This study proposed a wrist-worn device that displays 3-dimensional (3D) environmental airflow information during Virtual Reality (VR) experience by controlling the rotation of air outlets using servo motors and providing physical airflow using air pumps. Instead of air resistance as the consequence of strong airflow, we focused on reflecting the natural existence of the airflow itself, while maintaining the advantages of wearable devices including being lightweight and convenient to set up. We designed VR scenes to record users' perception accuracy of airflow in diferent directions and their level of confidence, and evaluated through a within-subject experiment. Based on the results, we discussed the methods for future advancement of haptic perception as well as possible VR application scenarios.
To address the lack of real-time haptic perception of Due to limited configurations of current commercial VR
VR experience, researchers have presented multiple ap- devices, they cannot provide real-time and lifelike haptic
proaches for simulating wind using air jets and pro- feedback along with the change of visual scene. Thereby,
pellers. Since the airflow is generally produced during approaches have been investigated to physically augment
the player’s two-degree-of-freedom (2-DoF) movements users’ feeling of motion and wind [
In this study, we aim to convey the information of [
By conducting a within-subject user experiment, we far, these studies have primarily focused on actuating the
evaluated if users were able to understand the wind direc- head to address the lack of tactile perception during the
tion correctly and confidently through the haptic device. high-speed activities in the virtual world, and
successfully promote the immersion of VR systems. However,
lots of people experienced discomfort such as dizziness
and nausea when using VR headsets due to receiving
conflicting visual and body movement information [
Considering that environmental airflow dynamics are
largely independent of human activities, the simulation
of environmental airflow on a HMD while participants
engage in autonomous movement may result in
inconsistencies between visual stimuli and head-driven haptic
stimuli. This condition could similarly cause discomfort
of information transfer, which increases the possibility of
demonstrating 3D surrounding environment by
airflowbased feedback. Hence, our goal is to retain the benefit
of wearable design while achieving the presentation of
3D airflow patterns for VR users.
3. System Implementation
3.1. Wearable Haptic Device
the air propulsion has been achieved through air jets [
To make sure that airflow from 3-DoF directions can make
contact with the skin surface, the haptic device has to
be worn in a body location that is not limited to a single
2D plane. Among the body parts that have both dorsal
and ventral side, the wrist has the highest sensitivity for
airflow recognition [
we implemented two VR scenes for 3D data recording: 1) Practice Scene, in which users can feel free to experience the feeling of airflow in diferent directions; 2) Test Scene, in which users indicate the airflow direction they perceive and answer their level of confidence.
The Practice Scene has 14 "Press me!" buttons posi- Figure 3: Bar chart of participants’ accuracy in perception tioned in evenly distributed 14 directions originating tests, and line chart of their considered dificulty of perception from the VR headset (Fig. 1 (b)). 6 of them are the posi- on a 5-point Likert scale, in terms of three dimensions. tive and negative directions of the X, Y, and Z axes (i.e. +X, -X, +Y, -Y, +Z, -Z), while the other 8 correspond to the diagonal directions of the 8 quadrants formed by 4.1. Study Procedure these axes. When a "Press me!" button is pressed, the VR system decides the status of air pumps and the rotation Participants sat on a non-swivel chair in a quiet and angles of servo motors that make the air outlets pointing closed area. Participants wore our haptic device on the to the corresponding position on the wrist skin originat- wrist of their non-dominant hand, and placed their hand ing from the haptic device, then send data to the haptic and elbow on 4.5cm stands on the table with palm facing device via communication with Arduino. inward. Participants then put on a Meta Quest Pro head
Following the same mechanism, the Test scene ran- set and a noise-canceling earphone with the researcher’s domly selects one of the 14 directions as a test question help, and hold a VR controller in the dominant hand. and controls the Arduino to present the corresponding Participants first tried "Press me!" buttons in the Pracairflow, with participants completing 28 tests in which tice Scene without a time limit, and then proceeded to the each direction occurs twice. There are 14 black buttons Test Scene. For each test, participants selected a black butin the same position as the 14 "Press me!" buttons in Prac- ton to indicate the perceived airflow direction, selected tice Scene, which means that the user’s answers are also confidence level and continued to the next test. After among the 14 directions. After the VR system receives completing 28 tests, participants filled out a questionthe answer, 5 button choices from "Very Unconfident" to naire about haptic perception experience, with responses "Very Unconfident" will be shown on the scene to require recorded anonymously. users’ level of confidence.
4. User Study We recruited 8 participants, 4 males and 4 females with age 22-28 (mean = 25.3, SD = 2.1). All participants were To evaluate users’ ability to understand the direction of right-handed and three had prior experience with tactile 3D airflow by wearing the haptic device, we conducted a experiments. 2 participants are familiar with VR. Each within-subject study on perception accuracy and level of participant received 1,500 yen as an honorarium. confidence. 4.3. Results changed from spherical rotation to 2D movement to create a clearer positional distinction between the front and back ends. To strengthen the intensity of wind perception, the air pumps could be replaced with more powerful alternatives. Regular short pauses could be implemented during the continuous display of airflow to facilitate users’ comprehension. A quantitative evaluation could be carried out to assess the precision of haptic display. Additionally, to further investigate the efect of wind contact on various skin areas during VR experience, the device could be worn on multiple diferent body locations other than the wrist.
Fig. 2 shows the distribution of airflow perception errors and participants’ level of confidence regarding each airflow direction. Participants (P2/5/6) explained that changes in the hit point on their skin helped them determine wind directions. P6 noted that her direction recognition also relied on the skin area where she felt particularly cool. P4/7/8 judged directions referring to the structure of the haptic device and movement of the air outlet around the wrist. Users’ average level of confidence is shown to be similar when deciding diferent airflow directions (mean = 3.32, SD = 0.27). P5 mentioned that her conifdence increased when airflow changed along a single dimension, e.g. Front-Right-Up to Back-Right-Up.
The accuracy and considered dificulty of airflow perception in three dimensions are shown in Fig. 3. Participants demonstrated the highest accuracy and the lowest dificulty in answering Left/Right direction among the three dimensions. In contrast, we observed their accuracy in determining Front/Back direction to be much lower compared to the other two dimensions, but participants considered figuring out Up/Down direction to be as dificult as figuring out Front/Back direction. In more detail, P3 described that she first determined the Left/Right side, then a more specific position (i.e. Up/Down and Front/Back on the Left/Right side). Moreover, P1 sometimes felt the airflow was too weak, and P6 suggested that increasing wind strength might make it easier to recognize the wind direction. P3/4 believed they could better understand airflow direction if it were intermittently displayed.
Compared to placing fans in physical rooms or equipping them with HMD, our wearable device could be more compact and convenient for experiencing the sensation of wind in VR. Our design, which supports continuous 3D reflection of the surrounding, could be particularly suitable for scenarios involving 3-DoF movements at special locations, e.g. high altitudes, forests and caves. In such situations, users are exposed to a 3D space encountering airflow from various dimensions, hence airflow plays an important role in users’ environmental cognition. Through future development, we aim to provide more precise directional information to help users better understand environmental changes in such contexts.
As a hands-free haptic device, we propose wearing it
while engaging in games like battle games that require
holding guns and knives, or applications like flight
simulators that require mastering steering wheels or control
5. Discussion joysticks. Based on these possibilities, we will examine
the device’s usability within such sample scenarios that
5.1. Improvement of Airflow Perception involve realistic body movements. Simultaneously, the
influence of visual content in HMD on users’ recognition
According to the feedback, we analyzed the factors influ- of airflow direction will be investigated.
encing user recognition of airflow direction. Considering Our device could be used together with plenty of haptic
that the accuracy of airflow perception from Left/Right, equipment worn in diferent body areas, such as on the
Up/Down and Front/Back directions decreases in turn, hands [
To address current user confusion, we will implement will explore various methods to improve the accuracy of various measures to improve the accuracy of 3D direc- airflow recognition, and apply this haptic design to VR tion recognition. The movement of air outlets could be scenarios with fluctuating wind conditions.