IEEE Consumer Electronics Magazine - May 2018 - 84

A set of simple peripheral LEDs
mounted on the helmet's frame
can easily allow a skier to be
aware of other skiers approaching
from behind.
LEDs mounted on the helmet's frame can easily allow a skier
to be aware of other skiers approaching from behind.

OFF-SLOPE EXPERIMENT

Average Detection Success Rate (%)

To evaluate the effectiveness of our prototype, we conducted
two experiments, with a total of 26 participants and skiing/
snowboarding enthusiasts. The first study was intended to validate the concept of our prototype by approximating the traverse-slope conditions in an outdoor (but not on-slope) setting,
where the subjects would simply walk a straight route wearing
the s-Helmet while a researcher would approach from behind
at different speeds. This was mainly for receiving feedback
about our prototype over the summer and improving it further
for on-slope experiments in the upcoming winter season.
For this first experiment we had 20 participants, with
an average age of ~ 30 ( M = 29.9, standard deviation 6SD@ =
6.488), ranging from 23 to 50 years old, whom we subjected
to three conditions: normal helmet (condition A), s-Helmet
(condition B), and s-Helmet while listening to music (condition C). In condition A, the subjects wore a normal helmet and walked a straight line while a researcher crossed
behind them in a random zigzag fashion. The participants
tried to identify from which side the researcher was
approaching by lifting their corresponding hand: the left
hand for the left side, right hand for right side, and both

100
80
60
40
20
∗

∗

0
A

B
Condition

FIGURE 2. Our participants' detection scores per condition for the
off-piste study. Condition A is with a typical ski helmet, condition
B is with the s-Helmet, and condition C is with the s-Helmet while
listening to music.
84 IEEE Consumer Electronics Magazine

MAy 2018

hands for the center. A supervisor registered the subjects'
successful attempts to identify the position of the moving
researcher. We permitted the participants to use any contextual cues to identify the position of the researcher behind
them, such as shadows and noises, but we did not allow
them to turn around and look. For condition B, everyone
followed the same process, only this time the participants
wore the s-Helmet, which offered the three LEDs to identify the position of the moving researcher behind them.
Finally, condition C involved the use of the s-Helmet while
the subjects listened to music via earplugs to simulate hearing interference.
After the end of each condition, we asked the participants to
complete a NASA Task Load Index (TLX) questionnaire [7] for
assessing self-reported mental and physical effort in percentages,
while we registered their successful detection scores (a maximum of ten for the ten random movements we performed) for
that condition. When the subjects completed all three conditions,
they filled out a System Usability Scale (SUS) questionnaire [8]
for assessing the overall perceived usability of the s-Helmet on a
scale from one to 100. Finally, the participants completed a
questionnaire inquiring about their skiing/snowboarding habits,
and we briefly interviewed them about their thoughts and feelings regarding the prototype and the experiment.

OFF-SLOPE RESULTS
While wearing the s-Helmet, participants were able to accurately detect the position of the researcher moving randomly
behind them 70% of the time, as opposed to 20% when wearing a typical helmet (see Figure 2). With the s-Helmet while
listening to music, our subjects detected the moving researcher just as accurately (70%) as in condition B. The average
workload scores were almost equal across all of the conditions, with 33.95% ^SD = 15.85h for condition A (normal
helmet), 32.7% ^SD = 14.65h for condition B (s-Helmet),
and 34.81% ^SD = 15.26h for condition C (s-Helmet while
listening to music). Interestingly, when no augmentation was
used (condition A), participants reported that they increasingly used contextual cues, such as shadows, for determining the
position of the moving researcher.
Overall, the subjects evaluated the s-Helmet on average
usability, with an average SUS score of 67.12 ^SD = 13.08h .
(A system that achieves an SUS score of >68 is considered to
be above average usability [9].) Novice skiers found the
s-Helmet significantly more usable than did expert skiers.
Moreover, people who would usually not wear a helmet on
the slopes reported a significantly higher workload than people who ordinarily wear one. During the exit interviews, participants also reported a range of important usability issues,
mainly regarding the feedback mechanism employed in our
prototype. The most commonly reported issue was the LEDs'
placement either too close to the wearer's eyes or, at times,
outside the wearer's field of view. Despite the issues reported, our subjects were positive about the s-Helmet's usefulness and thought the device could potentially increase overall
on-slope safety.

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