Vision-based approaches for mobile indoor localization do not rely on the infrastructure and are therefore scalable and cheap. The particular requirements to a navigation user interface for a vision-based system, however, have not been investigated so far.
Such mobile interfaces should adapt to localization accuracy, which strongly relies on distinctive reference images, and other factors, such as the phone’s pose. If necessary, the system should motivate the user to point at distinctive regions with the smartphone to improve localization quality.
We present a combined interface of Virtual Reality (VR) and Augmented Reality (AR) elements with indicators that help to communicate and ensure localization accuracy. In an evaluation with 81 participants, we found that AR was preferred in case of reliable localization, but with VR, navigation instructions were perceived more accurate in case of localization and orientation errors. The additional indica- tors showed a potential for making users choose distinctive reference images for reliable localization.
A Mobile Indoor Navigation System Interface Adapted to Vision-Based Localization
1. A Mobile Indoor Navigation System Interface
Adapted to Vision-Based Localization
Andreas Möller1, Matthias Kranz2, Robert Huitl1, Stefan Diewald1, Luis Roalter1
1Technische Universität München, Germany
2Luleå University of Technology, Department of Computer Science,
Electrical and Space Engineering, Luleå, Sweden
MUM 2012, Ulm, Germany
2. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Outline
§ Motivation for Vision-Based Navigation
§ Challenges
§ Proposed User Interface Concepts
§ Evaluation
§ Discussion and Conclusion
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 2
3. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Background and Motivation
§ Location information still the
most important contextual information
§ Indoor localization is a hot topic
and useful for a lot of scenarios:
□ Airports
□ Hospitals
□ Conference venues
□ Large environments
§ Various indoor localization methods
possible:
□ WLAN/cell-based localization
□ Sensor-based localization
□ Beacon-based localization
□ Vision-based localization
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 3
4. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Why vision-based localization?
Live features
§ Compare query image to reference data set
§ Use existing features in the environment
§ No additional augmentation needed
§ Modern devices are equipped with a camera
and powerful processor
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 4
5. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Challenges of Vision-Based Localization
§ Query images need to be distinctive
□ Repeating structures throughout the environment (e.g. corridors)
□ Little visual features (e.g. uniform walls)
§ Orientation of the user’s smartphone
□ Good candidates for query images (windows, adverts, posters) are
typically found in eye height
□ In a typical pose, user holds the device in a way that it points at
the floor
§ User interface needs to address those challenges
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 5
6. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
User Interface Concept
Our
contrib
§ No 1 Challenge: Localization inaccuracy ution:
Addres
□ When insufficient number of features problem s
from
in the query image a UI
perspe
ctive
§ Solution A: Adaption of the user interface
□ Augmented Reality (AR) with a live view of the environment
□ Virtual Reality (VR) with panorama images
(compare to Google Street View)
taken in advance
§ Solution B: Corrective actions
□ The user needs to point at interesting regions with the camera
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 6
7. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Augmented Reality (AR)
• Display the live camera image
• Impose navigation instructions
over the image
• Smartphone held in
eye height
• Direct match of information and
the real environment
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 7
8. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Virtual Reality (VR)
• Uses preloaded panorama
images
• Navigation instructions are
drawn in panorama
• Smartphone can be directed
to the ground
• User matches environment
and virtual reality
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 8
9. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Pointing Towards Discriminative Areas
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 9
10. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Highlighting Objects of Interest
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 10
11. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Research Questions
§ RQ1: Which concept (AR or VR) is preferable in terms of perceived
accuracy?
§ RQ2: Which concept (AR or VR) is preferred by users?
§ RQ3: Which visualizations are appropriate to acquire sufficient visual
features?
§ RQ4: Can object highlighting be improved with a soft border
visualization?
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 11
12. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Methodology
§ Online Survey
§ Videos and images of mockup system
§ 81 participants (18-59 years) recruited via Mobileworks
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 12
13. Virtual Reality
Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Perceived Accuracy of Virtual Reality and Augmented Reality Views
RQ1: Perceived accuracy
No Error Orientation Error Location Error Loc.+Ori. Error
Response Std.Dev. Response Std.Dev. Response Std.Dev. Response Std.Dev.
1: 2.5 0.9 0.8 2.0 1.7 1.5 0.6 2.0
§ Simulated error conditions 1.4
1.7 1.6 1.8 1.4 1.8 1.0 1.7
2:
□ No Error 1.0
2.4
1.7 1.5
0.2
1.1
2.1
1.8
1.2
□ Orientation Error
1.3
1.8
1.7
0.4
0.9
2.1
1.8
3: □ Location Error
2.3 1.0 -0.2 2.0 □ Both Errors 1.9
0.4 -0.5 1.9
1.8 1.4 1.4 1.6 1.4 1.7 0.9 1.8
Augmented Reality 1: The system seemed to know well my location.
2: The system seemed to know well my orientation.
Virtual Reality 3: I perceived the navigation instructions as correct.
Perceived Accuracy of Virtual Reality and Augmented Reality Views
No Error Orientation Error Location Error Loc.+Ori. Error
Response Std.Dev. Response Std.Dev. Response Std.Dev. Response Std.Dev.
2.3 1.0 -0.2 2.0 0.4 1.9 -0.5 1.9
1.8 1.4 1.4 1.6 1.4 1.7 0.9 1.8
Augmented Reality
-3: strongly disagree
Virtual Reality +3: strongly agree
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 13
14. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
RQ2: User Preferences
§ In total, AR more popular with users than VR
-3: strongly disagree
+3: strongly agree
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 14
15. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
RQ3: Feature Indicator
§ Clarity of the indicator‘s a)
meaning b)
c)
-3: strongly disagree d)
+3: strongly agree
a) b)
c) d)
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 15
16. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
RQ4: Highlighting
methods
§ Objects of interest
highlighting
□ Potentially feature-rich
□ Interaction points for
location-based services
§ Two types
□ Frame
□ Soft border
§ Both equally raise attention,
but Frame distracts more
-3: strongly disagree
+3: strongly agree
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 16
17. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Discussion
§ Accuracy perception through the interface
□ VR beneficial for lower localization accuracy
□ AR preferred for reliable estimate
§ Feature indicators and object highlights potentially contribute to good
reference images
§ Situational use
□ VR for 45° angle
□ Less features visible
□ Less accuracy required
□ AR when holding the phone up
□ Target to visual features
□ Highlight objects of interaction
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 17
18. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Future and
Ongoing Work
§ Evaluating interfaces in the
real world
§ Modeling the system’s
and user’s state
□ Localization accuracy
□ Location/environment
□ User context
□ User mental model
□ …
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 18
19. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Thank you for your attention!
Questions?
?
?
andreas.moeller@tum.de
www.vmi.ei.tum.de/team/andreas-moeller.html
This research project has been supported by the space agency of the German Aerospace Center with
funds from the Federal Ministry of Economics and Technology on the basis of a resolution of the German
Bundestag under the reference 50NA1107.
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 19
20. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Paper References
§ Please find the full paper at:
http://dx.doi.org/10.1145/2406367.2406372
§ Please cite this work as follows:
§ Andreas Möller, Matthias Kranz, Robert Huitl, Stefan Diewald, and Luis
Roalter. 2012. A mobile indoor navigation system interface adapted to
vision-based localization. In Proceedings of the 11th International
Conference on Mobile and Ubiquitous Multimedia (MUM '12). ACM, New
York, NY, USA, , Article 4 , 10 pages. DOI=10.1145/2406367.2406372
http://doi.acm.org/10.1145/2406367.2406372
§
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 20
21. Institute for Media Technology
Distributed Multimodal Information Processing Group Technische Universität München
Please use the following BibTex file:
§ @inproceedings{Moller:2012:MIN:2406367.2406372,
author = {M"{o}ller, Andreas and Kranz, Matthias and Huitl, Robert and Diewald,
Stefan and Roalter, Luis},
title = {A mobile indoor navigation system interface adapted to vision-based
localization},
booktitle = {Proceedings of the 11th International Conference on Mobile and
Ubiquitous Multimedia},
series = {MUM '12},
year = {2012},
isbn = {978-1-4503-1815-0},
location = {Ulm, Germany},
pages = {4:1--4:10},
articleno = {4},
numpages = {10},
url = {http://doi.acm.org/10.1145/2406367.2406372},
doi = {10.1145/2406367.2406372},
acmid = {2406372},
publisher = {ACM},
address = {New York, NY, USA},
keywords = {augmented reality, indoor navigation, user interface, virtual
reality, vision-based localization},
}
[
04.12.2012 A. Möller, M. Kranz, R. Huitl, S. Diewald, L. Roalter 21