2018ISMAR Comparing Different Augmented Reality Support Applications for Cooperative Repair...

目录

ABSTRACT

Introduction

Aim of this study

User study

2.2.1 Head mounted display AR: condition ”HMD”

2.2.2 Projection based AR: condition ”Projection”

2.2.3 AR on tablet computer: conditions ”Video”, ”Screenshot” and ”Tracking”

2.2.4 Audio-only: condition ”Phone”

Results

SUMMARY AND CONCLUSION

---

ABSTRACT

Digitization and the growing capabilities of data networks enable companies to perform tasks via remote support, which previously
required service personnel to travel. But which mixed reality method leads to better results regarding human factors, grounding and performance criteria? This paper reports on a collaborative user study, in which a local worker is guided by a remote expert with the help of different augmented reality methods, specifically see-through HMD, spatial projection, and video-mixing tablet. The task to perform is the exchange of a controller in a switch cabinet of an industrial robot, a task rather typical for failure detection within the field. Our study was conducted in collaboration with a technician school, where 50 technician apprentices participated in our study. Our results show clear advantages of using augmented reality (AR) versus traditional conditions (audio, video, screenshot) to enable remote support. It further gives significant indications for using a projection based AR method.
 

Introduction

 

This paper reports on the development and evaluation of three different AR-methods targeting the local collaborator at the production site who is assisted by a remote expert. The three systems are based on (1) tablet-based AR, (2) optical see-through AR, and (3) projection-based AR. We compare these approaches with typical usability measures of humancomputer interaction (e.g., QUESI, NASA-TLX, ISONORM, etc.) and highlight pros and cons of each approach.
 

Aim of this study
 

Before the experiment, we formulated the following hypotheses:
• Participants wearing augmented reality glasses (head mounted display, HMD) are likely to perform the task quicker than when using other applications, because the user does not need to put the tablet out of hand or bend back for the projection based method. That is also why we think that of all tested variants, the HMD will provide the best support for building up situation awareness and grounding.

• The participants might prefer to use the projection based method, because it is more intuitive and requires a lower task
load. We think that this is due to the fact that it provides the most stable information presentation and is easiest to use, as
the worker does not need to do much on his or her own. This will show in the values for the questionnaires.

• For the grounding we assume that the HMD is the best solution for supporting both workers. It is the only application which
provides a continuous view for the expert and we also presume that this first person viewpoint enables the expert to help the
participant in a better way.
 

User study

2.2 Experimental setup

2.2.1 Head mounted display AR: condition ”HMD”

A head mounted display (HMD) is a display device worn on the
head or as part of a helmet, that has a small display optic in front
of one (monocular HMD) or each eye (binocular HMD). For this
experiment we used the Epson Moverio BT200, which is shown in
Fig. 2. Those are glasses which have a LCD polysilicium display
integrated for each eye, illuminated by mini projectors on the rim.
The displays provide a resolution of 960x540 pixels and cover 23
percent of the field of view. There is an external control device with
a touchpad, using a 1.2GHz dual-core-processor and 1GB RAM. It
uses Android 4.0.4.
The Android application connects to the expert’s Desktop PC
application, which is displayed in Fig. 1 via Wi-Fi and tries to
visually track camera picture features at the location of the expert’s
annotations. We have published a separate paper about the tracking
algorithms [11].

2.2.2 Projection based AR: condition ”Projection”

The projection based augmented reality application (or spatial augmented reality application [2]) uses a Panasonic PT-VZ575N projector with a resolution of 1920x1200 pixels. The setup is displayed
in Fig. 3: Mounted rigidly to the projector is a PointGrey Blackfly camera that offers a static third person perspective of the work
surface. After the camera and the projector have been calibrated, a
Structured Light approach is used to gather a surface model of the
working site. With the help of this model, the expert can directly
send visual instructions to the surface of the working environment,
which is displayed in Fig. 4. The worker also can use a pointing
device tracked by a Polaris Spectra system in order to visualize his
view. More publications to this specific setup can be found at [9].

2.2.3 AR on tablet computer: conditions ”Video”, ”Screenshot” and ”Tracking”

Tablet PCs have been commercially available for several years and
are regularly used in everyday life as well as in industrial settings.
This is why this device serves as a representative of the current
common practice in industry, and its performance in the described
experiment is also meant to encourage more computer supported
collaborative work with a shared visual view in the field. In our prior
work, we used tablet computers for cooperative work in an active
production environment [1].
This is also one reason why we compared three different applications running on the tablet PC: a version which provides just the
camera picture, a version which enables the expert to make a screenshot and annotate it (see Fig. 5) and a permanent tracking approach
(see Fig. 6). The latter figure also shows that the tablet PC needs to
be placed on the ground during the actual repair work.
We use a ASUS MEMO ME302C Tablet with an 1.6GHz Intel
Atom Z2560 Processor with 2GB RAM. The device runs Android
4.3 and the display has a resolution of 1920x1200 pixels.

2.2.4 Audio-only: condition ”Phone”

In order to provide a baseline to the AR application experiments,
we chose contact between worker and expert without shared visual
context by introducing a ”telephone” condition. In our experimental
setup, both users cannot see each other (see Fig. 7) but can have a
normal conversation in the same room. This is the optimal condition
which eliminates all latency of the audio, and this audio condition
is used for all of the experiments. There is of course an influence
of the audio quality and latency on the quality of CSCW, but it
is intentionally left out in this experiment in order to compare the
visual conditions only.
 

 

 

Results

 

 

SUMMARY AND CONCLUSION
 

This study and its findings contribute to the improvement of remote maintenance tasks by using concepts of industry 4.0, namely computer supported collaborative work. In an industrial remote maintenance task, a worker is supported in a repair task by a remote supervisor. The service technician is there to execute maintenance orders while the supervisor trusts on the technician to obtain situational awareness. Using a practical repair task, research on the impact of information highlighting using different means of multimedia representation was carried out. Different optical tracking algorithms were compared and implemented according to their suitability for industrial remote maintenance tasks. A tablet PC, optical see-through AR-Glasses, and a spatial AR projector were used as visual output media.