The authors developed a simulation VR system for kyudo beginners to train the shooting form. Before kyudo beginners can practice with a (real) bow, they train the shooting form empty-handed and with a rubber practice bow. The system is expected to be used during this period. The user wears the HMD and practices in the training hall in the VR space empty-handed or with a rubber bow. The user's shooting form is measured, and when the form is appropriate (compared to the kyudo expert's form equipped in the system), the arrow hits the target in the VR space. Whether the arrow hits the target depends only on the user's forms, which reflects seisha hitchu (“true shooting, certain hitting”), i.e., the importance of the shooting form. In addition, when the arrow hits the target, the user feels a sense of fulfillment and is motivated to continue practicing kyudo.
When you first start kyudo (Japanese archery), you can’t shoot an arrow with a (real) bow. You must learn the shooting form, how to control your body for shooting, and so on. Therefore, you must train the shooting form empty-handed and with a rubber practice bow for six months to a year. During this training, you practice without arrows and a (real) bow. You may find it hard to continue practicing kyudo.
In this paper, the authors study the training of the shooting form in the period described above, referring to the previous works on practicing kyudo shooting form. The authors also focus on motivating the user in this period. The authors expect the user to be motivated when his/her appropriate form make the arrow hit the target. This reflects a famous kyudo phrase seisha hitchu (“true shooting, certain hitting”). Though the user doesn’t equip an arrow or a bow, s/he can make the arrow hit the target in the VR space. This idea will let the user know the importance of the shooting form and motivate him/her (moreover, this concept is a kind of gamification).
0000-0003-3972-4332 (K. Ushida) © 2023 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
Many studies on sports, including budo, deal with the practice and evaluation of the form. As for kyudo, there are also studies on the form. In [1], an evaluation of the user’s shooting form with motion capture using a Kinect sensor is developed. In [2], the advice for improving the form is superimposed on the user’s image in the display calculated from his/her motion data. In [3], the evaluation and the advice to improve the form are provided from the measured form in real time. In [4], the user examines his/her form by superimposing and comparing the model form and his/her form with an HMD. In [5], the user learns the shooting form by observing the images of the model form and adapting his/her form to the model. In [6], a self-training system of kyudo shooting posture using a computer vision method is proposed. In [7], a training method is proposed in which the user makes a sketch doll pose in the shooting form and poses by him/herself as making the doll pose.
VR-based kyudo shooting form training systems
have also been developed. In [8], a VR-based shooting
form training system using multi-modal information is
proposed. In [
In addition to measuring and evaluating the shooting form, the remarkable feature of the authors’ system is that the appropriate form makes the arrow hit the target in VR space. Hitting the target with the arrow is the real charm of kyudo. This feeling will lead to a sense of fulfillment and motivation to practice kyudo.
Those who train the shooting form empty-handed and with rubber bows before they are permitted to equip real bows are the expected users of the developed system. They have learned the overview of the shooting form.
The user wears an HMD, stands against the sensor, and trains the shooting form empty-handed or with a rubber bow. There is a training hall and a target in the VR space displayed in the HMD. At the top right of the view, the user’s skeleton measured by the sensor is shown (Figure 1).
The user trains following the eight-step movement of the shooting form (shaho hassetsu). After the user completes the setup for the shooting (ashibumi, dozukuri, and yugamae), the measurement starts. Thus, the last five steps (uchiokoshi, hikiwake, kai, hanare, and zanshin) are measured and evaluated. As the movement progresses, the evaluation indicators (at the top left of the view) turn on. The green lamp represents qualified, and the yellow lamp represents unqualified.
After the shooting motion (hanare and zanshin) is completed, the arrow in the VR space is shot. The arrow hits the target when the total score of the four movements (hanare and zanshin are evaluated into one) is higher than the qualifying score. Whether the arrow hits the target is closed-up in the view and indicated by “○” (hit) or “×” (miss) (Figure 2). Thus, whether the arrow hits the target depends on the appropriateness of the shooting form. Therefore, when the user qualifies for the current goal, i.e., his/her shooting form is appropriate, s/he can make the arrow hit the target in the VR space.
Finally, the detailed evaluation and the comments are displayed. The user can restart the training by pressing the button on the VR controller.
The authors developed the system using Unity on Windows. The HMD is Oculus Rift S, and the sensor is Microsoft Kinect v2.
The kyudo shooting form consists of eight steps (shaho hassetsu): ashibumi, dozukuri, yugamae, uchiokoshi, hikiwake, kai, hanare, and zanshin. In the developed system, the user trains mainly uchiokoshi (raising the bow), hikiwake (drawing apart), kai (the full draw), hanare (the release), and zanshin (the remaining body or mind). In the system, hanare and zanshin are lumped together because the time of hanare is very short.
The system displays the training hall to the user, as shown in Figure 1. In addition, in the training, the skeleton of the user measured by the sensor (at the top right of the view) and the evaluation indicator (at the top left of the view) are displayed. The four indicators represent the evaluation of uchiokoshi, hikiwake, kai, and hanare and zanshin, respectively. The user can see the position of the hands and the arrow which follows the position of the hands.
The user presses the button on the VR controller to start the training (the controller should be within reach). The user places the footing (ashibumi), forms the body (dozukuri), and readies the bow (yugamae). These setup steps are out of evaluation.
When the user starts uchiokoshi, the system begins to measure and evaluate the shooting form. The user’s shooting form is compared with the prerecorded model form performed by a 6-dan kyudo expert (Figure 3). The evaluation points are: the angle between the torso and the right upper arm from the side view, the angle between the torso and the left upper arm from the side view, the angle of the right elbow, the angle of the left elbow, the angle between the torso and the armpit, the angle between the back and the chest, the positions of both hands, the positions of the shoulders, and the inclination of the body. Most items are angles because they don’t depend on the user’s physique. The score is higher when the difference between the measured value of the user and that of the model is slighter.
The scores of each step (uchiokoshi, hikiwake, kai, and hanare and zanshin) are indicated by the lamps at the top left of the view. The green lamp represents qualified and the yellow lamp represents unqualified.
After hanare and zanshin motion, the arrow in the VR space is shot. When the total score of the four steps is higher than the qualifying score, the arrow hits the target. (when the total score is lower, the arrow doesn’t hit the target). The target is closed-up at the top right of the view, as shown in Figure 2. Whether the arrow hits the target is indicated by “○” (hit) or “×” (miss). In Figure (a) Uchiokoshi (b) Hikiwake
(c) Kai (d) Hanare and zanshin 2, though uchiokoshi, hikiwake, and hanare and zanshin are unqualified, the arrow hits the target because the total score is qualified.
Figure 4 shows a user training with the system and his view. The cyan line is the arrow and the hand positions are displayed near it. As shooting progresses, the indicators at the top left turn on. Finally, his form made the arrow hit the target.
A test drive was carried out by one of the authors who was experienced in kyudo. Figure 4 shows his training. Those who have mastered the kyudo shooting form can make the arrow hit the target because their shooting forms are appropriate.
The authors carried out a usability study. The authors checked that the system was easy for everyone to use before it was extensively tested by kyudo beginners and experienced kyudo players.
Ten participants aged 18–20 tried the kyudo training using the system. All of them had no kyudo experience, and they tried the training after the authors showed and explained the form. The authors observed their trials and received comments from them
All of them were able to train without problems. However, none of them were able to make the arrow hit the target. Their inability was inevitable because they had to acquire the shooting form to a sufficient level to make the arrow hit the target. Nevertheless, their smooth training indicated that the system’s usability was appropriate.
Two of them were familiar with VR experiences and gave detailed comments: The feedback (displayed hand positions and marking indicators) was useful for smooth training. It was better to display the model form (this indicates that integration of a presentation method like [4] will be useful for the users at a lower level than expected).
No usability problems were found. The next step is to evaluate the system by kyudo beginners and experienced kyudo players.
In this paper, the authors reported a VR simulation system for training the kyudo shooting form. The appropriate form makes the arrow hit the target in the VR space, which reflects seisha hitchu. The real charm of making the arrow hit the target will motivate the users who haven’t been permitted to equip the real bow.
Currently, the usability of the system has been tested. After this, improving the training quality and the practicality of the actual training will be studied. The former is to improve the method of measurement and evaluation referring to the related works. The latter is the trial and evaluation by kyudo beginners and experienced kyudo players. The effectiveness will also be studied in the long term.
Furthermore, considering the study in Section 5.2,
the system can be used as VR content to introduce kyudo
(such as [