International Conference on Artificial Reality and Telexistence
Eurographics Symposium on Virtual Environments (2024)
S. Hasegawa, N. Sakata and V. Sundstedt (Editors)
Influence of Virtual Reality Setup on Locomotion Technique Usage
during Navigation with Walking, Steering and Teleportation
Hugo Brument†1 , Renate Zhang2 and Hannes Kaufmann‡1
1VR & AR Research Unit, TU Wien, Austria
2 TU Wien, Austria
Abstract
The evaluations of Locomotion Techniques (LTs) provide information regarding the advantages and shortcomings of LTs for
navigating in Virtual Reality (VR). While the primary approach is to assess the LTs separately (e.g., comparing walking ver-
sus steering versus teleportation), little is known about how LTs can be used simultaneously (i.e., how users navigate when
several options are offered), especially in different VR setups. This paper aimed to investigate the influence of real and virtual
environment size on LT usage during VR navigation for the first time. We conducted a user study (n=24), where participants
had to explore a virtual garden and pick up mushrooms. Participants could choose to walk, steer, or teleport. We varied the
size of the virtual environment as well as the size of the user’s physical workspace. We found that users’ LT usage depends on
the VR setup. For instance, they tend to do more displacements with teleportation (which was users’ favorite technique overall)
but would rather walk or steer when the size of the virtual environment is the same as the workspace. This work contributes to
understanding user behavior in VR, particularly regarding LT usage, which tends to be an overlooked topic.
CCS Concepts
• Human-centered computing → Virtual reality; User studies;
Figure 1: Left - Participants could freely choose between three lo-
comotion techniques (walking, steering, teleportation) during the
navigation task. Right - User’s view when using the 2:1 Turn tech-
nique: top - the user is requested to perform a half-turn rotation to
gaze at a sphere; bottom - once the gaze is aligned with the sphere,
we ask the user to walk towards to reset the view of the VE.
† hugo.brument@tuwien.ac.at
‡ hannes.kaufmann@tuwien.ac.at
1. Introduction
Navigation is an essential interaction task in Virtual Environ-
ments (VEs) that consists of controlling the virtual position of
the user [LKMP17], involving two main subtasks: traveling, user’s
control of movement through the VE [CGBL98], and wayfinding,
the ability to update the self-position and orientation relative to
the known places in the environments and defining a path through
it [DS96]. This paper only focuses on the traveling part and how the
Locomotion Techniques (LTs) can influence the subcomponents of
traveling such as exploration, search, and maneuvering [BJH99].
We excluded wayfinding, which relies more on a cognitive pro-
cess. The evaluations of Virtual Reality (VR) interfaces can be
complicated because they involve additional factors than traditional
human-computer interfaces, such as the physical environment (e.g.,
size of the workspace, users sitting or standing), multi-modal input
generating a lot of activity logs, or users-related issues (e.g., VR
can cause sickness or fatigue, the user experience with VR can im-
pact the usability and comfort in VR). Besides, the design space
of LT evaluations is very vast, which means that it is unrealistic
to have exhaustive evaluations assessing every factor. This means
that the current and past evaluations of LTs can lead to an over-
generalization (i.e., where claims regarding the properties of LTs
are not further discussed) and lack reproducibility (i.e., making sure
that the observation remains with similar conditions) of results that
compare one or several techniques in a user study [ZW23].
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.
This is an open access article under the terms of the Creative Commons Attribution License, which
permits use, distribution and reproduction in any medium, provided the original work is properly
cited.
DOI: 10.2312/egve.20241377 https://diglib.eg.orghttps://www.eg.org
2 of 10 Brument, Zhang and Kaufmann / Influence of VR Setup on LT Usage during Navigation with Walking, Steering and Teleportation
Bowman et al. proposed one of the most used VR evaluation
methodologies in several publications [BKH97,BKH98,BDHB99],
where they describe the central methodology and required compo-
nents to assess LTs. Many studies focused on comparing the bene-
fits and shortcomings of LTs, particularly regarding walking, steer-
ing, and teleportation [ZMF18]. Those studies showed, in general,
that walking is the most ecological approach for exploring VEs but
is limited to the physical user’s workspace; steering enables contin-
uous trajectories but can lead to some cybersickness; teleportation
enables one to travel faster in larger VE but can lead to some disori-
entation. Most of those evaluations compared each LT separately
using within-subject design (e.g., testing walking, then steering,
then teleportation) and considered only one size of VE and one size
of workspace, which may influence how the user navigates with the
different LTs. Thus, little is known regarding what the users would
do if (1) we allow them to use every technique during a task (i.e.,
having the choice to switch between walking, steering, and tele-
portation) and (2) if we ask them to perform a similar task with
different VR setups.
This paper reports to our knowledge the first VR study compar-
ing simultaneous LTs with different real environments (REs) and
virtual environment sizes. We conducted a user study to investigate
how the size of VEs and REs could influence ’LT usage during
pickup tasks. We introduce the concept of user’s navigation state,
which represents which technique is used; we present a first ex-
ploratory analysis for modeling the user navigation state while per-
forming the task using discrete Markov chains and n-grams. Our
results contribute to understanding user locomotion behavior and
preferences during navigation in VR, and we discuss the design of
LTs, which could improve users’ experience.
2. Related Work
2.1. LTs for Navigating in VR
Walking is accepted to be the most ecological approach to nav-
igate in a VE, as it better matches real locomotion tasks. How-
ever, the limited size of the physical workspace often prevents us
from using it in most VR setups. Several ways exist to manipu-
late users’ rotations and translations to increase walking in the re-
stricted workspace. The main approach consists of increasing the
user’s translation, but they are not the focus of our study; for fur-
ther detail, please refer to [NSSN18]. Another approach is to re-
set the user position (i.e., to replace the user at the center of the
workspace) while trying not to break user immersion. For exam-
ple, [WNR∗07a] designed three overt techniques to fulfill this re-
quirement: freeze-backup, the freeze-turn, and the 2:1 turn. These
three techniques allow the users to infinitely walk in the VE while
being recentered at the center of the workspace when they reach
boundaries while minimizing the number of resets.
In contrast, steering techniques [BJH01] are by far the most
employed in most VR applications and only require a small
workspace. They are mainly characterized by how the user pro-
vides the navigation direction and speed [LKMP17]. The navi-
gation direction is typically provided by a user’s body segment,
such as the head, the hand, or the torso, and is defined by point-
ing or looking toward the desired direction [BOMA20]. Then, the
navigation speed is determined by a control law that defines the
navigation speed depending on the user’s inputs (discrete, e.g., a
button press, or continuous, e.g., a joystick). However, steering
and automatic techniques do not provide proprioceptive or vestibu-
lar information about walking and, therefore, can decrease spatial
awareness [CBCWS18] and the potential increase of motion sick-
ness [SCMW08].
To enable to reach longer distances in the VE faster than walk-
ing and steering, teleportation can be used. The user’s viewpoint
can be updated instantly, with a fixed speed to provide optical
cues [BMF18], or with reorientation mechanisms so that the user
would remain at the center of the workspace [FRK14, LPAZF18],
or even with jumping metaphor [BSB11]. While most teleportation
techniques use a hand-held controller to select the future direction,
other inputs can be considered (e.g., the feet [vWSM∗20]). Tele-
portation is one of the most common LTs for navigating VEs and
has been extensively studied. For further information about telepor-
tation in VR for single users, please refer to [PAF21].
2.2. Evaluations of Walking, Steering, and Teleportation
In the VR community, walking is considered the most ecolog-
ical approach for navigating in VEs since the users’ motion in
the real and the virtual environments are the same, thus provid-
ing an intuitive interface for users. However, research work has
been done to show the benefits of walking versus the use of vir-
tual locomotion techniques. First evaluations showed that walk-
ing increases the presence compared to steering [UAW∗99] thanks
to the proprioceptive and vestibular feedback provided by walk-
ing [SUS95]. Besides, walking induces fewer cybersickness symp-
toms than virtual steering techniques during straight navigation
tasks in VEs [JM01], which was confirmed in a more recent study
comparing walking and teleportation in see-through augmented re-
ality [SSH20]. However, the type of task can influence these re-
sults. [SFR∗09] showed that during short exposure (five minutes)
in a virtual complex 3D maze, walking increased cybersickness
scores more than head steering. Another work indicated that vir-
tual LTs increase spatial working memory demands compared to
walking and that locomotion with a lower FOV increases gen-
eral attentional demands [MKDO13]. Regarding performance in
achieving different tasks in a VE, Ruddle et al. conducted several
experiments regarding the benefits of walking compared to other
techniques. First, walking with complete body-based information
provided better search results than the other techniques [RL06].
They confirmed these results in another experiment where walking
yielded less imperfect search than the head steering or joystick con-
trol with monitor [RL09]. Last, they revealed potential limitations
of their original findings where steering techniques could provide
similar performance results than walking during a search naviga-
tion task [RBM∗10]. Regarding spatial awareness, walking gener-
ates less turning error than other LTs and provides lower distance
estimation [WKMW16].
Steering and teleportation are often compared. In one experi-
ment, participants had to go from one point to another (informa-
tion was given through a map) and collect tokens in a VE [CA17].
They reported that task completion was faster with teleportation
than steering and lower cybersickness scores. Similar results were
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.
Brument, Zhang and Kaufmann / Influence of VR Setup on LT Usage during Navigation with Walking, Steering and Teleportation 3 of 10
found in another study in a point to origin task [WKFK18]. In
terms of presence, they both seem to provide similar immersion,
but teleportation seems more usable than steering in goal-directed
tasks [BC19]. [LLS18] also demonstrated in their experiment that
users prefer teleportation over steering, which also provided higher
cybersickness. A recent study investigated the influence of head-
ing selection and control laws with steering techniques [GMTD21].
Participants had to perform a straight motion, and the results
demonstrated that linear control allow them to travel faster but
with less trajectory precision. On the contrary, the non-linear con-
trol law provided a slower travel timer but better trajectory pre-
cision. Concerning the virtual trajectories achieved with steering
techniques, [COMP13] showed that in the case of navigation with
different LTs (steering, leaning, keyboard), the control of the tra-
jectory was similar, even though continuous interfaces were more
conform with the kinematics of accurate walking trajectories than
virtual interfaces. LT navigation performance can also increase over
time. Nasisir et al. investigated how novice users navigate VR with
steering or teleportation over time [NPKR23]. Their results showed
that performance improved faster with steering than with telepor-
tation, but spatial performances improved faster with teleportation
than steering.
Since a vast number of factors, including the design of the LTs,
the type of task to perform, the metrics to assess the study, and
the human factors can impact users’ behavior while using LTs,
some work focused on taxonomies for describing the relationships
between LTs [PMW22]. While several taxonomies and classifica-
tions have been done to understand better the shared characteris-
tics of similar techniques, some meta-analysis of LTs showed that
some LTs could be identical and clustered to understand better the
similarity and the extensions of results conducted in some stud-
ies [DLSEGF21]. Walking, steering, and teleportation are three dis-
tinct techniques that achieve different purposes and provide dif-
ferent user behaviors regarding performance and spatial awareness
that can be considered to maximize one factor or another based on
the navigation task. However, a recent survey of LTs in commer-
cial VR applications showed that teleportation is still the most used
technique; steering is being more and more adopted by practition-
ers, and while walking is investigated a lot in academia, practition-
ers are not using them due to workspace limitations [ACST24].
3. User Study
Most of the evaluation studies of LTs focus on a between or within-
subject design where each LT is assessed separately. To the best
of our knowledge, no studies allow participants to use every LT
simultaneously to understand why participants would instead use
one technique or another based on the VR setup (i.e., VE and WS
sizes). We know the benefits and drawbacks of each LT (walking,
steering, teleportation), but we do not know whether participants
would take advantage of each technique to perform a given task.
3.1. Locomotion Techniques and Navigation States
Participants could perform the task with three LTs: Walking (in-
cluding a resetting technique), Head-Steering, and Teleportation.
Walking was an isometric mapping between the motion in the RE
and the VE (i.e., walking 1 meter in the RE resulted in walking
1 meter in the VE). To allow participants to perform the task only
with walking if they wanted to, we implemented an overt reset tech-
nique (2:1 Turn) based on [WNR∗07b]. Whenever the user faces
the borders of the tracked area, the environment freezes, and the
user is requested to do a half-turn (Figure 1). The turn is completed
if the player is gazing at a sphere, which is situated in the center of
the room, thus urging the player to make a turn. When the user is
looking at the sphere, the virtual environment unfreezes. It is worth
highlighting that the virtual world did not change in this scenario.
Hence, the user can continue to walk as usual. The player will stop
their natural walking path when colliding against the borders of the
tracked area to perform the necessary turning motion and can pro-
ceed afterward.
Regarding steering, the virtual movement was initiated using the
Meta Quest controller’s joystick, but the head orientation defined
the heading. Even though it has some limitations for searching
tasks, we chose the head as the heading direction to navigate in the
VE as it does not require additional hardware compared to torso
steering and is primarily implemented in VR setups. In addition,
we considered the joystick as a binary input as it is easier for users
to navigate with constant control law than linear [BOMA20]. When
the joystick is pressed, movement is initiated with an acceleration
of 10m/s2 until a maximum tangential speed of 2m/s is reached;
at this point, the speed remains constant. If the joystick is released,
a deceleration of −10m/s2 is applied to reach a speed of 0m/s in
approximately 0.30 seconds.
Teleportation allowed the user to change their position from one
point to another in the VE. By pressing the left controller trigger, a
virtual white ray appears to select the future destination. We added
a visual indication by ray casts that update the valid destinations
for the users to prevent collisions with the VE or going beyond the
virtual fences and leaving the VE when selecting the future virtual
position. The users were instantly moved to the selected VE desti-
nation by releasing the trigger. Users could cancel the teleportation
by pressing a button on the left controller. To have a fair compari-
son between the different LTs while keeping the advantages of each
technique, we set a maximum distance where users could teleport
to 4m (that corresponds to the width of the workspace in our exper-
iment; see subsection 3.2). This still enabled users to, in practice,
travel faster with teleportation than with walking or steering, but
would require to use the teleportation several times to reach the
boundaries of the VEs as in steering or walking.
Since participants could choose which LTs to select during the
task, we also designed an algorithm that detects in which Naviga-
tion State (NS) the user is across time. We defined four different
states: None (NS-None), Walking (NS-Walk), Steering (NS-Steer),
and Teleportation (NS-Teleport). The navigation state was updated
and recorded in every frame. NS-Steer corresponded to whether or
not the controller joystick is pressed. NS-Teleport was set whether
the controller’s trigger was pressed or not, encoding the duration to
choose the next destination. NS-Walking navigation state was de-
fined as a displacement of at least 10cm of the user’s headset over
the window time of 0.1 seconds. NS-None navigation state was set
should any of the above states were not set, resulting in an unmov-
ing user not using any of the LTs.
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.
4 of 10 Brument, Zhang and Kaufmann / Influence of VR Setup on LT Usage during Navigation with Walking, Steering and Teleportation
3.2. Experimental Design and Hypotheses
The navigation task was to search and gather mushrooms in a
garden-like environment while not colliding with the obstacles
present in the VE (Figure 2). We chose this pickup task as it re-
quires exploration of the entire VE to collect the mushrooms, and
such protocols have already been used in similar work [DHLA20].
Two types of mushrooms with different colors and shapes (cylin-
drical brown and spherical red) were present in the VE, and par-
ticipants should only pick up the red one to ensure that they were
paying attention to the task. To investigate the influence of physi-
cal and virtual environments on user navigation behavior, our ex-
periment had a two Workspace Size (WS-Size: WS-Small, WS-
Medium) x 3 Virtual Environment Size (VE-Size: VE-Small, VE-
Medium, VE-Large) within-subject design. WS-Size referred to the
size of the physical workspace. The Small WS-Size (referred after
as WS-Small) was 2 by 2 meters, as it is the minimal room-scale
workspace size required for running SteamVR applications. The
Medium WS-Size (WS-Medium) was 4 by 4 meters, which repre-
sents a recommended WS-Size by Meta and Steam. VE-Size re-
ferred to the size of the VE. We defined the Small VE-Size (VE-
Small) as a room-scale VE (4 by 4 meters), the Medium (VE-
Medium) as 10 by 10 meters (more extensive than WS-Medium),
and the Large (VE-Large) as 20 by 20 meters. The experiment com-
prised 6 blocks corresponding to a combination of WS-Size and
VE-Size. Each block was counterbalanced using a Latin square de-
sign. A trial consisted of navigating the VE for three minutes and
collecting as many brown mushrooms as possible.
Users could navigate in the VE by using any of the techniques
presented in subsection 3.1, which means they could alternate be-
tween those three techniques depending on their own will. Based
on our analysis, we hypothesized that the workspace and virtual en-
vironment size could affect participants’ navigation behavior and,
more precisely: [H1] - The VE-Size will affect the distribution of
usage and distance achieved per LTs. In particular: [H1.1] - the
bigger the VE-Size, the more inactive users will be (i.e., higher
NS-None detection); [H1.2] - the higher the VE-Size, the more
users will teleport; and [H1.3] the smaller the VE-Size, the more
users walk. [H2] - The WS-Size will affect the distribution of usage
and distance achieved per LTs. In particular: [H2.1] users will be
less active in WS-Small than WS-Medium (i.e., higher NS-None
detection); [H2.2] - users will less teleport in WS-Medium than
WS-Small; and [H2.3] users will walk more in WS-Medium than
WS-Small. [H3] - Users will subjectively prefer Teleportation over
Steering over Walking for navigating.
While most of the related work showed the benefits of using each
LT individually between Walking, Steering, and Teleportation, our
motivations are to investigate whether users would use more or less
some LTs if we offered them the freedom of choosing a technique.
Our hypotheses are similar to results already found in previous
studies (section 2), where we would say that users maximize the
advantages and shortcomings of each technique to perform the task
(i.e., traveling faster with teleportation in bigger VE, walking more
if the workspace is bigger...)
3.3. Participants and Apparatus
A total of 24 participants (14 males, 10 females) aged from 22 to
38 years old (M=29.5; SD=4.15) participated in the experiment. 16
participants had only one or no prior experience with VR, whereas
8 participants had regular experience with VR. They signed an in-
formed consent form and were naive to the purpose of the exper-
iment. No ethical approval was required by our institution to con-
duct the study, which conformed with the standards of the Declara-
tion of Helsinki. The headset used for the experiment was a Meta
Quest 1 with its controllers (Figure 1). The experimental platform
was guaranteed to run at the minimum of the headset’s frame rate
(72Hz). The VE was developed with Unity3D (version 2021.3.31f)
and was a square garden with fences to define the boundaries of
the VE (Figure 2). We used a grass texture for the floor and dif-
ferent obstacle elements such as trees, rocks, bushes, and flowers
to prevent linear trajectories. A procedural generation of the VE
was used to prevent learning effects. The mushrooms were put on
a one-meter-high tree stump so users would not have to crouch to
collect them from the ground. To avoid bias regarding the perfor-
mance task (i.e., collect as many mushrooms as possible), the den-
sity of mushrooms depended on the VE-Size. We also procedurally
placed the mushrooms at different locations, and we guaranteed
that the distribution of mushrooms would be equal between VEs,
which corresponded respectively to 4, 24, and 48 mushrooms for
the VE-Small, VE-Medium and VE-Large. The proportion of red
and brown mushrooms was equal (i.e., half red, half brown). When
a participant collected all red or brown mushrooms, we respawned
a new set at different randomized locations with the same density
property. The limits of the user’s workspace were represented by
a transparent black plane (see Figure 2). We turned off the Meta
Quest guardian system to prevent breaks in the user’s presence and
used our resetting technique instead (see subsection 3.1).
3.4. Procedure
Before starting the experiment, participants were briefed on the
study (i.e., instructions and explanations about the task, including
information about rotational gain), provided written consent, and
completed a demographics questionnaire (covering age, gender,
and experience of VR). Subsequently, participants were equipped
with the necessary hardware and engaged in a training session to
become familiar with the three different LTs and the picking mush-
room task. The experiment consisted of 6 trials, where one was a
combination of a WS-size and VE-Size. At the start of the trial,
the users had to calibrate by placing themselves at the center of
the workspace, represented by a small yellow plane on the ground.
Then, they pressed a button on their controller to start the trial, and
the calibration plane disappeared. The users performed the naviga-
tion task by collecting as many mushrooms as possible while mini-
mizing collisions with the VE. After three minutes, the trial ended,
and the user took off the headset and filled out a subjective ques-
tionnaire for this condition, including the following questions: (1)
"Which technique did you prefer the most to navigate in the virtual
environment? Sort the techniques from the most preferred to the
least preferred." "(2) Did you have a particular strategy to navigate
during the tasks? (e.g., using a particular technique for a given sit-
uation, such as the path to collect mushrooms)". After filling out
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.
Brument, Zhang and Kaufmann / Influence of VR Setup on LT Usage during Navigation with Walking, Steering and Teleportation 5 of 10
Figure 2: Left - Empty Top view of the VEs (overlaid). We can notice at the center the two transparent black workspace sizes (only one was
displayed depending on the condition). Middle - Example of a random generation of the VE-Large environment. Right- User’s point of view
in the VE.
the subjective questionnaire, a new WS-Size and VE-Size condi-
tions were set for the subsequent trial. Users were allowed and en-
couraged to take breaks during the experiment between trials. The
experiment lasted approximately 45 minutes, including pauses, ex-
planations, questionnaires, and the experimental protocol.
3.5. Data Processing and Analysis
We recorded data for each trial. Objective data were recorded for
each frame and included the time of the trial, the users’ head po-
sition and orientation in both RE and VE, the users’ navigation
state, the users’ traveled distance in VE and RE, the number of
mushrooms (red and brown) collected and the number of collisions
with the VE. Subjective data were collected after each condition
and included the questions to assess cybersickness, the ranking of
the techniques in terms of preferences, and open questions regard-
ing the strategy to navigate during the condition. We collected 144
trials, (24 users x 3 WS-Size x 2 VE-Size). As every trial lasted
three minutes and was recorded at 72Hz (resulting in 12 960 frames
recorded), we did not have to resample the data to compare them.
We computed the traveled distance over a trial as the sum of the
user’s headset displacement in both VE and RE and the distance
achieved per LT (walking, steering, teleportation). We also com-
puted the percentage of NS detection of each LT as the ratio be-
tween the number of occurrences of an NS during the trial and the
number of frames, then multiplied by 100.
For normally distributed metrics, assessed using the Shapiro-
Wilk test, we analyzed variance (ANOVA) with repeated measures
factors. Greenhouse-Geisser adjustments were applied to the de-
grees of freedom when the sphericity assumption was violated. For
metrics that deviated from a normal distribution, we used the non-
parametric Aligned Rank Transform (ART) test [WFGH11]. The
post-hoc analysis involved pairwise t-tests with Bonferroni correc-
tions for customarily distributed dependent variables or the multi-
factor contrast test procedure presented in [EKHW21] for the non-
normally distributed ones.
4. Results
4.1. Task Performance
Table 1 shows the mean and standard deviations for the objective
metrics related to the task (collisions, number of mushrooms col-
lected, distance achieved in RE and VE). A two way ART-ANOVA
with VE Size and WS Size as within subjects factors showed no
significant effect of VE Size (F2,115 = 1.27, p = 0.28) or WS Size
(F1,115 = 0.36, p = 0.55) on collisions with the VE. However, we
noticed a significant effect of VE Size on number of mushrooms
collected (F2,115 = 113.42, p< 0.001, η2
p = 0.69), where posthoc tests
showed that more mushrooms were collected in the VE-S than the
others. Regarding distance achieved in RE, we noticed a significant
effect of VE Size (F2,115 = 138.47, p< 0.001, η2
p = 0.71) on distance
achieved in RE, where posthoc tests showed that more physical
motion was completed in the VE-S than in the VE-M and VE-
S. Regarding distanced achieved in the VE, we noticed a signifi-
cant effect of VE Size (F2,115 = 143.59, p< 0.001, η2
p = 0.70), where
posthoc tests showed that more virtual motion was achieved in the
VE-L and VE-M than VE-S.
Table 1: Mean and Standard deviation of objective metrics related
to the task grouped by VE Size and WS Size. Main effects are re-
ported with ∗ for VE-Size (no effect was found for WS-Size). Post-
hoc tests are reported using superscripts. Two levels sharing the
same superscript are not significantly different.
Collisions Mushrooms∗ Dist RE (m)∗ Dist VE (m)∗
VE-S 3.35±2.06 48.71±8.491 17.81±4.211 75.39±10.221
VE-M 2.73±2.10 27.02±6.662 10.47±3.472 144.79±24.972
VE-L 3.23±2.07 25.12±5.543 9.16±2.172 139.81±28.352
WS-S 3.04±1.82 32.8±13.35 12.21±2.47 118.18±39.10
WS-M 3.16±2.23 34.40±12.25 12.10±6.90 121.81±38.8
4.2. Usage of techniques
Regarding the percentage of detection of navigation state (Fig-
ure 3), a three-way ART-ANOVA with NS, VE Size and WS
Size as within-subjects factors showed a significant effect of NS
(F3,529 = 862.43, p < 0.001, η2
p = 0.83), where post-hoc showed that
the navigation state distributions followed that following patterns:
%NS-None (48.23±7.72) > %NS-Walk (22.71±8.42) > %NS-
Steer (17.46±5.87) > %NS-Teleport (11.58±4.27). When we con-
sidered the NS individually, we conducted two-way ART-ANOVA
with VE Size and WS Size as within-subjects factors for the detec-
tion of NS. Table 2 summarizes the main effects found and pairwise
comparison: % NS-None - We observed an effect of WS-Size and
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.
6 of 10 Brument, Zhang and Kaufmann / Influence of VR Setup on LT Usage during Navigation with Walking, Steering and Teleportation
VE-Size, where post-hoc showed that NS-None was more detected
in the WS-Small than WS-Medium, and more in VE-Medium and
VE-Large than VE-Small. % NS-Walk - We observed an effect of
WS-Size and VE-Size, where post-hoc showed that NS-Walk was
more detected in the WS-Medium than WS-Small, and more in VE-
Small than VE-Medium and VE-Large. % NS-Steer - We observed
an effect of WS-Size and VE-Size, where post-hoc showed that
NS-Steer was more detected in the WS-Medium than WS-Small,
and more in VE-Large than VE-Medium than VE-Small. % NS-
Teleport - We observed an effect of WS-Size and VE-Size, where
post-hoc showed that NS-Teleport was more detected in the WS-
Small than WS-Medium, but no differences were observed between
VE-Small, VE-Medium and VE-Large.
Figure 3: Bar plots showing the mean and standard deviation of
Detection (in percentage) of NS per NS, WS-Size, and VE-Size.
Regarding the distance achieved per navigation state (Figure 4),
a three-way ART-ANOVA with NS, VE Size and WS Size as
within-subjects factors showed a significant effect of NS (F2,391 =
1645.03, p < 0.001, η2
p = 0.89), where post-hoc showed that users
achieved more distance with Teleportation (141.41±54.25) than
Steering (31.31±9.1) than Walking (12.15±5.24). We also con-
sidered the NS individually and conducted two-way ART-ANOVA
with VE Size and WS Size as within-subjects factors for the dis-
tance achieved per LTs. Table 2 summarizes the main effects found
with pairwise comparison: Dist NS-Walk - We did not observe an
effect of WS-Size, but we observed an effect of VE-Size, where
post-hoc showed that the distance achieved in walking was higher
in VE-Small than VE-Medium and VE-Large. In addition, we ob-
served an interaction effect of VE-Size and WS-Size on distanced
achieved in RE (F2,115 =166.41, p<0.001, η2
p =0.75), where posthoc
tests showed that the highest physical motion was achieved in the
WS-M, VE-S condition (M = 21.54; SD = 2.01). Dist NS-Steer -
We observed an effect of WS-Size and VE-Size, where post-hoc
showed that the distance achieved in steering was higher in the WS-
Small than WS-Medium, and higher in VE-Small and VE-Large
than VE-Medium. Dist NS-Teleport - We did not observe an effect
of WS-Size, but we observed an effect of VE-Size, where post-hoc
showed that the distance achieved in teleportation was higher in
VE-Large than VE-Medium than VE-Small.
4.3. Subjective Questionnaires
Regarding ranking the preferences between the techniques after
each block, Figure 5 shows the ranking distribution for each LT
(either ranked first, second, or third) per WS Size and VE Size.
We can observe that for both WS Size, the distribution of ranking
is similar participants ranked teleportation as the most preferred
technique (46 for both) followed by steering (45 for WS-Small, 53
for WS-Medium) and then walking (48 for WS-Small and 50 for
WS-Medium). Similar observations were found for the VE-Large
(40 first-rank teleportation, 39 second-rank steering, 45 third-rank
walking) and VE-Medium (36 first-rank teleportation, 36 second-
rank steering, 37 third-rank walking). However, for the VE-Small
environment, the distribution was different, where we observed a
similar count for the ranked technique (17 steering, 16 teleporta-
tion, and 15 walking).
5. Discussion
5.1. Influence of VE-Size on Locomotion Technique Usage
In this experiment, we wanted to see whether the same task within
different VE sizes would influence how users navigate. We chose
the pickup mushroom task as it involved exploration, recognition,
and collection of objects which are similar to what users would
do in some VR games (e.g., gathering, gardening, mining). As ob-
served on Figure 3, the most detected navigation state was NS-
None, followed by NS-Walk, NS-Steer, and NS-Teleport. Users
usually spent half of the trial not using the LTs. It does not mean
they were not doing anything. For instance, they could have gath-
ered information in the VE to find the mushrooms and decide the
next destination and LT to choose, or they could have taken some
time to update their spatial awareness. In particular, we noticed that
users were more in NS-None when the VE was bigger (Figure 3).
This makes sense as participants may need to pay more attention to
a larger area where the mushrooms were located before initiating
their next movement. We then confirm our hypothesis [H1.1].
However, while teleporting was the least NS detected (since it
requires less time to achieve distance compared to walking or steer-
ing), we noticed that users were traveling more with teleportation
than the other techniques (Figure 4) and that they were even more
traveling with teleportation with bigger VE-Size (Table 2). We can
perform the task entirely with only one LT if the user wants to.
Still, it seems pragmatically more efficient to use teleportation as
the size of the VE starts to be bigger than the workspace. This
confirms [H1.2], where teleportation has already shown its benefits
over steering and walking when achieving quick travel. Neverthe-
less, it is worth noticing that being fast does not necessarily mean
being more efficient since the number of mushrooms collected was
smaller in VE-Large (Table 1). We suggest that this difference is
due to the spatial disorientation that teleportation may produce, as
already observed in previous work [BC23].
Similarly, we observed that users tended to walk the most in the
VE-Small, where the distance was almost twice as much compared
to VE-Medium and VE-Large (Table 2). We suggest that the room-
scale size of the VE invites users to perform the task as they would
do in real life by naturally moving. This behavior has already been
observed between walking and teleportation [SSH20], which con-
firms [H1.3]. Overall, our work showed that users may navigate
differently depending on the VE-Size, where walking was more
used in VE-Small, steering in VE-Medium, and teleportation in
VE-Large, which confirms [H1].
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.
Brument, Zhang and Kaufmann / Influence of VR Setup on LT Usage during Navigation with Walking, Steering and Teleportation 7 of 10
Figure 4: Bar plots showing the mean and standard deviation of Distance achieved per LTs, WS-Size, and VE-Size.
Table 2: Mean and standard deviation, reported as M±SD, for usage of LTs metrics (% of NS detected and distance achieved by LTs),
grouped by WS-Size and VE-Size. The two effect columns report whether there was a significant effect of WS-Size and VE-Size. Post-hoc
tests for main effects are reported using superscripts. Two levels sharing the same superscript are not significantly different.
WS-Small WS-Medium Effect VE-Small VE-Medium VE-Large Effect
% NS-None 50.90±5.591 45.55±8.622 F1,115 =43.53, p<.001, η
2
p = .27 42.87±8.381 52.29±5.832 49.52±5.392 F2,115 =44.80, p<.001, η
2
p = .46
% NS-Walk 20.11±5.511 25.89±9.942 F1,115 =59.48, p<.001, η
2
p = .34 31.53±7.671 19.33±4.932 17.27±3.362 F2,115 =132.40, p<.001, η
2
p = .69
% NS-Steer 16.50±5.011 18.42±6.512 F1,115 =11.01, p<.01, η
2
p = .01 12.30±2.931 16.68±3.542 23.41±4.463 F2,115 =140.82, p<.001, η
2
p = .71
% NS-Teleport 12.47±4.401 10.70±3.971 F1,115 =9.64, p<.01, η
2
p = .08 12.29±4.561 11.69±3.611 10.78±3.901 F2,115 =12.18, p<.001, η
2
p = .17
Dist NS-Walk 12.21±2.76 12.10±6.90 F1,115 =0.07, p=0.78 17.81±4.201 9.48±3.472 9.16±2.172 F1,115 =138.4, p<.001, η
2
p = .70
Dist NS-Steer 33.77±7.951 28.86±9.382 F1,115 =22.43, p<.001, η
2
p = .16 27.68±6.611 39.44±7.982 26.83±6.151 F2,115 =60.27, p<.001, η
2
p = .51
Dist NS-Teleport 144.77±44.06 138.04±62.96 F1,115 =1.97, p=0.17 108.88±35.661 139.06±39.972 176.29±61.303 F2,115 =25.21, p<.001, η
2
p = .30
Figure 5: User subjective ranking of LTs per WS-Size (top row)
and VE-Size (bottom row).
5.2. Influence of WS-Size on Locomotion Technique Usage
Another factor that we were interested in our study was the
workspace size, which is a major limitation of navigation in VR
setups, as users cannot physically explore larger VEs. We won-
dered whether the users would navigate differently depending on
whether they have more physical workspace. As observed on Ta-
ble 2, participants had less NS-None detection in WS-Medium,
particularly with VE-Small. We suggest that providing a higher
workspace would motivate the users to move more or be less careful
about reaching the limits of the workspace and thus reduce the mo-
ment when they are idle. We thus validate [H2.1], but future work
is required to explore how workspace size and shapes can influence
user engagement in VR. Indeed, those results might not be gener-
alized, as our workspace and VE had the same shapes, which is
unlikely in most VR applications. As such work has already been
done for redirected walking [MHB19], we could imagine similar
studies to understand how we could influence users to use one tech-
nique over another based on the VR setup.
Regarding teleportation usage, we did not observe the differ-
ence in NS-Detection and distance achieved between WS-Small
and WS-Medium (Table 2). Contrary to the results for the VE-Size,
we expected less teleportation by providing a bigger workspace
that would increase walking (which we also did not observe). Still,
we saw higher steering usage in the WS-Small as an alternative to
teleportation. We rejected both [H2.2] and [H2.3], where telepor-
tation and walking behavior were similar across WS-Size. Future
analyses should deepen the interaction influence between WS-Size
and VE-Size as we observed that users walked the most when the
workspace size was more significant and the VE the smallest. One
suggestion would be that a bigger workspace would require fewer
resets for physically exploring the VEs. This may have encouraged
users to move more, explaining a higher NS-Walk detection for
WS-Medium than WS-Small.
5.3. Users Preferences and Navigation Behavior
Figure 5 showed that users preferred overall Teleportation, Steer-
ing, and Walking. This result aligns with the literature, where
people generally prefer virtual techniques as they do not require
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.
8 of 10 Brument, Zhang and Kaufmann / Influence of VR Setup on LT Usage during Navigation with Walking, Steering and Teleportation
physical movement in the workspace [LLS18]. We suggest that
participants are more familiar with Teleportation and Steering as
they are still the main techniques used in VR commercial applica-
tions [ACST24]. However, we noticed that the ranking differed for
the VE-Small condition, where each technique had the same num-
ber of votes for the first ranking. This means that walking could
remain a viable option when the size of the VE matches the size
of the workspace. Yet, most studies comparing LTs often have big-
ger VE to demonstrate the benefits of virtual techniques and not be
limited to room-scale applications. We then partially confirm [H3]
and we sum up hereafter the takeaway messages we noticed from
our user study for navigating with each LT:
Walking seems the most appropriate technique when the
workspace and VE match. Our results showed that walking had the
highest 1st ranking and distance achieved with an LT during the
WS-Medium×VE-Small condition (Figure 4). While we hypothe-
sized that might happen, we must say that this result still surprises
us as we expected users to be lazy and always prefer steering over
walking in case teleportation is irrelevant. In addition, from our ob-
servations, walking seems to be the most appropriate as users get
closer to a mushroom after teleporting or steering, as it seems more
accessible for users, to be more precise. However, further analyses
would be required to demonstrate which technique was most used
depending on the closest mushroom distance. Last, users reported
that they never intended to navigate by repeating the reset technique
while walking and reset only got triggered when people were not
aware of the limits of the VE anymore.
Steering was mainly used as an intermediate between Walking
and Teleportation, especially when the workspace is smaller. For
instance, when teleportation was not suitable (i.e., cannot find a
mushroom or too many obstacles), users would steer to explore
the VE without physical movement. This was reported as a rea-
sonable trade-off between faster movement and walking without
disorientation of teleportation. This is mainly observed in our re-
sults, where the distance achieved with Steering was higher in the
VE-Medium (Figure 4). One reason why steering was less used in
VE-Small might be because of the implementation (head-steering)
that was preventing participants from looking around to find mush-
rooms while navigating. Users’ strategy was reported as using this
technique more after teleportation to have finer control of the tra-
jectory to reach a mushroom. Further analysis would be required to
understand how steering could be more beneficial than walking in
the larger workspace.
Teleportation was, as expected, the technique that enabled faster
travel in larger environments. Even though we limited the maximal
distance of travel per teleportation to 4 meters, users still found this
LT the most suitable for exploring. Users mostly reported that their
strategy was to always consider teleportation as a potential first op-
tion after getting a mushroom if no other mushrooms were within
walking range. This explains why teleportation was less used as the
VE-Size decreased, as users would rather walk or steer (Figure 3).
We suggest that participants felt it was harder to perform the task
in smaller VEs to move precisely enough without disorientation to
get as many mushrooms as possible. Future work could investigate
how to improve the use of teleportation for room-scale use, like
redirected teleportation [LPAZF18].
Regarding the users’ navigation patterns, we often noticed the
following sequence: (1) teleportation to travel further than what the
workspace offers; (2) steering to the closest mushroom; (3) walking
to grab the mushroom. While we believe those sequences and pat-
terns could exist when users interact in VR, additional analysis in
future work would be required to use different LTs simultaneously
patterns using a similar protocol from our study.
6. Limitations and Future Work
First, we are aware that the choice and usage of LTs are likely
task-dependent, preventing our results from being generalized. Fu-
ture works must consider expanding the range of tasks and envi-
ronments beyond the searching and selecting tasks we assessed.
For instance, a higher cognitive load might influence how users
choose the LTs, and additional experiments could investigate it by
considering tasks involving remembering the environment’s lay-
out (e.g., escaping a maze) or using n-back tasks. Second, the
insights regarding the long-term effects of using LTs should be
investigated, too. In our experiment, participants only performed
three minutes per condition, which minimizes cybersickness and
prevents users from getting used to the task. The usage of tech-
niques might change over time based on experience or user com-
fort, and further studies considering the user adaption to LTs would
help to understand how users prefer to navigate over time as well
as how they become familiar with each LT. Last, the range of the
navigation state could be extended. In our experiment, we consid-
ered that a user was doing nothing as long as they were not us-
ing an LT. Future work could focus on improving the navigation
states to model user navigation states better, including user rota-
tion. While there exists a lot of work trying to predict the user’s tra-
jectory while navigating in VR to improve redirection-based tech-
niques [SBL22, MCFH24, ARB22, BBM∗21], there are very few
works tried to understand how and why users would prefer a LT
for different locomotion task (e.g., involving different size of envi-
ronments) to perform in VR [NK12]. Such user behavior modeling
could help VR practitioners understand when such LT is appropri-
ate for users.
7. Conclusion
Understanding user locomotion behavior and preferences has sig-
nificant practical implications for improving LTs in VR. Yet, the
analysis of simultaneous uses of LTs in the literature is limited,
particularly in understanding why and how VR users tend to use or
prefer LTs. Overall, our study showed that users basically use all
techniques that are available to them (as long as they make sense)
and integrate them into their workflow to navigate. This supports
earlier design suggestions to provide users with multiple different
interaction techniques and they will use what fits best depending on
the VR setup. However, this is only a tiny step toward the precise
understanding of navigation patterns between teleportation, steer-
ing, and walking, with limitations such as generalization to more
prolonged VR exposure, consideration of different models to un-
derstand the user’s navigation state and potential human factors
such as learning. This work opens new perspectives on evaluating
LTs and could help propose adapted LTs for the user’s workspace,
VE size, and tasks to perform to improve the user experience.
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.
Brument, Zhang and Kaufmann / Influence of VR Setup on LT Usage during Navigation with Walking, Steering and Teleportation 9 of 10
References
[ACST24] ANDERTON C., CREED C., SARCAR S., THEIL A.: From
teleportation to climbing: A review of locomotion techniques in the most
used commercial virtual reality applications. International Journal of
Human–Computer Interaction (2024), 1–21. 3, 8
[ARB22] ANNAKI I., RAHMOUNE M., BOURHALEB M.: Computa-
tional analysis of human navigation trajectories in the vr magic carpet™
using k-means. In International Conference on Electronic Engineering
and Renewable Energy Systems (2022), Springer, pp. 73–79. 8
[BBM∗21] BRUMENT H., BRUDER G., MARCHAL M., OLIVIER A. H.,
ARGELAGUET F.: Understanding, modeling and simulating unintended
positional drift during repetitive steering navigation tasks in virtual real-
ity. IEEE Transactions on Visualization and Computer Graphics 27, 11
(2021), 4300–4310. 8
[BC19] BUTTUSSI F., CHITTARO L.: Locomotion in place in vir-
tual reality: A comparative evaluation of joystick, teleport, and leaning.
IEEE Transactions on Visualization and Computer Graphics (2019), 1–
1. doi:10.1109/TVCG.2019.2928304. 3
[BC23] BUTTUSSI F., CHITTARO L.: Acquisition and retention of spatial
knowledge through virtual reality experiences: Effects of vr setup and lo-
comotion technique. International Journal of Human-Computer Studies
177 (2023), 103067. 6
[BDHB99] BOWMAN D. A., DAVIS E. T., HODGES L. F., BADRE
A. N.: Maintaining Spatial Orientation During Travel in an Immer-
sive Virtual Environment. Presence: Teleoper. Virtual Environ. 8, 6
(Dec. 1999), 618–631. URL: http://dx.doi.org/10.1162/
105474699566521, doi:10.1162/105474699566521. 2
[BJH99] BOWMAN D. A., JOHNSON D. B., HODGES L. F.: Testbed
evaluation of virtual environment interaction techniques. In Proceed-
ings of the ACM symposium on Virtual reality software and technology
(1999), pp. 26–33. 1
[BJH01] BOWMAN D. A., JOHNSON D. B., HODGES L. F.: Testbed
evaluation of virtual environment interaction techniques. Presence: Tele-
operators and Virtual Environments 10, 1 (2001), 75–95. URL: https:
//doi.org/10.1162/105474601750182333, arXiv:
https://doi.org/10.1162/105474601750182333,
doi:10.1162/105474601750182333. 2
[BKH97] BOWMAN D. A., KOLLER D., HODGES L. F.: Travel in im-
mersive virtual environments: an evaluation of viewpoint motion control
techniques. In Proc. of IEEE Annual International Symposium on Virtual
Reality (1997), pp. 45–52, 215. 2
[BKH98] BOWMAN D. A., KOLLER D., HODGES L. F.: A methodol-
ogy for the evaluation of travel techniques for immersive virtual envi-
ronments. Virtual Reality 3, 2 (June 1998), 120–131. URL: https:
//doi.org/10.1007/BF01417673. 2
[BMF18] BHANDARI J., MACNEILAGE P., FOLMER E.: Teleportation
without spatial disorientation using optical flow cues. In Proc.s of Graph-
ics Interface (2018), pp. 162–167. 2
[BOMA20] BRUMENT H., OLIVIER A.-H., MARCHAL M., ARGE-
LAGUET F.: Does the control law matter? characterization and eval-
uation of control laws for virtual steering navigation. In ICAT-EGVE
2020-International Conference on Artificial Reality and Telexistence &
Eurographics Symposium on Virtual Environments (2020), pp. 1–10. 2,
3
[BSB11] BOLTE B., STEINICKE F., BRUDER G.: The jumper metaphor:
an effective navigation technique for immersive display setups. In Proc.
of Virtual Reality International Conference (2011). 2
[CA17] CHRISTOU C. G., ARISTIDOU P.: Steering versus teleport loco-
motion for head mounted displays. In Proc. of International Conference
on Augmented Reality, Virtual Reality and Computer Graphics (2017),
Springer, pp. 431–446. 2
[CBCWS18] COOMER N., BULLARD S., CLINTON W., WILLIAMS-
SANDERS B.: Evaluating the effects of four vr locomotion
methods: Joystick, arm-cycling, point-tugging, and teleporting. In
Proc. of the 15th ACM Symposium on Applied Perception (2018),
pp. 7:1–7:8. URL: http://doi.acm.org/10.1145/3225153.
3225175, doi:10.1145/3225153.3225175. 2
[CGBL98] CHANCE S. S., GAUNET F., BEALL A. C., LOOMIS
J. M.: Locomotion mode affects the updating of objects en-
countered during travel: The contribution of vestibular and pro-
prioceptive inputs to path integration. Presence 7, 2 (1998),
168–178. URL: https://www.mitpressjournals.org/
doi/10.1162/105474698565659, doi:https://doi.org/
10.1162/105474698565659. 1
[COMP13] CIRIO G., OLIVIER A. H., MARCHAL M., PETTRÉ J.:
Kinematic evaluation of virtual walking trajectories. IEEE Transac-
tions on Visualization and Computer Graphics 19, 4 (2013), 671–680.
doi:10.1109/TVCG.2013.34. 3
[DHLA20] DEWEZ D., HOYET L., LÉCUYER A., ARGELAGUET F.:
Studying the inter-relation between locomotion techniques and embodi-
ment in virtual reality. In 2020 IEEE International Symposium on Mixed
and Augmented Reality (ISMAR) (2020), IEEE, pp. 452–461. 4
[DLSEGF21] DI LUCA M., SEIFI H., EGAN S., GONZALEZ-FRANCO
M.: Locomotion vault: the extra mile in analyzing vr locomotion tech-
niques. In Proceedings of the 2021 CHI Conference on Human Factors
in Computing Systems (2021), pp. 1–10. 3
[DS96] DARKEN R. P., SIBERT J. L.: Navigating large virtual
spaces. International Journal of Human–Computer Interaction 8,
1 (Jan. 1996), 49–71. URL: https://doi.org/10.1080/
10447319609526140, doi:10.1080/10447319609526140.
1
[EKHW21] ELKIN L. A., KAY M., HIGGINS J. J., WOBBROCK J. O.:
An aligned rank transform procedure for multifactor contrast tests. In
The 34th annual ACM symposium on user interface software and tech-
nology (2021), pp. 754–768. 5
[FRK14] FREITAG S., RAUSCH D., KUHLEN T.: Reorientation in virtual
environments using interactive portals. In 2014 IEEE Symposium on 3D
User Interfaces (3DUI) (Mar. 2014), pp. 119–122. doi:10.1109/
3DUI.2014.6798852. 2
[GMTD21] GAO B., MAI Z., TU H., DUH H. B.-L.: Evaluation of
body-centric locomotion with different transfer functions in virtual real-
ity. In 2021 IEEE Virtual Reality and 3D User Interfaces (VR) (2021),
IEEE, pp. 493–500. 3
[JM01] JAEGER B. K., MOURANT R. R.: Comparison of
Simulator Sickness Using Static and Dynamic Walking Sim-
ulators. Proc. of the Human Factors and Ergonomics Soci-
ety Annual Meeting 45, 27 (Oct. 2001), 1896–1900. URL:
https://doi.org/10.1177/154193120104502709,
doi:10.1177/154193120104502709. 2
[LKMP17] LAVIOLA J. J., KRUIJFF E., MCMAHAN RYAN P. BOW-
MAN D. A., POUPYREV I.: 3D User Interfaces: Theory and Practice.
Addison Wesley Longman Publishing Co., Inc., 2017. 1, 2
[LLS18] LANGBEHN E., LUBOS P., STEINICKE F.: Evaluation of lo-
comotion techniques for room-scale vr: Joystick, teleportation, and redi-
rected walking. In Proc. of the Virtual Reality International Conference-
Laval Virtual (2018), ACM, p. 4. 3, 8
[LPAZF18] LIU J., PAREKH H., AL-ZAYER M., FOLMER E.: Increas-
ing walking in vr using redirected teleportation. In Proceedings of the
31st annual ACM symposium on user interface software and technology
(2018), pp. 521–529. 2, 8
[MCFH24] MAYOR J., CALLEJA P., FUENTES-HURTADO F.: Long
short-term memory prediction of user’s locomotion in virtual reality. Vir-
tual Reality 28, 1 (2024), 65. 8
[MHB19] MESSINGER J., HODGSON E., BACHMANN E. R.: Effects of
tracking area shape and size on artificial potential field redirected walk-
ing. In 2019 IEEE Conference on Virtual Reality and 3D User Interfaces
(VR) (2019), IEEE, pp. 72–80. 7
[MKDO13] MARSH W. E., KELLY J. W., DARK V. J., OLIVER J. H.:
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.
10 of 10 Brument, Zhang and Kaufmann / Influence of VR Setup on LT Usage during Navigation with Walking, Steering and Teleportation
Cognitive Demands of Semi-Natural Virtual Locomotion. Presence:
Teleoperators and Virtual Environments 22, 3 (Aug. 2013), 216–234.
URL: https://doi.org/10.1162/PRES_a_00152, doi:10.
1162/PRES_a_00152. 2
[NK12] NESCHER T., KUNZ A.: Analysis of short term path prediction
of human locomotion for augmented and virtual reality applications. In
2012 International Conference on Cyberworlds (2012), IEEE, pp. 15–
22. 8
[NPKR23] NASIRI M., PORTER J., KOHM K., ROBB A.: Changes in
navigation over time: A comparison of teleportation and joystick-based
locomotion. ACM Transactions on Applied Perception 20, 4 (2023), 1–
16. 3
[NSSN18] NILSSON N. C., SERAFIN S., STEINICKE F., NORDAHL R.:
Natural Walking in Virtual Reality: A Review. Comput. Entertain. 16, 2
(Apr. 2018), 8:1–8:22. 2
[PAF21] PRITHUL A., ADHANOM I. B., FOLMER E.: Teleportation
in virtual reality; a mini-review. Frontiers in Virtual Reality 2 (2021),
730792. 2
[PMW22] PRINZ L. M., MATHEW T., WEYERS B.: A systematic litera-
ture review of virtual reality locomotion taxonomies. IEEE transactions
on visualization and computer graphics (2022). 3
[RBM∗10] RIECKE B. E., BODENHEIMER B., MCNAMARA T. P.,
WILLIAMS B., PENG P., FEUEREISSEN D.: Do we need to walk for
effective virtual reality navigation? physical rotations alone may suf-
fice. In International Conference on Spatial Cognition (2010), Springer,
pp. 234–247. 2
[RL06] RUDDLE R. A., LESSELS S.: For efficient navigational search,
humans require full physical movement, but not a rich visual scene. Psy-
chological Science 17, 6 (2006), 460–465. 2
[RL09] RUDDLE R. A., LESSELS S.: The Benefits of Using a Walk-
ing Interface to Navigate Virtual Environments. ACM Trans. Comput.-
Hum. Interact. 16, 1 (Apr. 2009), 5:1–5:18. URL: http://doi.acm.
org/10.1145/1502800.1502805, doi:10.1145/1502800.
1502805. 2
[SBL22] STEIN N., BREMER G., LAPPE M.: Eye tracking-based lstm
for locomotion prediction in vr. In 2022 IEEE Conference on Virtual
Reality and 3D User Interfaces (VR) (2022), IEEE, pp. 493–503. 8
[SCMW08] SHARPLES S., COBB S., MOODY A., WILSON J. R.: Vir-
tual reality induced symptoms and effects (VRISE): Comparison of head
mounted display (HMD), desktop and projection display systems. Dis-
plays 29, 2 (2008), 58–69. URL: https://www.sciencedirect.
com/science/article/pii/S014193820700100X, doi:
https://doi.org/10.1016/j.displa.2007.09.005. 2
[SFR∗09] SUMA E. A., FINKELSTEIN S. L., REID M., ULINSKI A.,
HODGES L. F.: Real Walking Increases Simulator Sickness in Navi-
gationally Complex Virtual Environments. In 2009 IEEE Virtual Real-
ity Conference (Mar. 2009), pp. 245–246. doi:10.1109/VR.2009.
4811037. 2
[SSH20] SAYYAD E., SRA M., HÖLLERER T.: Walking and telepor-
tation in wide-area virtual reality experiences. In 2020 IEEE inter-
national symposium on mixed and augmented reality (ISMAR) (2020),
IEEE, pp. 608–617. 2, 6
[SUS95] SLATER M., USOH M., STEED A.: Taking steps: the influence
of a walking technique on presence in virtual reality. ACM Transactions
on Computer-Human Interaction (TOCHI) 2, 3 (1995), 201–219. 2
[UAW∗99] USOH M., ARTHUR K., WHITTON M. C., BASTOS R.,
STEED A., SLATER M., BROOKS JR. F. P.: Walking > walking-in-
place > flying, in virtual environments. In Proc. of the 26th Annual
Conference on Computer Graphics and Interactive Techniques (1999),
pp. 359–364. 2
[vWSM∗20] VON WILLICH J., SCHMITZ M., MÜLLER F., SCHMITT
D., MÜHLHÄUSER M.: Podoportation: Foot-based locomotion in virtual
reality. In Proceedings of the 2020 CHI Conference on Human Factors
in Computing Systems (2020), pp. 1–14. 2
[WFGH11] WOBBROCK J. O., FINDLATER L., GERGLE D., HIGGINS
J. J.: The aligned rank transform for nonparametric factorial analyses
using only anova procedures. In Proc. of the ACM SIGCHI conference
on human factors in computing systems (2011), pp. 143–146. 5
[WKFK18] WEISSKER T., KUNERT A., FRÖHLICH B., KULIK A.: Spa-
tial Updating and Simulator Sickness During Steering and Jumping in
Immersive Virtual Environments. In 2018 IEEE Conference on Vir-
tual Reality and 3D User Interfaces (VR) (Mar. 2018), pp. 97–104.
doi:10.1109/VR.2018.8446620. 3
[WKMW16] WILSON P. T., KALESCKY W., MACLAUGHLIN A.,
WILLIAMS B.: Vr locomotion: Walking > walking in place
> arm swinging. In Proc. of the 15th ACM SIGGRAPH Con-
ference on Virtual-Reality Continuum and Its Applications in In-
dustry - Volume 1 (New York, NY, USA, 2016), VRCAI ’16,
ACM, pp. 243–249. URL: http://doi.acm.org/10.1145/
3013971.3014010, doi:10.1145/3013971.3014010. 2
[WNR∗07a] WILLIAMS B., NARASIMHAM G., RUMP B., MCNAMARA
T. P., CARR T. H., RIESER J., BODENHEIMER B.: Exploring Large
Virtual Environments with an HMD when Physical Space is Limited.
In Proc. of the 4th Symposium on Applied Perception in Graphics and
Visualization (2007), pp. 41–48. 2
[WNR∗07b] WILLIAMS B., NARASIMHAM G., RUMP B., MCNA-
MARA T. P., CARR T. H., RIESER J., BODENHEIMER B.: Exploring
large virtual environments with an hmd when physical space is limited.
In Proceedings of the 4th symposium on Applied perception in graphics
and visualization (2007), pp. 41–48. 3
[ZMF18] ZAYER M. A., MACNEILAGE P., FOLMER E.: Virtual loco-
motion: a survey. IEEE Transactions on Visualization and Computer
Graphics (2018), 1–1. 2
[ZW23] ZIELASKO D., WEISSKER T.: Stay vigilant: The threat of a
replication crisis in vr locomotion research. In Proceedings of the 29th
ACM Symposium on Virtual Reality Software and Technology (2023),
pp. 1–10. 1
© 2024 The Authors.
Proceedings published by Eurographics - The European Association for Computer Graphics.