Export page to Open Document format View page as slide show

Context-Awareness

Why context-awareness (CA)?

  • People react on same inputs differently depending on the context
  • To make use of Apps & services easier
    • Computers do the work for user
    • Easier interaction
    • E.g. Aid user find relevant information for a given context
  • Added inteligence for automating processes

Context

  • Many definitions by different researchers in the field
    • Dictionary –> context: surrounding conditions
  • Context adds semantic meaning to raw data
    • E.g. coordinate points to locations having contextual meaning
    • x=123,y=456 —> home
  • Context can be defined for different objects
    • E.g. Person, Place, Object, Activity

Different kinds of context

  • Location
    • Europe, Finland, Lappeenranta, …
  • Time
    • 2010, Autumn, October, Thursday, …
  • Activity
    • Meeting, sleeping, watching tv, …
  • Identity
    • Interests, preferences, associations, relations, …
  • Environmental
    • Light level, noise level, polution level, …
  • Behaviour
    • Heart rate, pulse, role, …

Context sources

  • Fetch context from user
    • Explicitly request customer
      • Customer should not be over-burdened
    • Implicitly
      • Via communication with user's device, calendar (meeting, work, vacation, …)
  • Gathering from different sources
    • Integrated sensors
      • Device speed, temprature, location, …
    • Environmental sensors
      • RFID tags providing information
  • Fetch context from context broker (may have done initial processing)
    • Location provider
    • Context producers for MUPE (Multi User Publishing Environment) platform
    • Environment context broker

Context Reasoning

  • Reasoning the context based on fetched raw data
    • Reasoning —> process of drawing conclusions from facts
    • Context reasoning can be explained as:
      • Task of deducing new information relevant to the use of applications & users from various sources of context data
      • E.g. 30C, humid & user is stationary —> at the pool
  • Can be conducted by application or context broker
  • Complex area with plenty of research attention
  • Determine new context from known context
    • User in bed & heart rate is low –> sleeping
    • User has calendar note stating @ meeting –> @work
  • Standardisation immaturity

Personal information as context

  • Application behaviour is altered based on user preferences
    • Preferences are the context
    • Personalisation
  • Preference information acquiring
    • Through registration.
    • From third party e.g. Liberty alliance
    • Profiling i.e. examining the customer behavior
      • Amazon.com
  • Privacy

Context-aware applications

  • Behavious of App changes depending on the context
    • Location Based Services (LBS)
    • Device capability based on Apps
  • E.g.
    • Location –> Mobile phone don't ring in the bathroom
    • Activity –> Mobile phone don't ring during meeting
    • Environment –> Screen light turns on the dark / sound volume increase in nosiy place
    • Target –> Friend is online message sent via IM else via email

Uses of context data

  1. Direct use
    • Get all the context data
    • Use all data as is / infer from it meaningful context to your App
  2. Indirect use
    • Request meaningful pre-processed context data from some context producer
    • Request service adaptation data already affected by the context

Uses of context data ...

  • E.g. personal preference
  • Direct use
    • Request preference from preference broker
    • Request context from context broker
    • Infer what are the preferences for this context
  • Indirect use
    • Request preferences from preference broker
    • Preference broker requests the context from context broker
    • Preference broker deliver the context based preference

Example context platform : Kontti

  • Research project on CA service platform by VTT
  • Kontti platform supports context based information sharing & context based messages
    • Used content is adapted based on user preference, device capabilities & network capabilities
  • Context
    • Time, place, social environment, technical environment, user deniable situations
    • Acquired from either user or E.g. LIF, MLP protocol for location & UAProf for device capabilities

Middleware for context-awareness

  • RCSM – Yauet al., Pervasive computing 1, 2002
  • Confab – Hong & Landay, MobiSys 2004
  • GAIA for smart spaces – Campbell et al, Univ of Illinois at Urbana-Champaign
  • CORTEX – Blair et al at Lancaster Univeristy

Considerations

  • Should my service / App take context into account
    • What context are important?
  • Can I benefit on existing research on context & their meaning?
    • Activity recognition areas, colours & lightning
  • What source I can use to gain context?
Last modified: 2013/07/01 14:42