Building Mobile Augmented Reality Services in Pervasive Computing Environment

1
Building Mobile Augmented Reality Servicesin
Pervasive Computing Environment

• Hiroaki Kimura hiroaki_at_dcl.info.waseda.ac.jp
• Eiji Tokunaga eitoku_at_dcl.info.waseda.ac.jp
• Tatsuo Nakajima tatsuo_at_dcl.info.waseda.ac.jp
• Distributed and Ubiquitous Computing Lab.
• Waseda University, Japan

2
Motivation - Too-Many-Controllers Problem

• Control panels of current appliances are
  difficult to use.
• Wireless remote controllers are hard to find,
  easy to lost.
• As services disappear, controlling them might be
  more difficult.
• People must remember too many mappings between
  controllers and appliances.

Disappeared?
3
Motivation - TUI and Universal Remote Controller

• Tangible User Interfaces
• Physical form to digital information
• Pros easy to recognize and use
• Cons hard to deploy and reconfigure

musicBottles (MIT)
Gesture with 3D VD (University of Tronto)

• Universal Controller Architecture
• One personal device as a universal controller
• Pros easy to deploy and reconfigure
• Cons hard to recognize and use

iCrafter(Stanford Univ.)
Pebbles (CMU)
4
Movie - Vidgets Virtual Tangible Widgets
5
Challenges
6
Design Issues

• Deployability and Scalability
• Deploy anytime, anywhere
• Update or replace easily and cheaply
• Reusability and High level abstraction
• Deal with high level user events
• Simple interaction style
• Easy to use
• Short setting up time

7
Approaches

• Mobile code
• Download controlling software automatically from
  those services to improve deployability
• Visual tag and Service mapping
• Weak mappings between tags and services code
  improve deployability
• Visual tag is representation of invisible service
• Controlling using real world interaction
• Provides high level sensor API
• Reinforces augmented reality
• Simple interaction cycle
• Defines three interaction stages
• Searching, Selecting, Using

8
Architecture - Overview
9
Architecture - Sequence
10
Implementation
11
Evaluation and Discussion

• Average size of each mobile code 14KB
• Latency of mobile code migration less than 419ms
• Divided attention between controllers and
  services

• Which view a user should look?
• The feedback view on the personal device?
• The application view?

12
Conclusion

• We proposed Vidgets which is a research for
  interaction design and building framework that
  seamlessly integrates personal devices and
  pervasive services
• We implemented a prototype system that confirms
  the effectiveness of our framework.
• Future work
• Detailed user studies (for developers, users)
• Deal with some issues
• Divided attention
• Security risks in mobile code
• Cell phone / PDA implementations

13
Merci beaucoup!
Hiroaki Kimura hiroaki_at_dcl.info.waseda.ac.jp