Page 89 - Kaleidoscope Academic Conference Proceedings 2022

Extended reality – How to boost quality of experience and interoperability

Table 3 – Step count validation The ARN module can be further enhanced by adding a text
to speech translation feature, which could be used to speak
Model Navigational Average Error % out the directions to users, and thereby improving the overall
Mode Estimated accessibility of the application.
steps
(out of 30) The incorporation of computer vision for landmark
Realme Dizo Walking 29.1 3% recognition and comparing the performance with MAR–PNS
Watch 2 Running 28.5 5% and GPS-based identification could be another future area of
work.
MI Band 3 Walking 29 3.34%
Running 28.8 4% REFERENCES
MAR-PNS Walking 29.2 2.67%
System Running 28 6.67% [1] T. Alsop, "Mobile augmented reality (AR) market
revenue worldwide from 2019 to 2025", Statista, 2021.
https://www.statista.com/statistics/897587/world-
The distance estimation sub-module was analyzed with Root augmented-reality-market-value/
Mean Squared Error (RMSE) as a metric to validate the
distance covered by the user. The experiments were carried
out locally, where the user with test application walked a [2] T. Alsop, "Number of mobile augmented reality (AR)
active user devices worldwide from 2019 to 2024",
stretch of road whose distance was already known. The
RMSE for the distance deviation was observed and is listed Statista, 2021.
https://www.statista.com/statistics/1098630/global-
in Table 4.
mobile-augmented-reality-ar-users/
Table 4 – Distance estimation experiments
[3] E. Cranmer, M. tom Dieck, and P. Fountoulaki,
Actual Distance Calculated Squared error "Exploring the value of augmented reality for tourism",
(m) Distance (m) (m) Tourism Management Perspectives, vol. 35, July 2020,
doi: 10.1016/j.tmp.2020.100672.
25 23.4 2.56
30 27.6 5.76 [4] A. Dirin and T.H. Laine, User Experience in Mobile
52 53.2 1.44 Augmented Reality: Emotions, Challenges,
Opportunities and Best Practices. Computers 2018, doi:
70 69.3 0.49 10.3390/computers7020033.
100 99.4 0.36
RMSE 1.45 [5] Google Maps offline navigation.
https://support.google.com/maps/answer/6291838?hl=
5. CONCLUSION AND FUTURE WORK en&co=GENIE.Platform%3DAndroid
The proposed MAR–PNS system provides means for [6] Google Live View.
mitigating the drawbacks of traditional pedestrian navigation https://support.google.com/maps/answer/9332056?hl=
systems and existing MAR-based navigational systems. The en&co=GENIE.Platform%3DAndroid/
major contribution of this work is its novel methodology for
discovering nearby landmarks whose locations are unknown [7] M. Garcia Puyol, D. Bobkov, P. Robertson, and T. Jost,
by just pointing the phone’s camera at those landmarks. This "Pedestrian Simultaneous Localization and Mapping in
improves the user experience as well as opens new junctures Multistory Buildings Using Inertial Sensors," in IEEE
for landmark recognition without using the computationally Transactions on Intelligent Transportation Systems,
intensive traditional methods which rely on image or video vol. 15, no. 4, pp. 1714-1727, August 2014, doi:
processing. Another contribution of this work is cutting 10.1109/TITS.2014.2303115.
down on the reliance of the navigational systems on GPS and
providing a local way of estimating the user’s position by [8] A. Abadleh, E. Al-Hawari, E. Alkafaween, and H. Al-
using their step count. Sawalqah, "Step detection algorithm for accurate
distance estimation using dynamic step length,” 2017
The step count-based user localization module requires the 18th IEEE International Conference on Mobile Data
user to hold the phone in their hand while walking. This can Management (MDM), 2017, pp. 324-327, doi:
be improved in future by applying an additional filter which 10.1109/MDM.2017.52.
mitigates the effect of noise when the user is keeping the
phone in other positions such as their pockets or shaking the [9] Android Motion Sensor documentation.
phone while walking. Integrating this module with a https://developer.android.com/guide/topics/sensors/se
stationary device such as a smart watch can further improve nsors_motion
its accuracy.

– 43 –