IEEE Consumer Electronics Magazine - May 2018 - 97

Sixteen participants (12 males and four females) took part
in the study, all with normal or corrected vision and no colorvision deficiency. In the experiment, a person stood in front
of a tripod with a OnePlus phone facing the mockup factory
model. We adjusted the height of the tripod for each subject,
but the distance to the model was fixed.
We introduced the participants to the study with a written
description that included the statement, "You have to detect
the presence or absence of a machine that requires maintenance." After each trial, a dialog asked if the subject perceived the target to have been present (a yes-or-no answer). A
feedback message stated whether the answer was correct or
not. We provided a training phase of 16 trials for each experimental condition. Then the study began.

RESULTS
The hue condition counted 1,920 trials, with 66 wrong
answers. The error mean in hue/control conditions 1) and 2)
was 0.04 [standard deviation (SD) = 0.2], and the error mean in
the hue/experimental condition 3) equaled 0.03 (SD = 0.16).
The form phase consisted of 1,920 trials, with 20 wrong
answers. The error mean in form/control conditions 4) and 5)
was 0.01 (SD = 0.08), and the error mean in form/experimental
condition 6) equaled 0.01 (SD = 0.12). The experiment's error
mean, considering both the hue and form trials, was 0.02.
We conducted a paired-samples t-test to compare the error
means in the hue and form conditions. There was a significant difference in the hue condition scores (mean = 0.035,
SD = 0.03) versus those of the form condition (mean = 0.01,
SD = 0.021), with t(31) = −4.52 and p < 0.001.

CONCLUSION
This article presented a novel metaphor to bring a machine to
life, as it were, representing sensor data that determined the
machine's status in a nature-inspired metaphor, using features
of a virtual tree. In a world with machines becoming increasingly complex, the metaphor we propose is a step toward
compelling, calm interfaces.
Two studies support our assumptions that BioAR is understandable, compelling, and scalable. The first showed that the
nature-inspired encodings are readable; that is, participants
understood the status of the machine just from the BioAR metaphor. In this study, the subjects stated that they felt an urge to
take care of the machine, thanks to the BioAR metaphor.
The second study showed that the metaphor scales to a
larger number of machines (49) and changes regarding hue
and form features are preattentively perceived. Contrary to
our expectations, participants made significantly more errors
with the hue condition than the form condition. Two explanations are plausible. Healey et al. experimented with shapes on
a constant background. But the AR background interferes
with the visualization, and, in the face of such interference,
variations in hue are less differentiable than variations in
form. The trees used in our experiment have arguably more
complex shapes than the simple 2-D geometric shapes used
in the original experiments. Furthermore, in terms of real-

world experience, humans expect to see a tree change its
aspect, but it is not common that a tree transforms into a different species. This discordance may have attracted early
attention and caused early detection.
The results in both studies encourage us to continue
exploring the BioAR nature metaphor as an innovative way
to communicate real-time sensor information. Following this
vision, a household full of IoT consumer electronics would
communicate itself to its inhabitants as a beautiful garden.
The BioAR nature metaphor can be applied to many other
scenarios where multivariate real-time sensory data are available, e.g., in industry 4.0 plant floors, smart cities, and smart
buildings in the areas of air pollution, traffic, energy consumption, and the like.
The studies helped us identify the following interesting
avenues for research:
▼ Personalized encoding for the sensor data needs to
be developed.
▼ We must address the issue that a large amount of sensor
information integrated in a single visualization decreases
understandability, especially when differentiating between
intermediate states.
▼ A tree has only a few properties that can be used to encode
sensory data. What are the limits of the metaphor?
▼ The connection between the BioAR nature metaphor and
the IoT device needs to be improved (e.g., overlaying
the BioAR onto the device and enhancing the interaction modalities).
We applied this metaphor so users could perceive the machine
as a living being and, by doing so, learn to take better care of
it. The act of caring for the machine rewards the user with an
esthetic and pleasing tree, which enriches the user's environment. Moreover, we expect users' productivity and well-being
to improve from contact with natural (or nature-like) elements
[29]. We surmise that the metaphor positively reinforces the
target behavior with a digital reward, which motivates the user
to establish or strengthen the desired behavior. The first
impressions of our study participants support this assumption.
To date, we performed two studies to validate the metaphor design, concentrating on the perceptual aspects of
decoding information in a nature-inspired metaphor. However, there are other aspects of this metaphor that need to be
explored. Our future work will address the issues of affect
and quality of maintenance. In addition, we should consider
the habituation issue, i.e., when the nature metaphor is introduced, users may feel engaged, but the effect could vanish as
the novelty wears off. We intend to conduct a long-term study
of participants' behavior to investigate our hypotheses.

ACKNOWLEDGMENTS
This work was funded by the LiTech K-project and by
Know-Center GmbH. Both are, in turn, funded by the Austrian Competence Centers for Excellent Technologies
(COMET) program, under the auspices of the Austrian Federal Ministry of Transport, Innovation, and Technology; the
Austrian Federal Ministry of Economy, Family, and Youth;
MAy 2018

IEEE Consumer Electronics Magazine

Table of Contents for the Digital Edition of IEEE Consumer Electronics Magazine - May 2018