Context Aware Computing

1
Context Aware Computing

• Shawn Outman and Charles Vermette
• CEN 5531 Term Paper Presentation
• November 20, 2003

2
Context Defined

• Indirect input
• In contrast to direct user input, such as
  clicking on a menu item
• Examples
• Users location
• Time of day
• Available resources

3
Divisions of Context

• Proximate selection
• Use of location to automate a choice
• Automatic reconfiguration
• Includes resource discovery and mobile profiles
• Contextual commands
• Use of context info to define or aid a computation

4
Types of Context

• Computing
• Hardware configuration of device, power level
• User
• Mood, preferences, biometrics
• Physical
• Temperature, light level
• Time
• Year, Day, Minute, etc.

5
Motivation For A Framework

• Support generic acquisition and use of context
• Allows for component reuse and compatibility
  between various sensors
• Reduce burden on application to support context
• Reduce development time and resources

6
Solution Widget

• Familiar paradigm for direct input
• Context consumer queries widget for information
• Independent of device
• Program can query the pointer to find where the
  user is pointing
• This pointer can be an optical mouse, trackball,
  pen, keyboard cursor, etc.

7
Differences from Classical Widgets

• Context requires some changes from the
  established direct input widget model
• New obstacles/requirements
• Distributed environment
• Widget lifecycle
• Multilayered abstraction

8
Distributed Environment

• Need common exchange medium, since sensors,
  various consumers may all be different platforms
• Solution client/server with pluggable transport
  and encoding mechanisms
• Default HTTP and XML for maximum portability

9
Widget Lifecycle

• Widget needs to exist beyond the application
  lifecycle
• May be used by many applications
• Needs to be able to give past data if needed by
  new subscriber
• Solution client/server
• History supported with persistent storage
• Default storage database

10
Multilayered Abstraction

• Raw sensor data may be useless to many
  application application
• Application often needs interpreted context data
• May be difficult/expensive/hard to update
  interpretation code in mobile agent
• Solution Multilayered framework

11
Layers of Framework

• Widget handles one sensor or piece of context
• Example data from proximity sensor
• Server aggregates widget data
• Example data from all proximity sensors in
  building
• Interpreter takes other pieces and forms a
  higher abstraction of the data
• Example rooms where people are congregating

12
Benefits of Framework

• Hide details of context gathering from
  application
• Less code executed, power used on mobile device
• Sensors loosely coupled to application
• Easy to change vendor, or even type
• Reduced overhead to add context support
• Encourages reuse of context components

13
Open Problems

• Handling of streaming data
• Client/server polling implicitly discretizes data
• Lifecycle management not integrated into
  framework
• Need watchdog to spot when a sensor needs reboot
• Context discovery not transparent
• Requires configuration by application
• No history keeping by interpreters?

14
Application

• Wearable conference assistant
• Published in 1999
• Used framework to add context to wearable
  application
• Provides relevant info for participants as they
  attend conferences
• Allows them to focus on conference content
  instead of planning

15
Components

• User server
• organizes all context about one user
• Presentation server
• aggregates all info about one presentation
• Registration widget
• gets user's info at start of conference
• Memo widget
• tracks user note taking
• Question widget
• tracks question askers

16
Components (cont.)

• Content widget
• archives presentation slides and materials
• Recording widget
• tracks if a presentation is being recorded
• Location widget
• used to track attendees at a presentation
• Recommender interpreter
• suggests which presentations a user would be
  interested in

17
Component Relationships
18
Operation

• Widgets collect data about users, presentations
  and report this back to each user
• Allows users to plan their conference activities
• stick around or avoid people
• spread out to maximize conference coverage

19
Hardware

• RFID tags for user tracking
• Sensors and context components running on
  standard desktops
• PC, Mac and Solaris
• Wireless network
• User interface on laptops and handhelds
• no actual wearable devices

20
Wearable Characteristics Met

• Portable while operational
• Hands free use
• Sensors
• Requirement intended to put sensors on body, but
  authors argue that having the sensor data on the
  user good enough
• Proactive
• Always-on

21
Context Aware Capabilities Supported

• Contextual sensing
• Contextual adaptation
• Contextual resource discovery
• Contextual augmentation

22
Bibliography

• 1 Bill Schilit, Norman Adams, and Roy Want.
  Context-Aware Computing Applications. IEEE
  Workshop on Mobile Computing Systems and
  Applications, Santa Cruz, CA, US, 1994
• 2 Guanling Chen and David Kotz. A Survey of
  Context-Aware Mobile Computing Research.
  Technical Report TR2000-381, Dept. of Computer
  Science, Dartmouth College, November, 2000.
• 3 A. K. Dey, G. D. Abowd, and D. Salber, A
  Context-Based Infrastructure for Smart
  Environment, In Proceeding of the 1st
  International Workshop on Managing Interactions
  in Smart Environments (MANSE '99), December 1999,
  pp. 114-128.
• 4 A. Dey, M. Futakawa, D. Salber, and G. Abowd,
  The Conference Assistant Combining
  Context-Awareness with Wearable Computing, In
  Proceedings of the 3rd International Symposium on
  Wearable Computers (ISWC '99), San Francisco, CA,
  USA, IEEE Computer Society Press, pp. 21-28,
  October 1999.
• 5 B Rhodes. The Wearable Remembrance Agent A
  system for augmented memory. Proceedings of the
  1st International Symposium on Wearable
  Computers, pp. 123-128, October 1997.
• 6 J Pascoe. Adding generic contextual
  capabilities to wearable computers. Proceedings
  of the 2nd International Symposium on Wearable
  Computers, pp. 92-99, October 1998.

23
Questions?

• Go ahead and ask
• We cant sit down until you do