training support system [A-1] Hardware configuration of the system The training support system according to the present embodiment will be described with reference to FIGS. 1 and 2. The training support system according to the present embodiment includes a moving image acquisition unit, a storage unit for storing various data, a processing unit for executing various calculations, an image output unit (display), and an audio output unit. Typically, the video acquisition unit is composed of a video camera, and the storage unit, processing unit, image output unit, and audio output unit include one or more computers (input unit, arithmetic unit consisting of CPU, RAM, ROM). It is composed of a storage unit and an output unit).
[A-2]モーションキャプチャ
身体運動の時系列データを取得する手段は、いわゆるモーションキャプチャであって、モーションキャプチャは、1つの態様では、対象の動作の動画データを取得する動作取得部(ビデオカメラ)と、動画データに基づいて対象の動作を表す時系列データを取得する動作取得部から構成される。カメラによって取得された動画データやモーションキャプチャによって取得された当該対象の姿勢の時系列データは記憶部に記憶される。
[A-2] Motion capture The means for acquiring time-series data of physical exercise is so-called motion capture, and in one aspect, motion capture is a motion acquisition unit (video camera) that acquires moving image data of a target motion. ) And an action acquisition unit that acquires time-series data representing the target action based on the moving image data. The moving image data acquired by the camera and the time-series data of the posture of the target acquired by motion capture are stored in the storage unit.
対象の動作は、モーションキャプチャによって取得された当該対象の姿勢の時系列データによって規定される。対象の姿勢は、対象の身体上の複数の特徴点(典型的には関節)によって特定され、各フレームにおいて、複数の特徴点の3次元座標値を取得することで、複数の特徴点の3次元座標値の時系列データから対象の動作を規定する。本発明に用いられるモーションキャプチャの種類は限定されず、特徴点を特定する光学式マーカを用いた光学式モーションキャプチャ、加速度センサやジャイロスコープ、地磁気センサなどのいわゆる慣性センサを対象の身体に装着して、対象のモーションデータを取得する方式、光学式マーカやセンサを装着しない、いわゆるマーカレスモーションキャプチャ等を例示することができる。本実施形態では、光学式マーカを用いた光学式モーションキャプチャによって動作計測を行ったが、対象の自然な動作を妨げないという観点からは、マーカレスモーションキャプチャが有利である。マーカレスモーションキャプチャとしては、カメラと深度センサを備えたシステム(Kinectに代表される)を用いたモーションキャプチャ、あるいは、深層学習を用い、1視点あるいは複数視点からのRGB画像を解析してモーションデータを取得するモーションキャプチャを例示することができる。 The movement of the target is defined by the time series data of the posture of the target acquired by motion capture. The posture of the target is specified by a plurality of feature points (typically joints) on the target body, and by acquiring three-dimensional coordinate values of the plurality of feature points in each frame, three of the plurality of feature points are obtained. The target operation is defined from the time-series data of the dimensional coordinate values. The type of motion capture used in the present invention is not limited, and an optical motion capture using an optical marker that identifies a feature point, a so-called inertial sensor such as an acceleration sensor, a gyroscope, or a geomagnetic sensor is attached to the target body. Therefore, a method of acquiring the target motion data, a so-called markerless motion capture in which an optical marker or a sensor is not attached, and the like can be exemplified. In the present embodiment, motion measurement is performed by optical motion capture using an optical marker, but markerless motion capture is advantageous from the viewpoint of not interfering with the natural motion of the target. As markerless motion capture, motion capture using a system equipped with a camera and a depth sensor (represented by Kinect) or deep learning is used to analyze RGB images from one or more viewpoints and motion data. A motion capture to acquire can be exemplified.
[A-3]身体運動の動作シーケンス
身体運動の動作シーケンスは、所定の身体部位、または連携した複数の身体部位が、他の身体部位との相対的な位置関係の変化として現す全身運動におけるイベント群が、順序関係および時間関係(タイミング)を持つものとして説明される。すなわち、動作シーケンスは、異なる時刻で発生する複数のイベントによって規定される。イベントは、1つ以上の身体部位の速度(最大速度、最大角速度、最小角速度等)や位置(角度等)等によって規定される。動作シーケンス取得手段は、各イベントの発生時刻を取得する手段を備えている。イベント発生時刻取得手段によって、学習者の動作シーケンスを規定する各イベントの発生時刻が取得され、各イベントの発生時刻及び発生時刻から得られるイベントの順序が記憶部に記憶される。
[A-3] Motion sequence of physical exercise The motion sequence of physical exercise is an event in whole body exercise in which a predetermined body part or a plurality of linked body parts express as a change in the relative positional relationship with other body parts. The group is described as having an order relationship and a time relationship (timing). That is, the operation sequence is defined by a plurality of events that occur at different times. The event is defined by the velocity (maximum velocity, maximum angular velocity, minimum angular velocity, etc.), position (angle, etc.) of one or more body parts. The operation sequence acquisition means includes means for acquiring the occurrence time of each event. The event occurrence time acquisition means acquires the occurrence time of each event that defines the learner's operation sequence, and stores the occurrence time of each event and the order of the events obtained from the occurrence time in the storage unit.
例えば、5つのイベントE、E、E、E、Eがこの順序で生起する身体運動において、参照となる動作シーケンスは、E、E、E、E、Eの生起順序及び発生タイミングで規定することができる。各イベントE、E、E、E、Eの発生時刻を取得することで、発生時刻からイベントの順序がわかる。また、各イベントの発生時刻から、各イベントの発生タイミング(相対的なタイミング)を取得することができる。同じ身体運動(例えば、ゴルフスイング)であっても、開始から終了までの時間には個人差があるので、身体運動におけるイベントの発生間隔の時間は相対的なものである。よって、身体運動の開始時点、終了時点を取得し、身体運動の全体の時間を共通化させて発生タイミングを相対化することで、各イベントの発生タイミングを比較することができる。 For example, in a physical exercise in which five events E 1 , E 2 , E 3 , E 4 , and E 5 occur in this order, the reference motion sequences are E 1 , E 2 , E 3 , E 4 , E 5 and so on. It can be specified by the order of occurrence and the timing of occurrence. By acquiring the occurrence time of each event E 1 , E 2 , E 3 , E 4 , and E 5 , the order of the events can be known from the occurrence time. In addition, the occurrence timing (relative timing) of each event can be acquired from the occurrence time of each event. Even for the same physical exercise (for example, golf swing), the time between the occurrence of events in the physical exercise is relative because the time from the start to the end varies from person to person. Therefore, it is possible to compare the occurrence timings of each event by acquiring the start time point and the end time point of the physical exercise, sharing the entire time of the physical exercise, and relativizing the occurrence timing.
[A-4]イベント発生時刻取得手段
イベント発生時刻取得手段において、どのような情報を用いてどのようにイベントの発生時刻を検出するかはイベントによって異なり得る。イベント発生時刻取得手段は、対象の動作の時系列データを用い、イベントを規定する1つあるいは複数の部位の所定の角度、角速度、角加速度、速度、回転速度、加速度、位置(他の部位に対する相対的位置、身体以外の物に対する相対的位置等)、姿勢、重心等を条件として用いて所定のイベントの発生時刻を取得する。
[A-4] Event Occurrence Time Acquisition Means In the event occurrence time acquisition means, what kind of information is used and how the event occurrence time is detected may differ depending on the event. The event occurrence time acquisition means uses time-series data of the target motion, and uses a predetermined angle, angular velocity, angular acceleration, velocity, rotation speed, acceleration, and position (relative to other parts) of one or more parts that define the event. Acquire the occurrence time of a predetermined event using the relative position, relative position with respect to an object other than the body, etc.), posture, center of gravity, etc. as conditions.
[A-5]音源データ
記憶部には音源データが格納されている。音源データの用い方としては、動作シーケンス(参照となる動作シーケンス)を規定する各イベントに音が割り当てられる場合、参照となる動作シーケンスと対象者の動作シーケンスの差に対して音が割り当てられる場合がある。前者では、例えば、各イベントに、音源データとして格納された音が予め割り当てられており、動作シーケンスの音響情報変換手段によって、動作シーケンスを音の時系列情報に変換して、音声出力部から出力する。各イベントに割り当てられた各音は、互いに聴別できる音、また、音のシーケンスを識別しやすい音であることが望ましい。また、参照となる動作シーケンスに対応する音のシーケンス(音の時系列情報)として、馴染みのある既知の音源を用いてもよい。後者においては、各イベントに割り当てられた音とは独立して(各イベントに音が割り当てられていない場合を含む)、差に対して音を割り当てる場合、あるいは、各イベントに割り当てられた音を用いて、その音を変化させて提示する(例えば、音高を半音上げたり下げたりして提示する)場合が例示される。
[A-5] Sound source data is stored in the sound source data storage unit. As for how to use the sound source data, when sound is assigned to each event that defines the operation sequence (reference operation sequence), or when sound is assigned to the difference between the reference operation sequence and the target person's operation sequence. There is. In the former, for example, a sound stored as sound source data is assigned to each event in advance, and the operation sequence is converted into time-series information of the sound by the acoustic information conversion means of the operation sequence and output from the audio output unit. To do. It is desirable that each sound assigned to each event is a sound that can be distinguished from each other and that the sequence of sounds can be easily identified. Further, a familiar and known sound source may be used as a sound sequence (sound time series information) corresponding to the reference operation sequence. In the latter, the sound assigned to each event is independent of the sound assigned to each event (including the case where no sound is assigned to each event), and the sound is assigned to the difference, or the sound assigned to each event is assigned. An example is the case where the sound is presented in a changed manner (for example, the pitch is raised or lowered by a semitone).
[A-6]参照となる動作シーケンス
記憶部には、参照となる動作シーケンスを規定する各イベントの発生時刻及び順序が記憶されている。参照となる動作シーケンスは、手本・模範となる理想的な運動データから取得する。動作のシーケンスが確立されている身体運動について、手本となる対象の身体運動をビデオ撮影して運動データを取得し、手本となる運動データの時系列データを取得する。取得した時系列データを用いて身体運動の動作シーケンスを規定する複数のイベントの発生時刻を取得し、各イベントの発生時刻と発生順序が、参照となる動作シーケンスとして記憶部に記憶される。身体運動の動画や姿勢の時系列データも記憶部に記憶される。
[A-6] The reference operation sequence storage unit stores the time and order of occurrence of each event that defines the reference operation sequence. The reference motion sequence is obtained from ideal motion data that serves as a model / model. For the physical exercise for which the motion sequence has been established, the physical exercise of the target to be a model is videotaped to acquire the exercise data, and the time-series data of the exercise data to be the model is acquired. Using the acquired time-series data, the occurrence times of a plurality of events that define the movement sequence of physical exercise are acquired, and the occurrence time and the occurrence order of each event are stored in the storage unit as a reference movement sequence. Video of physical exercise and time-series data of posture are also stored in the storage unit.
複数の異なる参照となる動作データ(動作シーケンスを規定するイベントセット及びイベントの発生時刻及び順序)を用意して、複数の候補から参照として用いる動作シーケンスを選択するようにしてもよい。手本となる身体運動や理想的な身体運動は必ずしも1つとは限らないので、複数の参照動作シーケンスを取得して、記憶部に格納しておいてもよい。また、ある動作シーケンスを規定するイベントのセットは必ずしも一通りではなく、参照となる同じ身体運動データから複数種類の動作シーケンス(少なくとも一部のイベントが異なる)を取得して、記憶部に格納しておいてもよい。 It is also possible to prepare a plurality of different reference operation data (event set that defines the operation sequence and the occurrence time and order of events), and select an operation sequence to be used as a reference from a plurality of candidates. Since the model physical exercise and the ideal physical exercise are not necessarily one, a plurality of reference motion sequences may be acquired and stored in the storage unit. In addition, the set of events that define a certain motion sequence is not always one set, and multiple types of motion sequences (at least some events are different) are acquired from the same physical exercise data as a reference and stored in the storage unit. You may keep it.
[A-7]動作シーケンスの音楽情報への変換
動作シーケンスを音響情報に変換するステップを図3に例示する。先ず、対象が身体運動を行った時の動作の動画を取得し、動画データから対象の動作の時系列データを取得する。対象の動作の時系列データに基づいて、身体運動の動作シーケンスを規定する各イベントの発生時刻及び発生順序を取得する。各イベントに音を割り当てることで、動作シーケンスが音響情報(音の時系列情報)に変換される。対象が身体運動を行った直後に動作シーケンスを音響情報に変換して、音響情報を対象に提示することができる。
[A-7] Conversion of operation sequence to music information The step of converting an operation sequence into acoustic information is illustrated in FIG. First, a moving image of the movement when the target performs physical exercise is acquired, and time-series data of the movement of the target is acquired from the moving image data. Based on the time-series data of the target motion, the occurrence time and the occurrence order of each event that defines the motion sequence of the physical exercise are acquired. By assigning a sound to each event, the operation sequence is converted into acoustic information (sound time series information). Immediately after the subject performs physical exercise, the motion sequence can be converted into acoustic information and the acoustic information can be presented to the subject.
音のシーケンス(音の時系列情報)における音の間隔は、実際のイベントと同じ間隔を用いてもよく、あるいは、相対的な間隔を維持して時間を長くしてもよい。音のシーケンスを提示する場合に、実際のイベントのシーケンスと同じタイミングで音を提示することが聴き分けにくい場合には、イベントと同じタイミングで音を提示する必要はない。すなわち、例えばゴルフスイングにおいて、スイングのシーケンスを規定する各イベントの時間幅と提示される音のシーケンスの時間幅は異なってもよい(音に合わせてスイングするのではない)。あるいは、手本となる運動データから得られたイベントの発生間隔を等間隔に変換し、手本を音声表現する時には、所定の音を所定の順序で等間隔で提示してもよい。  As the sound interval in the sound sequence (sound time series information), the same interval as the actual event may be used, or the relative interval may be maintained and the time may be lengthened. When presenting a sound sequence, if it is difficult to distinguish the sound at the same timing as the actual event sequence, it is not necessary to present the sound at the same timing as the event. That is, for example, in a golf swing, the time width of each event that defines the swing sequence and the time width of the presented sound sequence may be different (not swinging according to the sound). Alternatively, when the event occurrence intervals obtained from the exercise data as a model are converted into equal intervals and the model is expressed by voice, predetermined sounds may be presented in a predetermined order at equal intervals.
このような音響情報を用いたトレーニングの支援としては参照となる動作シーケンスに対応する音響情報と、自己の動作シーケンスに対応する音響情報を学習者が聴き比べることを例示することができる。図4に示すように、先ず、参照となる動作シーケンス(例えば、ゴルフスイング)の音響情報を学習者に提示する。次いで、学習者の身体運動(例えば、ゴルフスイング)を行い、学習者の動作から動作シーケンスを取得し、取得した動作シーケンスを音響情報に変換して学習者に提示する。なお、学習者の動作シーケンスに対応する音響情報を提示する際に、参照となる動作シーケンスと順序が異なるイベントに対応する音の音高を変化させてもよい(図17「例2」)。 As support for training using such acoustic information, it can be exemplified that the learner listens and compares the acoustic information corresponding to the reference motion sequence and the acoustic information corresponding to his / her own motion sequence. As shown in FIG. 4, first, the learner is presented with acoustic information of a reference motion sequence (for example, a golf swing). Next, the learner's physical exercise (for example, golf swing) is performed, a motion sequence is acquired from the learner's motion, and the acquired motion sequence is converted into acoustic information and presented to the learner. When presenting the acoustic information corresponding to the learner's motion sequence, the pitch of the sound corresponding to the event having a different order from the reference motion sequence may be changed (FIG. 17 “Example 2”).
この例では、参照となる動作に対応する音響情報と自己の動作に対応する音響情報を学習者に対比させることで差を認識させる。参照となる動作に対応する音響情報、自己の動作に対応する音響情報を学習者に提示するタイミングや回数は限定されない。例えば、スイングの都度、学習者の動作シーケンスを変換した音響情報を参照となる音響情報と対比することで、聴こえ方の差異の程度によって、どのスイングが参照となるスイングに近いか遠いかを直感的に判断することができる。 In this example, the learner is made to recognize the difference by comparing the acoustic information corresponding to the reference motion and the acoustic information corresponding to the own motion to the learner. The timing and number of times that the acoustic information corresponding to the reference motion and the acoustic information corresponding to the own motion are presented to the learner are not limited. For example, by comparing the acoustic information obtained by converting the learner's motion sequence with the reference acoustic information for each swing, it is intuitive to know which swing is closer to or far from the reference swing depending on the degree of difference in hearing. Can be judged.
[A-8]動作シーケンスの比較
動作シーケンス比較手段によって、学習者の動作シーケンスを規定する各イベントの発生時刻及び順序と、参照となる動作シーケンスを規定する各イベントの発生時刻及び順序が比較され、動作シーケンスの差(違い)が取得され、動作シーケンスの差の情報が記憶部に記憶される。動作シーケンスの差には、限定されないものの、イベントの発生順序の違い、イベントの発生タイミングの差が含まれる。発生タイミング(発生間隔)の差異は、例えば、身体運動の全体時間に対する各発生間隔の割合を対比することで得ることができる。身体運動によっては、イベントの発生順序とイベントの発生タイミングの両方を比較した方が望ましい場合があり、イベントの発生順序のみを比較すればよい場合もあり、あるいは、イベントの発生タイミングのみを比較すればよい場合もあり得る。なお、参照となる動作と学習者の動作について、動作シーケンス以外の事項を比較してもよい。参照となる動作を規定する姿勢の時系列データ、学習者の動作を規定する姿勢の時系列データを利用して様々な事項を比較することができる。例えば、参照動作と学習者の動作の速度(例えば、ゴルフスイングにおけるヘッドスピード)を比較してもよい。
[A-8] Comparison of Action Sequences The action sequence comparison means compares the occurrence time and order of each event that defines the learner's action sequence with the occurrence time and order of each event that defines the reference action sequence. , The difference (difference) in the operation sequence is acquired, and the information on the difference in the operation sequence is stored in the storage unit. The difference in the operation sequence includes, but is not limited to, the difference in the event occurrence order and the difference in the event occurrence timing. The difference in the occurrence timing (occurrence interval) can be obtained, for example, by comparing the ratio of each occurrence interval to the total time of physical exercise. Depending on the physical exercise, it may be desirable to compare both the event occurrence order and the event occurrence timing, and in some cases it may be necessary to compare only the event occurrence order, or only the event occurrence timing. In some cases it may be good. It should be noted that items other than the motion sequence may be compared between the reference motion and the learner's motion. Various items can be compared by using the time-series data of the posture that defines the movement as a reference and the time-series data of the posture that defines the movement of the learner. For example, the speeds of reference movements and learner movements (eg, head speed in a golf swing) may be compared.
図5には、参照となる動作シーケンスと学習者の動作シーケンスを比較して差を取得し、差を出力情報(例えば、音響情報)に変換し、出力情報(例えば、音響情報)を提示するステップを示している。先ず、学習者の動作の画像を取得し、画像データから学習者の動作の時系列データを取得する。学習者の動作の時系列データを用いて、学習者の動作シーケンスを規定する各イベントの発生時刻と発生順序を取得する。学習者の動作シーケンスを規定する各イベントの発生時刻及び発生順序と、参照の動作シーケンスを規定する各イベントの発生時刻及び発生順序を比較することで、動作シーケンスの差を取得する。動作シーケンスの差は、例えば、イベントの発生順序における差、イベントの発生タイミングにおける差である。発生順序の差の情報(順序のズレから得られたズレ情報)には、限定されないものの、間違えたイベントの位置、各イベントの位相差、最大位相差のイベントが含まれる(図18参照)。 In FIG. 5, the reference motion sequence and the learner's motion sequence are compared to obtain the difference, the difference is converted into output information (for example, acoustic information), and the output information (for example, acoustic information) is presented. Shows the steps. First, an image of the learner's movement is acquired, and time-series data of the learner's movement is acquired from the image data. Using the time-series data of the learner's movements, the occurrence time and the occurrence order of each event that defines the learner's movement sequence are acquired. By comparing the occurrence time and order of each event that defines the learner's action sequence with the occurrence time and order of each event that defines the reference action sequence, the difference in the action sequence is obtained. The difference in the operation sequence is, for example, a difference in the event occurrence order and a difference in the event occurrence timing. The information on the difference in the order of occurrence (displacement information obtained from the difference in the order) includes, but is not limited to, the position of the wrong event, the phase difference of each event, and the event of the maximum phase difference (see FIG. 18).
[A-9]動作シーケンスの差の音楽情報への変換
学習者の動作シーケンスと参照となる動作シーケンスの差は、例えば、動作シーケンスの差の音響情報変換手段によって音響情報に変換され、音声出力部から出力できるようになっている。取得した差(一部あるいは全部)をどのように出力情報(音響情報)に変換するかについては限定されないが、例えば、学習者に音のシーケンスを提示するにあたり、音のシーケンスの中に差に関する情報を付加することが挙げられる。
[A-9] Conversion of difference in motion sequence to music information The difference between the learner's motion sequence and the reference motion sequence is converted into acoustic information by, for example, an acoustic information conversion means of the difference in motion sequence, and is output as voice. It can be output from the unit. There is no limitation on how to convert the acquired difference (part or all) into output information (acoustic information), but for example, when presenting a sound sequence to a learner, the difference is related to the sound sequence. Adding information can be mentioned.
例えば、学習者の動作後に音のシーケンスを提示する順序は、実際の学習者の動作シーケンスに対応する順序とし、参照の動作シーケンスと順番が異なるイベントに対応する音を変化させて提示する。より具体的には、参照となる動作シーケンスに対応する音を聴かせること、自己の動作シーケンスに対応する音を聴かせること、において、参照となる動作シーケンスを規定するイベントの順序と自己の動作シーケンスを規定するイベントの順序に差が生じた場合に、イベントの順序は学習者の動作シーケンスの順序(誤った順序)とし、誤ったイベントに対応する音を変化させて提示する。音の変化は、典型的には音高を変化させることで、半音上げること、あるいは、半音下げることを例示することができる(図17例2)。例えば、参照となる動作シーケンスに対応する複数の音を同時に鳴らした時に和音となるように各音を選択しておくことで、自己の動作シーケンスに対応する複数の音を同時に鳴らした時に不協和音(音高を変えたことによって)となることによって、自己の動作完了後に各イベントに対応する音を同時に鳴らすことによって、参照となる動作シーケンスと自己の動作シーケンスとの差の有無を直ちに認識することができる。 For example, the order in which the sound sequence is presented after the learner's action is the order corresponding to the actual learner's action sequence, and the sounds corresponding to the events whose order is different from the reference action sequence are changed and presented. More specifically, in listening to the sound corresponding to the reference action sequence and listening to the sound corresponding to the own action sequence, the sequence of events and the self action that define the reference action sequence are specified. When there is a difference in the order of the events that define the sequence, the order of the events is the order of the learner's action sequence (wrong order), and the sound corresponding to the wrong event is presented in a different manner. The change in sound can be exemplified by raising or lowering a semitone by typically changing the pitch (Fig. 17, Example 2). For example, by selecting each sound so that it becomes a chord when multiple sounds corresponding to the reference action sequence are played at the same time, a dissonant sound (dissonance) when multiple sounds corresponding to the own action sequence are played at the same time. By changing the pitch), the sound corresponding to each event is played at the same time after the self-motion is completed, so that the presence or absence of a difference between the reference motion sequence and the self-motion sequence is immediately recognized. Can be done.
あるいは、学習者の動作後に音を提示する順序は、参照の動作シーケンスに対応する順序とし、参照の動作シーケンスと順番が異なるイベント(例えば、順番が異なる複数のイベントの中で最初に生起するイベント)に対応する音を変化させて(例えば、ピッチベンド)提示してもよい。あるいは、音のシーケンスにおいて、参照の動作シーケンスに対応する順序を用いて、位相差が最大のイベントの音高を変えて提示するようにしてもよい(図19参照)。すなわち、最大位相差のイベントとその位相差を差情報として用いることができる。最も誤差の大きいイベントを選択することで、誤差の大きいイベントがなんであったかを学習者に提示して認識させることができる。具体的には、最大位相差の音響情報での位置を用い、理想時の順番と同じ場所の音に位相差の値を情報として付加する。最大位相差のイベントが複数ある場合には、得られた順番中1番前にあるものを採用する。常に1番前にあるものを優先することで意識するものを固定化することができる。さらに、位相差の大きさに合わせて音高を変えるようにしてもよい。 Alternatively, the order in which the sounds are presented after the learner's actions is the order corresponding to the reference action sequence, and the event that occurs in a different order from the reference action sequence (for example, the event that occurs first among a plurality of events in different orders). ) May be changed (for example, pitch bend) and presented. Alternatively, in the sound sequence, the pitch of the event with the maximum phase difference may be changed and presented by using the order corresponding to the reference operation sequence (see FIG. 19). That is, the event of the maximum phase difference and the phase difference thereof can be used as the difference information. By selecting the event with the largest error, the learner can be made aware of what the event with the largest error was. Specifically, the position of the maximum phase difference in the acoustic information is used, and the value of the phase difference is added as information to the sounds in the same place as the ideal order. If there are multiple events with the maximum phase difference, the one that is the first in the obtained order is adopted. By always giving priority to the one in front of you, you can fix what you are conscious of. Further, the pitch may be changed according to the magnitude of the phase difference.
参照となる動作シーケンスに対応する音のシーケンスを等間隔に変換して学習者に提示してもよい(図16参照)。学習者の動作シーケンスを規定するイベントの順序が参照の動作シーケンスと同じ場合であって、かつ、1つ以上のイベントの発生タイミングが異なる場合には、参照動作シーケンスに対応する音のシーケンスの発生タイミングに対して変化させてもよい(図17例1)。例えば、手本データの音のシーケンスを等間隔(0.2秒)で鳴らすのに対して、評価対象データの音のシーケンスを、あるイベントのタイミングで遅れている場合には例えば等間隔に対して0.1秒遅らせ、早い場合には等間隔に対して0.1秒早めて鳴らすようにしてもよい。なお、イベントの発生順序に差異がある場合には、イベントの発生間隔の差異の評価は行わないように設定してもよく、あるいは、イベントの発生順序に差異がある場合であっても、発生間隔の差異の評価を行ってもよい。 The sound sequence corresponding to the reference operation sequence may be converted into equal intervals and presented to the learner (see FIG. 16). When the order of the events that define the learner's action sequence is the same as the reference action sequence, and the timing of occurrence of one or more events is different, the sound sequence corresponding to the reference action sequence is generated. It may be changed with respect to the timing (Fig. 17, Example 1). For example, the sound sequence of the model data is sounded at equal intervals (0.2 seconds), while the sound sequence of the evaluation target data is delayed at the timing of a certain event, for example, at equal intervals. It may be delayed by 0.1 seconds, and if it is early, it may be sounded 0.1 seconds earlier than the equal interval. If there is a difference in the event occurrence order, it may be set not to evaluate the difference in the event occurrence interval, or even if there is a difference in the event occurrence order, it occurs. The difference in intervals may be evaluated.
動作シーケンスの差の音響情報への変換は、上記例に限定されるものではなく、例えば、学習者の運動(例えば、スイング)中に、動作シーケンスのイベントの順序が参照となる動作シーケンスのイベントの順序と違った時点でリアルタイムで音を鳴らしてもよい。また、参照となる動作シーケンスと比較してイベントの発生順序のズレが最も大きくなっていたイベントについて、そのイベントを特定する呼称や対応する身体部位を音声で提示してもよく、また、そのイベント発生時のフレームの画像を提示してもよい。 The conversion of the difference in the motion sequence into acoustic information is not limited to the above example, and for example, during the learner's movement (for example, swing), the event of the motion sequence is referred to as the event of the motion sequence. The sound may be played in real time at a time different from the order of. Further, for the event in which the deviation of the event occurrence order is the largest as compared with the reference operation sequence, the name for specifying the event and the corresponding body part may be presented by voice, and the event may be presented. An image of the frame at the time of occurrence may be presented.
[A-10]視覚情報の提示
各種情報を学習者に視覚的に提示することは有用である。例えば、学習者の身体運動の動画データや参照となる動画データを画像出力部から提示してもよく、あるいは、学習者の姿勢の時系列データを筋骨格モデルに当てはめて表示してもよい。各イベントに対応するフレーム画像を表示してもよく、また、各イベントに対応するフレーム画像を参照フレーム画像と学習者のフレーム画像で対比して表示してもよい。各イベントに対応する画像を骨格情報や筋骨格情報を備えた画像として単独あるいは参照と対比して表示してもよい。
[A-10] Presentation of visual information It is useful to visually present various information to the learner. For example, moving image data of the learner's physical exercise or moving image data as a reference may be presented from the image output unit, or time-series data of the learner's posture may be applied to a musculoskeletal model and displayed. The frame image corresponding to each event may be displayed, or the frame image corresponding to each event may be displayed in comparison with the reference frame image and the learner's frame image. The image corresponding to each event may be displayed alone or in comparison with the reference as an image having skeletal information and musculoskeletal information.
学習者の姿勢の時系列データを筋骨格モデルに当てはめて表示する場合には、リアルタイムで逆運動学計算、逆動力学計算、筋張力推定を行うことで得られる筋骨格モデルをディスプレイ表示するシステム(Akihiko Murai, Kosuke Kurosaki, Katsu Yamane and Yoshihiko Nakamura. “Musculoskeletal-see-throuhg mirror: Computational modeling and algorithm for whole-body muscle activity visualizetion in real time”. Progress in Biophysics and Molecular Biology, Vol. 103, pp. 310-317, 2010)を用いて、スイングにおける一連の骨格の運動を提示してもよい。このように、スイングの一連の運動における全身の骨格の動きを可視化することで、プレイヤーに自分のスイングの理解を促すことができる。 A system that displays the musculoskeletal model obtained by performing inverse kinematics calculation, inverse kinematics calculation, and muscle tension estimation in real time when the time-series data of the learner's posture is applied to the musculoskeletal model and displayed. (Akihiko Murai, Kosuke Kurosaki, Katsu Yamane and Yoshihiko Nakamura. “Musculoskeletal-see-throuhg mirror: Computational modeling and algorithm for whole-body muscle activity visualization in physiology in real time” 310-317, 2010) may be used to present a series of skeletal movements in a swing. In this way, by visualizing the movement of the skeleton of the whole body in a series of swing movements, it is possible to encourage the player to understand his / her swing.
参照となる動作シーケンスと比較してイベントの発生順序のズレが最も大きくなっていたイベントについて、その時のフレームの画像を提示して、その身体部位の動作を注視するように促してもよい。学習者の動作シーケンスと参照となる動作シーケンスの対比や差を図やグラフを用いて視覚的に学習者に提示してもよい。学習者の身体運動の動作分析結果(参照となる運動データとの比較を含んでいてもよい)を図やグラフを用いて表示してもよい。学習者の動作データから取得したイベント順序の入れ替わりのパターンとスイングの癖や修正方法の対応表を用意しておき、得られた動作シーケンス(イベントの順序)にしたがって、学習者に修正方法を提示(視覚的ないし聴覚的に)してもよい。 For the event in which the deviation of the event occurrence order is the largest as compared with the reference motion sequence, an image of the frame at that time may be presented to encourage the user to pay close attention to the motion of the body part. The contrast or difference between the learner's motion sequence and the reference motion sequence may be visually presented to the learner using a diagram or a graph. The learner's physical exercise motion analysis results (which may include comparison with reference exercise data) may be displayed using figures and graphs. Prepare a correspondence table of the change pattern of the event order acquired from the learner's motion data and the swing habit and the correction method, and present the correction method to the learner according to the obtained motion sequence (event order). It may be (visually or audibly).
[A-11]対象となる身体運動
対象となる身体運動は、限定されないものの、ゴルフスイング、野球のバットのスイング、野球の投球動作、空手の型、ダンス、リハビリテーション、その他の動作のシーケンスが確立されている運動やルーティーンが決まっている運動等を例示することができる。例えば、ゴルフは動作のシーケンスが確立されているスポーツの1つである。ダンスの練習やリハビリテーションにおいても動作のシーケンスを獲得するところからトレーニングが始まると考えられる。
[A-11] Target physical exercise The target physical exercise is not limited, but a sequence of golf swing, baseball bat swing, baseball pitching motion, karate kata, dance, rehabilitation, and other motions is established. It is possible to exemplify exercises that are being performed and exercises that have a fixed routine. For example, golf is one of the sports for which a sequence of movements has been established. In dance practice and rehabilitation, training starts from the point of acquiring a sequence of movements.
後述する実施形態に示すように、例えばゴルフスイングを例にとると、その動作シーケンスは、(1)左膝伸展開始、(2)左手首固定解除、(3)腰のブレーキ、(4)手の速さ最大、(5)ヘッドの速さ最大、(6)肩のブレーキの6つのイベントからなり、所定の部位の角度変化の時刻や所定の部位の最大速度の時刻等を利用して各イベントの時刻を取得する。後述する実施形態では、光学式モーションキャプチャにより測定された各マーカの3次元位置情報の時系列データから、角度変化の時刻や速さ最大の時刻を利用して、各イベントの時刻を算出する。ゴルフのスイングにおいて、着目するイベントのセットの正解は必ずしも1つではなく、上記6つのイベントは例示である。これらの一部のイベントから参照動作シーケンスを構成してもよく、あるいは他のイベントから参照動作シーケンスを構成してもよい。キネマティックシーケンスを特定する他のイベントのセットを選択し得ることが当業者に理解される。また、対象の姿勢の時系列データが得られていれば、いかなるイベントであっても、その発生時刻を特定することが可能である。 As shown in the embodiment described later, for example, taking a golf swing as an example, the operation sequences are (1) start of left knee extension, (2) release of left wrist fixation, (3) waist brake, and (4) hand. It consists of 6 events: maximum speed of (5) maximum speed of head, (6) brake of shoulder, and each uses the time of angle change of a predetermined part and the time of maximum speed of a predetermined part. Get the time of the event. In the embodiment described later, the time of each event is calculated from the time series data of the three-dimensional position information of each marker measured by the optical motion capture, using the time of the angle change and the time of the maximum speed. In a golf swing, the correct answer for a set of events of interest is not necessarily one, and the above six events are examples. The reference operation sequence may be composed of some of these events, or the reference operation sequence may be composed of other events. It will be appreciated by those skilled in the art that a set of other events that identify the kinematic sequence may be selected. Further, as long as the time series data of the posture of the target is obtained, it is possible to specify the occurrence time of any event.
[B]ゴルフスイングへの適用
ゴルフは動作のシーケンスが確立されているスポーツの1つである。ゴルフでは、スイング動作のシーケンスはKinematic Sequence(キネマティック・シーケンス)として知られている(非特許文献7、図20)。なお、後述する説明においては右利きのスイングにしたがって説明するが、左利きのスイングについては、左右を入れ替えて議論すればよい。
[B] Application to golf swing Golf is one of the sports in which a sequence of movements is established. In golf, the sequence of swing movements is known as the Kinematic Sequence (Non-Patent Document 7, FIG. 20). In the explanation described later, the explanation will be given according to the right-handed swing, but the left-handed swing may be discussed by exchanging the left and right sides.
[B-1]特徴点と各時刻の定義
本実施形態では、光学式モーションキャプチャによりスイング時の身体運動を計測し、全身及ゴルフクラブに貼付したマーカ位置の時間変化に基づいてスイングの運動データを取得する。本実施形態では、全身及びゴルフクラブにおいて注目する特徴点及び、ゴルフスイングの工程について、それぞれ以下のように定義する。これらの特徴点の選択や定義は例示であって、本発明の内容を限定するものではない。
[B-1] Definition of feature points and each time In this embodiment, the body movement during the swing is measured by optical motion capture, and the swing movement data is based on the time change of the marker position attached to the whole body and the golf club. To get. In the present embodiment, the feature points of interest in the whole body and the golf club and the golf swing process are defined as follows. The selection and definition of these feature points are examples and do not limit the contents of the present invention.
[B-1-1]全身とゴルフクラブの特徴点の定義
全身から特定の身体特徴点を選出するにあたり、ゴルフ指導において運動を評価する上で一般的に注目する要素について、プロのゴルフ指導者に聞き取り調査を行った。本実施形態では、表1に示した身体上の特徴点(計測したマーカ位置(図6))について、各身体特徴点の中心位置(図7、図8)を以下のように定義し、それらを結ぶことで身体部位を定義した(表1、表2)。

Figure JPOXMLDOC01-appb-T000001

Figure JPOXMLDOC01-appb-T000002
[B-1-1] Definition of whole body and golf club feature points Professional golf instructors focus on the factors that are generally noted in evaluating exercise in golf coaching when selecting specific physical feature points from the whole body. We conducted an interview survey. In the present embodiment, with respect to the feature points on the body (measured marker positions (FIG. 6)) shown in Table 1, the center positions (FIGS. 7 and 8) of the body feature points are defined as follows, and they are defined as follows. Body parts were defined by connecting (Tables 1 and 2).

Figure JPOXMLDOC01-appb-T000001

Figure JPOXMLDOC01-appb-T000002
[肩]
右肩前部位置xrf,sh及び右肩後部位置xrb,shの平均をとり、右肩中心位置xr,shと定義する。左肩中心位置xl,shについても同様に、左肩前部位置xlf,sh及び左肩後部位置xlb,shに基づいて定義する。この2点を結んだものを肩の線分とする(図9)。
Figure JPOXMLDOC01-appb-I000003
 
[shoulder]
The average of the right shoulder front position x rf, sh and the right shoulder rear position x rb, sh is taken and defined as the right shoulder center position x r, sh . Similarly, the left shoulder center position x l, sh is defined based on the left shoulder front position x lf, sh and the left shoulder rear position x lb, sh . The shoulder line segment connects these two points (Fig. 9).
Figure JPOXMLDOC01-appb-I000003
[腰]
骨盤左側及び右側の位置をそれぞれ左腰位置xl,hi、右腰位置xr,hiと定義し、この2点を結んだ線分を腰の線分とする(図9)。
 
[Waist]
The left and right positions of the pelvis are defined as the left hip position x l, hi and the right hip position x r, hi, respectively, and the line segment connecting these two points is defined as the waist line segment (Fig. 9).
[上体軸]
上記の肩の線分の中点xu,ax及び腰の線分の中点xl,axの2点を結んだ線分を上体軸とする(図9)。
Figure JPOXMLDOC01-appb-I000004
[Upper body axis]
The upper body axis is the line segment connecting the two points of the midpoint x u, ax of the shoulder line segment and the midpoint x l, ax of the waist line segment (Fig. 9).
Figure JPOXMLDOC01-appb-I000004
[両腕]
右手首外側xro,wr、内側xri,wr、左手首外側xlo,wr、内側xli,wrの計4箇所の位置を平均し、その値を手首の中心位置xwrと定義する(図8)。右肩中心xr,shと手首中心xwrを結んだ線分を右腕、左肩中心xl,shと手首中心xwrを結んだ線分を左腕とする(図9)。
Figure JPOXMLDOC01-appb-I000005
[Both arms]
The right wrist outside x ro, wr , inside x ri, wr , left wrist outside x lo, wr , inside x li, wr are averaged, and the value is defined as the center position of the wrist x wr ( FIG. 8). The line segment connecting the right shoulder center x r, sh and the wrist center x wr is the right arm, and the line segment connecting the left shoulder center x l, sh and the wrist center x wr is the left arm (Fig. 9).
Figure JPOXMLDOC01-appb-I000005
[ゴルフクラブ]
ゴルフクラブのヘッド2箇所(フェース外側xo,fa、内側xi,fa)の位置を平均し、その値をフェースの中心位置xfaと定義する。そして、手首中心xwrとゴルフクラブのフェース中心xfaの2点を結んだ線分をゴルフクラブとする(図9)。
Figure JPOXMLDOC01-appb-I000006
[Golf club]
The positions of the two heads of the golf club (outside the face x o, fa , inside x i, fa ) are averaged, and the value is defined as the center position x fa of the face. Then, the line segment connecting the two points of the wrist center x wr and the golf club face center x fa is defined as the golf club (Fig. 9).
Figure JPOXMLDOC01-appb-I000006
[左脚]
左膝外側xlo,kn及び内側xli,knの位置を平均し、その位置を左膝の中心位置xl,knと定義する.左足首外側xlo,an及び内側xli,anの位置を平均し、その位置を左足首の中心位置xl,anと定義する.左腰xl,hiと左膝中心xl,knの2点を結んだ線分を左大腿とし、左膝中心xl,knと左足首中心xl,anの2点を結んだ線分を左下腿とする(図10)。
Figure JPOXMLDOC01-appb-I000007
[Left leg]
The positions of the lateral x lo, kn and medial x li, kn of the left knee are averaged, and the positions are defined as the central position x l, kn of the left knee. The positions of the lateral x lo, an and medial x li, an of the left ankle are averaged, and the positions are defined as the central position of the left ankle x l, an . The line segment connecting the left hip x l, hi and the left knee center x l, kn is the left thigh, and the line segment connecting the left knee center x l, kn and the left ankle center x l, an . Is the left lower leg (Fig. 10).
Figure JPOXMLDOC01-appb-I000007
[左前腕]
左肘外側xlo,el及び内側xli,elの位置を平均し、その位置を左肘の中心位置xl,elと定義する。左手首外側xlo,wr及び内側xli,wrの位置を平均し、その位置を左手首の中心位置xl,wrと定義する。左肘中心xl,elと左手首中心xl,wrの2点を結んだ線分を左前腕とする(図11)。
Figure JPOXMLDOC01-appb-I000008
[Left forearm]
Hidarihijisotogawa x lo, el and the inner x li, averaged position of el, defines its position center position x l left elbow, el a. The positions of the outer side x lo, wr and the inner side x li, wr of the left wrist are averaged, and the position is defined as the center position x l, wr of the left wrist. The line segment connecting the two points of the center of the left elbow x l, el and the center of the left wrist x l, wr is the left forearm (Fig. 11).
Figure JPOXMLDOC01-appb-I000008
[クラブシャフト]
ゴルフクラブのシャフト部分について、シャフト右側の位置xrg,sh及びシャフト左側の位置xlg,shを平均してシャフト位置xshと定義する。この点からクラブフェース中心位置xfaを結んだ線分をクラブシャフトの方向とする(図11)。
Figure JPOXMLDOC01-appb-I000009
[Club shaft]
For the shaft part of a golf club, the position x rg, sh on the right side of the shaft and the position x lg, sh on the left side of the shaft are averaged and defined as the shaft position x sh . From this point, the line segment connecting the center position of the club face x fa is defined as the direction of the club shaft (FIG. 11).
Figure JPOXMLDOC01-appb-I000009
[B-1-2]ゴルフスイングにおける各時刻の定義
ゴルフクラブのフェース中心位置xfaに基づき、ゴルフスイングの一連の運動を複数工程に分割し、開始時刻、終了時刻、及び各工程の境界時刻を以下のように定義した。以下の4つの時刻の間をそれぞれ、バックスイング、ダウンスイング、フォロースルーと定義する。
[B-1-2] Definition of each time in the golf swing Based on the face center position x fa of the golf club, a series of movements of the golf swing is divided into a plurality of steps, and the start time, the end time, and the boundary time of each step are divided. Was defined as follows. The following four times are defined as backswing, downswing, and follow-through, respectively.
(1)アドレス時刻tadd
初期姿勢を決定し、ゴルフクラブを後方に振り上げはじめる時刻。すなわち、フェース中心位置xfaが静止状態から打球反対方向へ移動し始める時刻と定義する。
(1) Address time t add
The time to determine the initial posture and start swinging the golf club backwards. That is, it is defined as the time when the face center position x fa starts to move from the stationary state in the opposite direction to the hit ball.
(2)トップ時刻ttop
ゴルフクラブの後方への振り上げが終了し、ボールへ向かってゴルフクラブの振り下ろしを開始する時刻。
(2) Top time t top
The time when the golf club finishes swinging backward and the golf club starts swinging down toward the ball.
(3)インパクト時刻timp
ゴルフクラブの振り下ろしが完了し、ボールとゴルフクラブのヘッド前面が衝突する時点。フェース中心位置がアドレス時の位置xfa(tadd)との距離が最小になった時点として定義する。
(3) Impact time t imp
When the golf club has been swung down and the ball collides with the front of the golf club head. The face center position is defined as the time when the distance from the address x fa (t add ) is the minimum.
(4)フィニッシュ時刻tfin
スイング運動の一連の動作の終了時刻。インパクト後のフェース中心位置xfaの移動速度が最も小さくなる時刻と定義する。
(4) Finish time t fin
The end time of a series of swing movements. It is defined as the time when the movement speed of the face center position x fa after impact is the smallest.
[B-3]身体部位に着目した動作解析
ゴルフ指導者1名に対して、光学式モーションキャプチャを用いてゴルフスイングの運動計測実験を実施した。本実施形態では、光学式モーションキャプチャに加えて、筋電計及び床反力計を用いてゴルフスイングの運動計測実験を実施した。筋電計及び床反力計は、ゴルフスイング時の動作シーケンスを取得することには用いないが、例えば、筋骨格モデルをディスプレイ表示したい場合に用いられる。光学式モーションキャプチャに用いるマーカ及び筋電位センサを装着した部位を表3に示す。

Figure JPOXMLDOC01-appb-T000010
[B-3] Motion analysis focusing on body parts A golf swing motion measurement experiment was conducted for one golf instructor using optical motion capture. In this embodiment, in addition to the optical motion capture, a golf swing motion measurement experiment was carried out using an electromyogram and a floor reaction force meter. The electromyogram and the floor reaction force meter are not used to acquire the motion sequence during a golf swing, but are used, for example, when a musculoskeletal model is to be displayed on a display. Table 3 shows the parts where the markers and myoelectric potential sensors used for optical motion capture are attached.

Figure JPOXMLDOC01-appb-T000010
ゴルフボールを実際に打ち、その運動における全身のマーカ位置を計測した。使用するドライバーにマーカを貼付することで、クラブの運動も同時に計測した。また、計測を行ったショット数は12球で、各ショットについて弾道測定器を用いて打球の計測も行った。各ショット後にショットの打感及びスイング姿勢について、自己評価をしてもらい、この自己評価の点数が高いものを成功したショット、低いものを失敗したショットとした。点数が最も高かった試行については、手本のショットとした。ゴルフの指導において、ショットを視覚的に評価する際に最も注目する身体部位に絞り、動作解析を行った。その上で成功スイングと失敗スイングとで比較し、スイングの出来に強く影響する要素を採用する。 The golf ball was actually hit and the marker position of the whole body in the movement was measured. By attaching a marker to the driver to be used, the movement of the club was also measured at the same time. In addition, the number of shots measured was 12, and the number of hit balls was also measured using a ballistic measuring device for each shot. After each shot, the shot feeling and swing posture were self-evaluated, and the one with a high self-evaluation score was regarded as a successful shot, and the one with a low self-evaluation score was regarded as a failed shot. The trial with the highest score was used as a model shot. In golf instruction, we focused on the body parts that are most noticeable when visually evaluating shots, and performed motion analysis. After that, the success swing and the failure swing are compared, and the factors that strongly affect the swing performance are adopted.
上記の通り定義された各身体部位について、ゴルフスイングの身体運動と各身体部位の動作の関連性を検討した。動作分析にあたり、骨盤右上端のアドレス姿勢からの高さ変化、骨盤左上端のアドレス姿勢からの高さ変化、肩のアドレス姿勢からの角度変化、腰のアドレス姿勢からの角度変化、上体軸のアドレス姿勢からの角度変化、上体軸の鉛直方向からの傾き、ゴルフクラブのシャフトの鉛直方向からの傾き、右肩と手首との距離のアドレス姿勢からの変化、左肩と手首との距離のアドレス姿勢からの変化、肩と腰のアドレス姿勢からの角度変化の差分の計10項目を採用した。 For each body part defined as described above, the relationship between the body movement of the golf swing and the movement of each body part was examined. In the motion analysis, the height change from the address posture of the upper right corner of the pelvis, the height change from the address posture of the upper left corner of the pelvis, the angle change from the address posture of the shoulder, the angle change from the address posture of the waist, the upper body axis. Angle change from address posture, tilt of upper body axis from vertical direction, tilt of golf club shaft from vertical direction, change from address posture of distance between right shoulder and wrist, address of distance between left shoulder and wrist A total of 10 items were adopted: the difference from the posture and the difference between the angle change from the address posture of the shoulder and the waist.
スイング中の各身体部位の動作を調査した。インパクトの直前に腰が鉛直上方向に移動し、フォロースルーで打球方向へ移動していくこと、それに伴って上体軸が打球方向に対して反対方向(学習者に対して右方向)に倒れていくこと、両腕とゴルフクラブとのなす角はアドレスからダウンスイングのある時点までほぼ一定の値で維持されているが、その時点から急激に角度が変化していることが観測された。腰の上昇は左膝の伸展によるものであり、インパクトの直前に左脚で床を踏み込むことで腰の回転運動にブレーキをかけ、全身の重心位置を上方向に移動させることで手首及びゴルフクラブのヘッドの速度を増加させる役割を担っているのではないかと考えられる。上体軸の打球反対方向への倒れこみは、この全身の回転運動へのブレーキ及び腰の移動によってもたらされる変化であり、回転運動によって全身の重心が打球方向への過度な移動を防ぐ意味も有していると考えられる。手とゴルフクラブのなす角の変化についても、ゴルフクラブのヘッドの速度を増加させる意味合いが示唆される。インパクト直前まで手首が曲がった状態で固定することで、回転半径を小さく保ち、回転速度を高くできる。その状態でインパクトの直前に左手首の固定を解除することで、回転速度が高い状態で回転半径が大きくなり、ヘッドの速度が増加していると考えられる。 The movement of each body part during the swing was investigated. Immediately before the impact, the hips move vertically upward and follow through to move in the hitting direction, and the upper body axis falls in the opposite direction to the hitting direction (to the right with respect to the learner). Going forward, the angle between both arms and the golf club was maintained at an almost constant value from the address to a certain point in the downswing, but it was observed that the angle changed sharply from that point. The rise of the hips is due to the extension of the left knee. Immediately before the impact, the left leg steps on the floor to brake the rotational movement of the hips, and the center of gravity of the whole body is moved upward to move the wrist and golf club. It is thought that it plays a role in increasing the speed of the head. The collapse of the upper body axis in the opposite direction to the hitting ball is a change brought about by the braking and the movement of the waist to the rotational movement of the whole body, and it also has a meaning to prevent the center of gravity of the whole body from excessive movement in the hitting direction by the rotational movement. it seems to do. The change in the angle between the hand and the golf club also suggests the implication of increasing the speed of the golf club head. By fixing the wrist in a bent state until just before the impact, the radius of gyration can be kept small and the rotation speed can be increased. It is considered that by releasing the fixation of the left wrist immediately before the impact in that state, the radius of gyration increases while the rotation speed is high, and the speed of the head increases.
[B-2]ゴルフスイングにおける動作シーケンスの評価
身体部位の動作及びその動作を引き起こしている別の身体部位の動作の時間変化について、それぞれの動作上の転換点(イベント)の間隔や順序についての調査を行う。今回の解析では、Kinematic Sequence(キネマティック・シーケンス)において着目される腰、肩、腕、ゴルフクラブの回転動作について、それぞれの動作の時系列的な関係を調査した。また、前節の解析によって重要性が示唆された、手首の固定及びその解除の動作、左膝を伸展させ床を蹴り上げる動作についても同様の調査を行った。特定部位の運動上の転換点(イベント)のタイミング評価6項目(左膝の角度最大となる時刻、左手首の固定が解除される時刻、腰の回転角速度が最大となる時刻、肩の回転角速度が最大となる時刻、手首の速度が最大となる時刻、ゴルフクラブのヘッド速度が最大となる時刻) を採用した。これら6つのイベントは、ゴルフ指導者である学習者への聞き取り調査及びゴルフの教則(非特許文献2)に基づいて選定したものであるが、例示であって、本発明を限定するものではない。各イベントは学習者に取り付けられた身体およびゴルフクラブ上のマーカの3次元位置座標から計算される。
[B-2] Evaluation of motion sequence in golf swing Regarding the temporal changes in the motion of a body part and the motion of another body part causing the motion, the interval and order of turning points (events) in each motion. Conduct a survey. In this analysis, we investigated the time-series relationships of the rotational movements of the waist, shoulders, arms, and golf clubs, which are the focus of attention in the Kinematic Sequence. In addition, the same investigation was conducted on the movements of fixing and releasing the wrist and the movement of extending the left knee and kicking up the floor, which were suggested to be important by the analysis in the previous section. 6 items of timing evaluation of the turning point (event) in the movement of a specific part (time when the angle of the left knee becomes maximum, time when the fixation of the left wrist is released, time when the rotation angular velocity of the waist becomes maximum, rotation angular velocity of the shoulder The time when the maximum speed, the time when the wrist speed is maximum, and the time when the head speed of the golf club is maximum) are adopted. These six events were selected based on an interview survey with a learner who is a golf instructor and a golf instruction (Non-Patent Document 2), but are examples and do not limit the present invention. .. Each event is calculated from the three-dimensional position coordinates of the body attached to the learner and the markers on the golf club.
(1)腰の回転
前節で定義した腰の線分(例えば、骨盤左側と骨盤右側に取り付けられたマーカを結ぶ線分)について、アドレス時のデータ(線分)を基準とし、任意の時刻における腰の線分と基準の線分のそれぞれを水平面に射影し、各時刻における2線分のなす角を計算し、その値を腰の回転角と定義する。本実施形態では、腰の回転角速度が最大となる時刻を注目イベントとする。
(1) Waist rotation For the waist line segment defined in the previous section (for example, the line segment connecting the markers attached to the left side of the pelvis and the right side of the pelvis), at an arbitrary time based on the data (line segment) at the time of addressing. Each of the waist line segment and the reference line segment is projected onto the horizontal plane, the angle formed by the two line segments at each time is calculated, and the value is defined as the rotation angle of the waist. In the present embodiment, the time when the rotational angular velocity of the waist becomes maximum is set as the attention event.
(2)肩の回転
腰の回転と同様に、前節で定義した肩の線分(右肩前部位置と右肩後部位置の平均を右肩中心位置とし、左肩前部位置と左肩後部位置の平均を左肩中心位置とする。これら右肩中心位置と左肩中心位置を結ぶ線分)について、基準(アドレス時刻の線分)と任意の時点における線分の水平面への射影それぞれのなす角から肩の回転角を定義する。本実施形態では、肩の回転角速度が最大となる時刻を注目イベントとする。
(2) Rotation of the shoulder Similar to the rotation of the waist, the line segment of the shoulder defined in the previous section (the average of the front position of the right shoulder and the rear position of the right shoulder is the center position of the right shoulder, and the front position of the left shoulder and the rear position of the left shoulder The average is the center position of the left shoulder. Regarding the line segment connecting the center position of the right shoulder and the center position of the left shoulder), the shoulder is formed from the angle formed by the reference (line segment at the address time) and the projection of the line segment on the horizontal plane at any time. Define the rotation angle of. In the present embodiment, the time when the rotational angular velocity of the shoulder becomes maximum is set as the event of interest.
(3)腕の運動
前節で定義した手首の中心位置(右手首外側と内側、左手首外側と内側の4箇所の位置の平均を手首中心位置)について、任意の時点における速度の大きさを計算し、手首の運動の速さが最大となる時刻を注目イベントとする。
(3) Arm exercise Calculate the magnitude of speed at any time point for the center position of the wrist defined in the previous section (the average of the four positions on the outside and inside of the right wrist and the outside and inside of the left wrist is the center position of the wrist). However, the time when the wrist movement speed is maximized is the notable event.
(4)手首の固定及びその解除
スイング中の両腕とゴルフクラブとの角度の維持は、左手首の固定によって実現されている。そこで、手首の固定を表現するパラメータとして、左前腕とクラブシャフトとのなす角を考える。任意の時点における左前腕の方向とゴルフクラブの方向とのなす角を計算し、左手首の角度と定義する。この角度について、インパクトの直前に角度変化の符号が変わる瞬間を手首の固定が解除された時刻を左手首固定解除と定義し、注目イベントとする。
(4) Fixing and releasing the wrist Maintaining the angle between both arms and the golf club during the swing is realized by fixing the left wrist. Therefore, consider the angle between the left forearm and the club shaft as a parameter that expresses the fixation of the wrist. Calculate the angle between the direction of the left forearm and the direction of the golf club at any point in time, and define it as the angle of the left wrist. Regarding this angle, the moment when the sign of the angle change changes immediately before the impact is defined as the time when the wrist is released from the fixation, and the time when the wrist is released is defined as the event of interest.
(5)左膝の伸展
左大腿と左下腿とのなす角を左膝の角度と定義する。左膝については、床を蹴り上げる直前の最も膝が曲がった状態、すなわち左膝の角度が最小となる時刻を注目イベントとする。
(5) Extension of the left knee The angle between the left thigh and the left lower leg is defined as the angle of the left knee. For the left knee, the state in which the knee is bent most immediately before kicking up the floor, that is, the time when the angle of the left knee is minimized is set as a noteworthy event.
(6)ゴルフクラブのヘッドの運動
前節で定義したゴルフクラブのヘッドの位置(フェース中心位置)について、任意の時点における速度の大きさを計算し、ヘッドの運動の速さが最大になる時刻を注目イベントとする。
(6) Golf club head movement For the golf club head position (face center position) defined in the previous section, calculate the magnitude of the speed at any time point, and set the time when the head movement speed becomes maximum. Make it a noteworthy event.
着目したイベントの発生タイミングは、左膝の伸展開始、左手首の固定解除、腰のブレーキ、手の速さ最大、ゴルフクラブのヘッドの速さ最大、インパクト、肩のブレーキの順で発生しており、インパクトの瞬間はゴルフクラブのヘッドの速さが最大となる時刻とほとんど一致していた。また、それぞれのショットにおいて左膝の伸展開始から肩の戻りまでに経過した時間はほぼ一致したが、左手首の固定解除のタイミングはばらつきが見られた。特に成功したショットにおいては、左手首の固定解除のタイミング以外はほぼ同じタイミングでイベントが発生していた一方で、失敗したショットでは、左膝の伸展開始から腰の戻りまでの間隔が短くなっていた。この結果から、抽出したイベントについてのタイミングの調査がゴルフスイングの評価に有効であることが確認された。 The timing of the event of interest is the start of extension of the left knee, the release of the left wrist, the brake of the waist, the maximum speed of the hand, the maximum speed of the golf club head, the impact, and the brake of the shoulder. The moment of impact almost coincided with the time when the speed of the golf club head was maximum. In each shot, the time elapsed from the start of extension of the left knee to the return of the shoulder was almost the same, but the timing of releasing the fixation of the left wrist was uneven. Especially in successful shots, the event occurred at almost the same timing except for the timing of releasing the fixation of the left wrist, while in the unsuccessful shot, the interval from the start of extension of the left knee to the return of the hip was shortened. It was. From this result, it was confirmed that the investigation of the timing of the extracted events is effective for the evaluation of the golf swing.
[B-3]学習者のゴルフスイングの評価
ゴルフスイングの動作シーケンスを規定する6つのイベントについて、発生時刻及び発生順序を解析した結果を、手本のデータと比較し、聴覚に対して音声のフィードバックをおこなう。まず、手本となるスイングにおける各イベントの発生時刻と発生順序をそれぞれtref i 、nref iとする。
[B-3] Evaluation of Learner's Golf Swing For the six events that define the movement sequence of the golf swing, the results of analyzing the time of occurrence and the order of occurrence are compared with the model data, and the sound is heard with respect to hearing. Give feedback. First, let t ref i and n ref i be the occurrence times and order of each event in the model swing.
次に、手本のショットの各イベントの発生間隔を等間隔tにするために以下のような行列Mを設定し、評価対象のスイングにおいても、各イベントの発生時刻tiにMによる処理を施す(図16参照)。
Figure JPOXMLDOC01-appb-I000011
ただし、Mは以下で表される行列である。
Figure JPOXMLDOC01-appb-I000012
Next, the following matrix M is set in order to make the occurrence interval of each event of the model shot equal to t, and even in the swing to be evaluated, the processing by M is performed at the occurrence time t i of each event. Apply (see FIG. 16).
Figure JPOXMLDOC01-appb-I000011
However, M is a matrix represented by the following.
Figure JPOXMLDOC01-appb-I000012
次に、手本のスイングについて、それぞれのイベントに異なる周波数fref i音を対応させる。ここで、href iは、Aの音(周波数440Hz)からの音程差を半音の数で表現したものである。
Figure JPOXMLDOC01-appb-I000013
本実施形態では、上記6音を同時に鳴らすと、C11と呼ばれる和音になるように設定した。各イベントへの音の割り当てを表4に示す。
Figure JPOXMLDOC01-appb-T000014
Next, for the model swing, each event is associated with a different frequency f ref i sound. Here, h ref i expresses the pitch difference from the sound of A (frequency 440 Hz) by the number of semitones.
Figure JPOXMLDOC01-appb-I000013
In the present embodiment, when the above six sounds are played at the same time, a chord called C11 is set. Table 4 shows the assignment of sounds to each event.
Figure JPOXMLDOC01-appb-T000014
続いて、評価対象のスイングについて、各イベントに音を対応させる。ここで、手本スイングと比較して、そのイベントの発生順序が早くなっていた場合に音程を半音上にずらし、遅くなっていた場合に音程を半音下にずらして対応させる。以上のデータを利用して、周波数fref iの音声を時間tだけ順番に鳴らしていくことで手本のスイングのタイミング評価を提示し、手本のタイミング評価提示終了時刻から時間t´i後に周波数fiの音声が再生されるようにする。 Next, for the swing to be evaluated, the sound is associated with each event. Here, as compared with the model swing, if the event occurrence order is earlier, the pitch is shifted up by a semitone, and if it is late, the pitch is shifted down by a semitone. Using the above data, the timing evaluation of the swing of the model is presented by playing the sound of the frequency f ref i in order for the time t, and after the time t ´ i from the end time of the presentation of the timing evaluation of the model. Make the sound of frequency f i play.
このように、手本と評価対象のスイングについて音声が再生される順序や再生時間を変化させることで、各タイミングのズレの直観的な提示を行う。その後、評価対象のスイングにおける6つの音声を同時に再生する。これによって、イベント発生順序が手本と異なっている場合に不協和音として音響情報を提示することになり、学習者に順序のズレを直感的に提示することができる。 In this way, by changing the order in which the sounds are reproduced and the reproduction time for the model and the swing to be evaluated, the deviation of each timing is intuitively presented. After that, the six sounds of the swing to be evaluated are simultaneously reproduced. As a result, when the event occurrence order is different from the model, the acoustic information is presented as a dissonance, and the learner can intuitively present the order deviation.
図17において、「手本」は参照となる動作シーケンスのイベントの時系列である。「例1」は、左手首が遅れ、肩ブレーキが早かった場合であり、順序誤差は無い。2音目が遅れ、6音目が早まっている。「例2」は、手首とヘッドの順序が逆転した場合であり、イベント間隔は同じに設定している。4音目が半音低く、5音目が半音高くなっており、順序が逆転している(間隔はすべて0.2秒のまま)。 In FIG. 17, a "model" is a time series of events in a reference motion sequence. "Example 1" is a case where the left wrist is delayed and the shoulder brake is early, and there is no order error. The second note is delayed and the sixth note is earlier. "Example 2" is a case where the order of the wrist and the head is reversed, and the event intervals are set to be the same. The 4th note is a semitone lower and the 5th note is a semitone higher, and the order is reversed (all intervals remain 0.2 seconds).
プロのゴルフ指導者に対して、ゴルフスイングにおける運動計測及び解析を行うことで、運動における特徴的な身体部位及びその動作の確認し、ゴルフ技能の評価する上での評価項目の選定を行い、姿勢評価10項目および、特定部位の運動上の転換点のタイミング評価6 項目を採用した。上記の選定に基づいて、ゴルフ技能を向上させるための評価結果について、手本との比較を聴覚情報を用いて提示した。本発明では、解析結果(全動作の全身骨格運動の時系列情報)をもとに、理想的なキネマティックシーケンスとの差を計算で求め、それを音響情報として動作後直ちに提示することで、学習者自らの動作を自覚させ、修正を行うことができる直感的な方法を考案した。 By measuring and analyzing the movement of a golf swing for a professional golf instructor, the characteristic body parts and their movements in the movement are confirmed, and the evaluation items for evaluating the golf skill are selected. Ten items of posture evaluation and six items of timing evaluation of the turning point in the movement of a specific part were adopted. Based on the above selection, the evaluation results for improving golf skills were presented in comparison with the model using auditory information. In the present invention, the difference from the ideal kinematic sequence is calculated based on the analysis result (time-series information of the whole body skeletal movement of all movements), and it is presented as acoustic information immediately after the movement. We devised an intuitive method that allows the learner to be aware of his or her own movements and make corrections.
本発明は、人間の運動計測、動作解析に基づいて自身の動作シーケンスと理想的な動作シーケンスとの差を、リアルタイムあるいはニア・リアルタイムで計算し、音響情報に変換して直ちに伝えることで、自らの動作の修正すべき点を直感的に把握させて学習者のトレーニング効果を高めるものであり、運動トレーニング、ゴルフトレーニング、ダンストレーニング、リハビリテーション、運動支援等に利用可能である。
 
 
 
 
The present invention calculates the difference between its own motion sequence and the ideal motion sequence based on human motion measurement and motion analysis in real time or near real time, converts it into acoustic information, and immediately conveys it. It is intended to intuitively grasp the points to be corrected in the movement of the learner and enhance the training effect of the learner, and can be used for exercise training, golf training, dance training, rehabilitation, exercise support, and the like.



Claims (16)

  1.  身体運動の第1動作シーケンスを参照として用意し、
     学習者の身体運動の運動データから第2動作シーケンスを取得し、
     第1動作シーケンスと第2動作シーケンスの差を音響情報として学習者に提示する、
     トレーニング支援方法。
    Prepare the first movement sequence of physical exercise as a reference,
    Obtain the second motion sequence from the exercise data of the learner's physical exercise,
    Presenting the difference between the first motion sequence and the second motion sequence to the learner as acoustic information,
    Training support method.
  2.  第1動作シーケンスと第2動作シーケンスの差を取得することを含み、
     取得した差を音響情報として学習者に提示する、
     請求項1に記載のトレーニング支援方法。
    Including acquiring the difference between the first operation sequence and the second operation sequence
    Present the acquired difference to the learner as acoustic information,
    The training support method according to claim 1.
  3.  動作シーケンスは、異なる時刻で発生する複数のイベントによって規定される、
     請求項1、2いずれか1項に記載のトレーニング支援方法。
    The operation sequence is defined by multiple events that occur at different times.
    The training support method according to any one of claims 1 and 2.
  4.  第1動作シーケンスと第2動作シーケンスの差を取得することを含み、
     前記差には、イベントの発生順序の差、および/あるいは、イベントの発生タイミングの差が含まれる、
     請求項3に記載のトレーニング支援方法。
    Including acquiring the difference between the first operation sequence and the second operation sequence
    The difference includes a difference in the order of occurrence of events and / or a difference in the timing of occurrence of events.
    The training support method according to claim 3.
  5.  前記各イベントには、音が割り当てられており、
     前記第1動作シーケンス、前記第2動作シーケンスを、音響情報に変換可能である、
     請求項3、4いずれか1項に記載のトレーニング支援方法。
    Sounds are assigned to each of the above events.
    The first operation sequence and the second operation sequence can be converted into acoustic information.
    The training support method according to any one of claims 3 and 4.
  6.  第1動作シーケンスと第2動作シーケンスの差を音響情報で学習者に提示することは、
     第1動作シーケンスに対応する第1音響情報を提示すること、
     第2動作シーケンスに対応する第2音響情報を提示すること、を含み、
     第1音響情報と第2音響情報を学習者が対比することで差を認識させることを含む、
     請求項5に記載のトレーニング支援方法。
    Presenting the difference between the first motion sequence and the second motion sequence to the learner with acoustic information is
    Presenting the first acoustic information corresponding to the first motion sequence,
    Including presenting the second acoustic information corresponding to the second motion sequence,
    Including making the learner recognize the difference by comparing the first acoustic information and the second acoustic information.
    The training support method according to claim 5.
  7.  前記差を音響情報に変換する手段を備えており、
     音響情報に変換された差が学習者に提示される、
     請求項1~6いずれか1項に記載のトレーニング支援方法。
    It is equipped with a means for converting the difference into acoustic information.
    The difference converted into acoustic information is presented to the learner,
    The training support method according to any one of claims 1 to 6.
  8.  第1動作シーケンスと第2動作シーケンスの差を視覚的に認識可能に学習者に提示することを含む、
     請求項1~7いずれか1項に記載のトレーニング支援方法。
    Including presenting the learner with a visually recognizable difference between the first and second motion sequences.
    The training support method according to any one of claims 1 to 7.
  9.  身体運動は、ゴルフスイングである、請求項1~8いずれか1項に記載のトレーニング支援方法。 The training support method according to any one of claims 1 to 8, wherein the physical exercise is a golf swing.
  10.  身体運動の第1動作シーケンスを参照として記憶する手段、
     学習者の身体運動の運動データから第2動作シーケンスを取得する手段と、
     第1動作シーケンスと第2動作シーケンスの差を音響情報として学習者に提示する手段と、を備えた
     トレーニング支援装置。
    A means of storing the first motion sequence of physical exercise as a reference,
    A means of acquiring a second motion sequence from the learner's physical motion data,
    A training support device including a means for presenting the difference between the first motion sequence and the second motion sequence to the learner as acoustic information.
  11.  第1動作シーケンスと第2動作シーケンスの差を取得する手段を含み、
     前記提示手段は、取得した差を音響情報として学習者に提示する、
     請求項10に記載のトレーニング支援装置。
    Including means for obtaining the difference between the first operation sequence and the second operation sequence.
    The presenting means presents the acquired difference to the learner as acoustic information.
    The training support device according to claim 10.
  12.  動作シーケンスは、異なる時刻で発生する複数のイベントによって規定され、
     各イベントの発生時刻を取得するイベント発生時刻取得手段を備えている、
     請求項10、11いずれか1項に記載のトレーニング支援装置。
    The operation sequence is defined by multiple events that occur at different times.
    It is equipped with an event occurrence time acquisition means for acquiring the occurrence time of each event.
    The training support device according to any one of claims 10 and 11.
  13.  前記差には、第1動作シーケンスと第2動作シーケンスのイベントの発生順序の差、および/あるいは、イベントの発生タイミングの差が含まれ、
     イベントの発生順序の差、および/あるいは、イベントの発生タイミングの差を取得する手段を含む、
     請求項12に記載のトレーニング支援方法。
    The difference includes a difference in the event occurrence order between the first operation sequence and the second operation sequence, and / or a difference in the event occurrence timing.
    Including means for obtaining the difference in the order of occurrence of events and / or the difference in the timing of event occurrence,
    The training support method according to claim 12.
  14.  前記各イベントには、音が割り当てられており、
     前記第1動作シーケンス、前記第2動作シーケンスを、音響情報に変換する手段を備えている、
     請求項12、13いずれか1項に記載のトレーニング支援方法。
    Sounds are assigned to each of the above events.
    A means for converting the first operation sequence and the second operation sequence into acoustic information is provided.
    The training support method according to any one of claims 12 and 13.
  15.  前記差を音響情報に変換する手段を備えており、
     音響情報に変換された差が学習者に提示される、
     請求項10~14いずれか1項に記載のトレーニング支援方法。
    It is equipped with a means for converting the difference into acoustic information.
    The difference converted into acoustic information is presented to the learner,
    The training support method according to any one of claims 10 to 14.
  16.  コンピュータを、請求項10~15に記載の各手段として機能させるコンピュータプログラム。 A computer program that causes a computer to function as each means according to claims 10 to 15.
PCT/JP2020/021085 2019-05-30 2020-05-28 Training assisting method and device WO2020241738A1 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
JP2019-101841 2019-05-30
JP2019101841A JP2020195431A (en) 2019-05-30 2019-05-30 Training support method and device

Publications (1)

Publication Number Publication Date
WO2020241738A1 true WO2020241738A1 (en) 2020-12-03

Family

ID=73552439

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2020/021085 WO2020241738A1 (en) 2019-05-30 2020-05-28 Training assisting method and device

Country Status (2)

Country Link
JP (1) JP2020195431A (en)
WO (1) WO2020241738A1 (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117859153A (en) * 2021-08-26 2024-04-09 索尼集团公司 Information processing device, information processing method, and program

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4414475B1 (en) * 2009-02-07 2010-02-10 田村 邦博 Downswing instruction device for golf swing. (Exercise teaching device that measures the effect of untwisting in the striking motion with a sensor and assists it with a computer)
JP2011000367A (en) * 2009-06-22 2011-01-06 Yamaha Corp Music reproduction control device
US20170061817A1 (en) * 2015-08-28 2017-03-02 Icuemotion, Llc System for movement skill analysis and skill augmentation and cueing
JP2018050884A (en) * 2016-09-28 2018-04-05 カシオ計算機株式会社 Notification device, notification method and program

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4414475B1 (en) * 2009-02-07 2010-02-10 田村 邦博 Downswing instruction device for golf swing. (Exercise teaching device that measures the effect of untwisting in the striking motion with a sensor and assists it with a computer)
JP2011000367A (en) * 2009-06-22 2011-01-06 Yamaha Corp Music reproduction control device
US20170061817A1 (en) * 2015-08-28 2017-03-02 Icuemotion, Llc System for movement skill analysis and skill augmentation and cueing
JP2018050884A (en) * 2016-09-28 2018-04-05 カシオ計算機株式会社 Notification device, notification method and program

Also Published As

Publication number Publication date
JP2020195431A (en) 2020-12-10

Similar Documents

Publication Publication Date Title
US9656121B2 (en) Methods for analyzing and providing feedback for improved power generation in a golf swing
US9878206B2 (en) Method for interactive training and analysis
JP5641222B2 (en) Arithmetic processing device, motion analysis device, display method and program
CA2819067C (en) Systems and methods for performance training
JP2852925B2 (en) Physical exercise proficiency education system
US9025824B2 (en) Systems and methods for evaluating physical performance
US20110270135A1 (en) Augmented reality for testing and training of human performance
Kelly et al. A virtual coaching environment for improving golf swing technique
JP5867680B2 (en) Exercise instruction device, exercise analysis device, exercise instruction program, and recording medium
JP2007520282A (en) System and method for measuring and evaluating the performance of physical skills and the equipment used to perform this physical skill
JP2009297240A (en) Learning support apparatus and method
Bezodis et al. Contributions of the non-kicking-side arm to rugby place-kicking technique
JP2017029460A (en) Diagnostic server, diagnostic system, diagnostic method, diagnosis program, and recording medium
TW201411507A (en) Method and system for motion comparison
US20220161117A1 (en) Framework for recording and analysis of movement skills
JP2005237494A (en) Actual action analysis system and program
Jakus et al. A system for efficient motor learning using multimodal augmented feedback
JP2017006192A (en) Advice generation method, advice generation program, and advice generation system
JP2016116615A (en) Motion analysis device, motion analysis system, motion analysis method, and program
WO2020241738A1 (en) Training assisting method and device
JP6029369B2 (en) Swing simulation method
KR20220098064A (en) User customized exercise method and system
JP7442258B2 (en) Evaluation method and evaluation system
JP2021058301A (en) Fitting device of golf club
Atack The biomechanics of rugby place kicking

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20813402

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20813402

Country of ref document: EP

Kind code of ref document: A1