
Conference acronym ’XX, June 03–05, 2018, Woodstock, NY Han et al.
(1)
We introduce PORTAL, designed for the user to select and
manipulate distant objects with no oset and a better depth
perception of remote locations.
(2)
We conduct two user studies and report our ndings to
evaluate the eciency and usability of PORTAL.
(3)
We provide a thorough discussion on our ndings, the po-
tential and limitations of PORTAL.
2 RELATED WORKS
2.1 Remote Object Interaction in IVE
To interact with objects in IVEs, the control and motor spaces
should be considered [
2
,
23
,
30
]. The control space refers to the
spatial range in which the user is aordable to control objects,
while the motor space is the physical space available for the user
to operate the objects (i.e., arm-reach distance). When the control
space corresponds to the motor space (Fig. 2 A), the user cannot
interact with an out-of-reach object. To interact with it, the user
has to move to the remote position rst [
5
,
10
] using a navigation
technique (e.g., Teleportation [
9
].) To avoid this, some interaction
techniques provide a bigger control space than the user’s arm-
reach to select out-of-reach objects. An example is ray-casting, a
widely adopted selection technique in many commercial VR and
AR devices [
31
,
32
,
50
]. It shoots a ray from a controller or the
user’s head to select a remote object that the ray hits. Another
example is an image-plane pointing technique, also known as Sticky
Finger [
37
]. It allows the user to select a distant object that is
occluded by using one of the index ngers. However, they provide
limited support for object manipulation.
There are two categorical approaches to fully support remote 6
degree-of-freedom (DOF) object manipulation (xyz + yaw, pitch,
roll) in IVEs. The rst approach is oset techniques (Fig. 2 B), ex-
panding the control space by multiplying the motor space by a scale
factor that impacts on the control-display (CD) ratio. The CD ratio
determines how the input device’s movements (
Δ
x) are mapped to
the virtual cursor’s movement (
Δ
X). It is dened as (
Δ
x) / (
Δ
X). For
example, if the CD ratio is 1, the control space is equivalent to the
motor space (i.e., simple virtual hand as Fig. 2 A). If the it is smaller
than 1 (
Δ
x <
Δ
X), the user can reach remote objects as the control
space is expanded by 1 / CD ratio. Examples of this oset technique
are Direct HOMER [
7
],Go-Go [
39
], Linear Oset [
24
,
25
], Voodoo
Dolls [
38
], and World in Miniature (WIM) [
45
]. These techniques
have two limitations. 1) precise control of out-of-reach objects is
dicult due to the sensitive CD ratios. 2) for some techniques, one
hand must be tied to the widgets (e.g., Voodoo Dolls and WIM).
The second approach is the clutching mechanism [
2
] that relo-
cates the control space to nearby target objects (Fig. 2 C). It keeps 1
CD ratio for the user to interact with out-of-reach objects directly
within a limited space. PORTAL adopts the clutching mechanism
to allow the user to precisely interact with remote objects while
keeping both hands free in order to overcome the limitations of the
aforementioned techniques.
2.2 Secondary View Interaction in IVE
The secondary view is an additional view that displays a dierent
perspective than the user’s primary view to provide additional
interactions for the user. The Magic Lens technique [
4
,
48
], for
Figure 2: (A) Simple hand metaphor can not reach remote
objects because they are located out of reach. (B) Oset tech-
niques extends the control space by multiplying the motor
space by a scale factor. (C) PORTAL relocates a portion of
the control space near the remote objects.
example, reveals hidden information when it is overlaid on part of
the primary view. Its applicable scenario is the medical training [
11
,
41
] to allow the user to explore anatomy (e.g., bones and organs)
by overlaying the lens on a virtual human body.
Various types of secondary view interaction have been intro-
duced. Photoportals [
22
] promotes telepresence between remote
users in a virtual environment projected on 3D powerwalls to take
photos and videos together. Nam et al. [
34
] introduce Worlds-in-
Wedges for visual comparison by rendering dierent VR scenes in
multiple widgets.
Secondary views are also studied for 3D object. Li et al. [
26
]
proposed vMirror, an interactive widget leveraging reection of
mirrors to select remote and occluded objects. But it lacks object
manipulation. Martin et al. [
29
] also adopted a mirror metaphor
providing a new perspective around an object for accurate object
alignment. Stoev and Schmalstieg [
46
] introduced Through-The-
Lens (TTL) to allow the user to interact with remote objects through
TTL with the limited DOF. To our best knowledge, only Tunnel
Window [
19
,
21
] fully supports remote 6-DOF manipulation. It
uses ray-casting as its selection tool. Compared to Tunnel Window,
PORTAL uses a simple virtual hand metaphor allowing the user to
interact with the target directly [
51
] through the secondary view.
In addition we conducted two user studies to evaluate its eciency
and usability on the remote target selection and manipulation tasks
while no evaluation was provided for Tunnel Window.
3 PORTAL: A REMOTE INTERACTION WIDGET
PORTAL is an interactive widget leveraging a secondary view. It
is comprised of the primary portal and the secondary portal as
shown in Fig. 3. The portals work as a spatial tunnel to connect
dierent locations in IVEs to gives the user an impression that
remote objects are right in front of him or her. The source codes
and data are available at our GitHub repository 1.
3.1 Design Considerations
We design PORTAL with two considerations: 1) the user’s control
space should reach remote objects and support direct selection and
manipulation of them; 2) it should provide precise depth perception
to the user to facilitate the ne-grained controls to the remote
objects. We satisfy the rst design consideration by adopting the
1https://github.com/VIZ-US/PORTAL