This repo contains the essential Unity code scripts used to implement the HEAD-BOB, ARM-SWING, LEG-LIFT and FULL-BODY walking-in-place (WIP) methods used in the paper on "Understanding User Experiences Across VR Walking-in-place Locomotion Methods".
A quick overview of the essential details of the paper can be see at the pre-recorded CHI'22 presentation below.
A preview of the experience can also be seen in the following video.
chi22-157-figure-480p.mov
The paper can be accessed at ACM DL and you can reference us in LaTeX with the following bibtex:
@inproceedings{10.1145/3491102.3501975,
author = {Tan, Chek Tien and Foo, Leon Cewei and Yeo, Adriel and Lee, Jeannie Su Ann and Wan, Edmund and Kok, Xiao-Feng Kenan and Rajendran, Megani},
title = {Understanding User Experiences Across VR Walking-in-Place Locomotion Methods},
year = {2022},
isbn = {9781450391573},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3491102.3501975},
doi = {10.1145/3491102.3501975},
booktitle = {CHI Conference on Human Factors in Computing Systems},
articleno = {517},
numpages = {13},
keywords = {Virtual Reality, Immersion, Walking-In-Place, Locomotion},
location = {New Orleans, LA, USA},
series = {CHI '22}
}
Or copy the formatted text into your document of choice:
Chek Tien Tan, Leon Cewei Foo, Adriel Yeo, Jeannie Su Ann Lee, Edmund Wan, Xiao-Feng Kenan Kok, and Megani Rajendran. 2022. Understanding User Experiences Across VR Walking-in-place Locomotion Methods. In CHI Conference on Human Factors in Computing Systems (CHI '22). Association for Computing Machinery, New York, NY, USA, Article 517, 1–13. https://doi.org/10.1145/3491102.3501975
Essential details (extracted from the paper) for each algorithm are given in the subsections below. Please read the paper for more details on the science behind these implementations, as well as all the associated references that these implementations were built upon.
Implementation details can be found within the code (and comments) itself, and feel free to post any issues you find in this repo.
Position and orientation of the VR HMD (a HTC Vive Pro) are tracked using its embedded accelerometer and gyroscope sensor. The vertical translations of the HMD translate to forward speeds to move the user forward in virtual space, while the orientation of the HMD translate to the virtual movement direction. The orientation also determines where the user will be facing which means that the user has to look in the direction of intended movement. This implementation is based on the assumption that users will naturally bob their bodies (and hence their heads) when they walk and will do the same when they perform WIP.
The swinging action of the arms are tracked using a consumer-friendly setup that includes beacons (SteamVR Base Stations) and IMU-based tracking devices (HTC Vive Trackers) strapped to the arms. Movement direction is determined from the same motion sensors, as they are strapped in a fixed orientation relative to the user’s intended movement direction. This implementation is based on the assumption that users swing their arms during WIP, similar to how they would in real-life walking.
The lifting action of the legs are tracked using the same tracking setup but with the trackers now strapped to the ankles. Movement direction is deter- mined from an additional tracker on the waist as the leg sensors will not be able to reliably maintain a fixed orientation relative to the user’s intended movement direction. This implementation is based on the assumption that users lift their legs during WIP, similar to how they would in real-life walking.
This is a combination of the Head-bob, Arm-swing and Leg-lift methods, to investigate any experiential differences provided by a sensor fusion approach. For example, one possible advantage was for users to freely look around (versus in pure Head-bob where they have to look at where they are going) while still incorporating the bobbing action tracked by the HMD for translation into walking velocities. This method collects tracked positional and orientation data from the various sensors used in all the previous methods and translate the combined averaged values to virtual speeds and directions in the virtual environment. In general, this provides another source of comparison to solicit experiential feedback from participants.
For all the four implementations, the HTC Vive Pro HMD and Vive Trackers were used in a SteamVR Tracking setup within two SteamVR Base Stations that demarcated an 4m by 4m VR play- space for participants. The Vive Trackers were paired with different straps for the arms, ankles and waist as needed for each setup.
The study in the paper focuses on WIP setups that only require tracking components that are easily accessible to end consumers, and hence can be reasonably applied to designs using similar off-the-shelf tracking mechanisms. For this study, a SteamVR Tracking setup was used, which combines both tracking from internal IMUs in the HTC Vive Trackers and external tracking with beacons to determine rotation and position data to enable WIP. Alternative systems using similar forms of tracking such as the StonX system will likely produce similar results to the studies presented here. Systems that use only internal IMUs for tracking, e.g., SlimeVR, may also infer some insights from this study, but in a more cautious manner due to technical differences. For example, the lower tracking precision may not provide the same extent of affordances mentioned in Section 4.3.
The code in the unity-locomotion-scripts folder are Unity C# component scripts that developers can use in a Unity3D project.
There is also a sample-unity-project that uses these scripts in a "Hello World"-ish interactive VR environment.
The code can also be treated as a source of information for the implementation details of the various WIP methods discussed in the paper.