A Survey on Context-Aware Systems Center for E-Business Technology Seoul National University Seoul,...

A Survey on Context-Aware Systems

Center for E-Business TechnologySeoul National University

Seoul, Korea

Sangkeun Lee

Intelligent Database Systems LabSchool of Computer Science & EngineeringSeoul National University, Seoul, Korea

Copyright 2008 by CEBT

Current Situation

Advanced network & Computing technology

Wide spread of mobile devices

Rapidly growing number of services

Too much contents

Realizing context-aware systems is a key driver for solv-ing these problems

Introduction

UserGPS Navigator

Cell phone

Home Network

PC

Digital camera

Voice recorder

PDAUser’s Context

Service

Service

ServiceService

Content

Content

Content

Content

How can we provide the most appropriate service or content to

users?


Agenda

Introduction

Definitions

Characteristics of Context-aware Systems

Requirements of Context-Aware Systems

History of Context-Aware Systems

Trend analysis : 1st, 2nd, 3rd generation context-aware systems

Summarize the characteristics of several representative context-aware sys-tems

Comparison

Conclusion

Introduction


Definitions

Context

any information that can be used to characterize the situa-tion of entities that are considered relevant to the interac-tion between a user and an application, including the user and the application themselves (Dey and Abowd ,2001)

Context-Aware Systems

define context-aware systems as the systems that use con-text to provide relevant information and/or services to the user, where relevancy depends on the user’s task (Dey and Abowd ,2001)

Definitions


A General Context-Aware Service Process

Acquiring Context Information

Storing Context Information

Controlling Abstraction Level of Context Information

Utilizing Context Information for Services or Applications


Acquiring Context Information

Storing Context In-formation

Controlling Abstrac-tion Level of Con-text Information

Utilizing Context In-formation for Ser-

vices or Applications

Physical Sensors

Other Sources (Virtual Sensors)

Context Models

Context Aggrega-

tion

Context In-terpretation

Context as Triggering Condition

Context as Additional

Information


Context Acquiring

A Context-aware system is responsible for gathering context informa-tion of entities (e.g. people, devices, places, etc)

Physical Sensors

Hardware components that converts physical context information into com-putable signals

Various physical sensors are available

GPS, Active Badge System, RFID

Temperature, Pressure, Humidity, Light, Ultrasonic Sensors

Microphones, Cameras

Acquiring Con-text Information


Controlling Ab-straction Level of Context Informa-

tion

Utilizing Context Information for Services or Ap-

plications

Physical Sensors


Context Models

Context Aggrega-

tion

Context Inter-pretation

Context as Trigger-ing Condi-

tion


Information



Context Acquiring

Virtual Sensors

Using physical sensors is not the only way of gathering context in-formation of entities

Software components that retrieve context information various source

– Getting weather information using a web service

– Achieving user’s preference on music by analyzing user’s playlist

– Acquiring user’s current status by getting user’s explicit submit




tion


plications

Physical Sensors


Context Models

Context Aggrega-

tion



tion


Information



Context Storing

Efficient storage and manage-ment of context information is crucial

Closely related to context mod-els

Criterias: Expressiveness, Flexibility, Generality, Com-putational Cost

Context Model Examples

Key-Value Model Schilt’s Approach, Context Toolkit

Logic-based Model McCarthy’s Ap-proach

Mark-up Scheme Model

CC/PP, UAProf, CSCP, GPM

Object-oriented Model

Hydrogen

Graphical Model Context extension of ORM, Vector space model

Ontology Model SOCAM, CoBrA, CASS, CoCA




tion


plications

Physical Sensors


Context Models

Context Aggrega-

tion



tion


Information



Controlling Abstraction Level of Context In-formation

Context Aggregation

To reduce the number of low-level context information for better performance

E.g. Temperatures of previous 10 minutes -> Very Cold

Context Interpretation

To grant high-level semantics to sensed context information (Context Inter-pretation)

E.g. GPS signal -> City name

If separate context-abstraction and context-aware application, applica-tion doesn’t have to the know details of sensors but still can utilize con-text-data




tion

Utilizing Context Information for Enhanced Ser-

vices or Applica-tions

Physical Sensors


Context Models

Context Aggrega-

tion



tion


Information



Utilizing Context Information for Services & Ap-plications

Two ways of utilizing Context Information

Context Information as Service Triggering Condition

– E.g. If a user is hungry and thirsty, then trigger food delivery service

Context Information as Additional Information for Enhanced Service

– E.g. Providing tour-guide information based on user’s current location




tion



Physical Sensors


Context Models

Context Aggrega-

tion



tion


Information



Utilizing Context Information for Services & Ap-plications

Context Information can be used for various types applications

Context-Aware Personalization

– Providing personalized contents or information to user based on the user’s current context

Automatic Device Configuration

– Automatically setting up device’s configuration according to user’s situation

Context-aware User Interface

– Automatically Optimize device’s user interface based on user’s preference or con-text

Context-aware Suggestion

– Suggests or recommend the most suitable action to users based on current situa-tion




tion



Physical Sensors


Context Models

Context Aggrega-

tion



tion


Information



Context-Aware System Architecture Style

A Context-Aware System can be implemented in many ways

Device

Sensors

Con-text

Device

Sensors

Con-text

Device

Sensors

Con-text

Device

Sensors

Con-text

Device

Sensors

Con-text

Device

Sensors

Central Con-text Server

Context

Device

Sensors

Device

Sensors

Sensors Sensors Sensors

Stand-alone Architec-ture

Distributed Architec-ture


Centralized Architec-ture




Simple architecture

Directly accessing sen-sors to get context-data

No context information sharing or communica-tion among devices

Device

Sensors

Con-text





Distributed Architecture

Context information can be stored in multiple locations

No centralized server

High degree of failure-tolerance

Less congestion Problem

Complex communication proto-col required

Centralized Architecture

All context information are man-aged in a centralized context server

Low degree of failure-tolerance

Congestion(Bottleneck) Problem

Easier to Implement

Device

Sen-sors

Con-text

Device

Sen-sors

Con-text

Device

Sen-sors

Con-text

Device

Sen-sors

Con-text

Device

Sen-sors

Central Con-text Server

Context

Device

Sen-sors

Device

Sen-sors

Sen-sors

Sen-sors

Sen-sors

Distributed Archi-tecture

Centralized Archi-tecture



History of Context-Aware Sys-tems

We review several existing context-aware systems and summa-rize the history of context-aware systems to analyze the trend

Categorize representative context-aware systems into three groups based on their characteristics

1st generation context-aware systems

– The Active Badge location System, Cyberguide

2nd generation

– Watson Project, Context Toolkit, Hydrogen, Aura, Gaia

3rd generation

– CoBrA, SOCAM, CASS, SOCAM, Confab, CORTEX, CoCA

History of Context-aware Systems


1st Generation Context-Aware Systems

Characteristics

Period : 1990 ~ 2000

Location-aware system & Domain-specific application

Limited types of context were used – Location & Time

Focus on how to achieve location information – GPS/ Active Badge / IR / Ultra-sonic

Example Systems

The Active Badge Location System

Cyber guide



The Active Badge Location System

A location-aware system

Active Badge

A tag that emits a unique ID for approximately a tenth of a second every 15 seconds

a network of sensors placed around the host building picks up the signals from active badges

A master station polls the sensors makes the sensed context data available to clients

Operations

A command interpreter allows simple investigations to be performed on the system

FIND (name), WITH (name), LOOK (location), NOTIFY (name), HISTORY (name)

Application

Forwarding phone calls to nearest phone to the user

Semantic Tech & Context - 18




Cyberguide

A domain-specific context-aware application

Limited context types

Location and time

Context Acquiring

Indoor Location : IR Sensor

Outdoor Location : GPS Sensor

Functionality

A mobile context-aware tour guide

Providing information based on tourist’s current location and time

Tour diary: keeping log of places tourist have visited





Cyberguide





2nd Generation Context-Aware Systems

Characteristics

Period : 2000 ~ 2003

Earlier attempt to build context-aware frameworks for rapid prototyping vari-ous context-aware applications

Support for various context types (More than location and time)

Example Systems

Context Toolkit, Hydrogen, Gaia

– Too simple context model (Context Toolkit, Hydrogen)

– Resource discovery not supported (Context Toolkit)

– Support only for specific types of context (Hydrogen)

– Only use built-in sensors (Hydrogen)

– Lack of privacy protection (Context Toolkit, Hydrogen, Gaia)

– Lack of performance concern (Context Toolkit, Hydrogen, Gaia)



Context Toolkit

Aim is to develop reusable solution to make development of context aware applications easier

Inspired by the success of GUI development kits

GUI Toolkits

Hide underlying complexity

Manage the details of interaction

Provide reusable building blocks

A distributed architecture + central discoverer

Widgets [Context Widget]

Hide the complexity of actual sensors used from application

Abstract context information to suit the expected needs

Provide reusable building blocks for context sensing

Aggregators

Aggregate context information

Hide even more complexity about the context

Interpreters

Used to abstract or interpret low-level context information

Resource discovery not supported

Context Model Attribute – Value Model

22



Context Acquiring

Context Abstraction


Hydrogen

Context-awareness framework on mobile devices

A distributed layered architecture Application layer

Management layer

– Providing and retrieving contexts and sharing context information with other devices using P2P communication

Adaptor Layer (Context Acquiring)

– Separating context storing, sens-ing from other layers

– Responsible to get information from sensors

– Providing same context informa-tion to multiple applications

All layers are located on one device

All application have access to any context data by querying the Con-text Server

Context Sharing Via WLAN, Bluetooth

Context Model Object-oriented Context Model

Resource discovery is not required Only use built-in sensors



Copyright 2008 by CEBT Semantic Tech & Context - 24

Gaia

A ubiquitous software platform

encapsulates the heterogeneity of active spaces, and presents them as a programmable environment, instead of a collection of individual and disconnected heterogeneous devices

Context-aware extension to traditional OS

Active Space

an extension to physical space capable of sensing user actions and equipped with a large variety of devices will assist users with dif-ferent tasks

A distributed architecture




Gaia

Context Acquiring & Abstraction

Context Provider : Acquires context information from sensors or external sources (e.g. stock price)

Synthesizer Context Provider : Infer high-level context from low-level context

First Order Logic

– Context(Number of people, Room 2401, >, 4) ANDContext(Application, Power-point, is, Running) => Context(Social Activity, Room 2401, Is, Presentation)

Context Model

4-ary predicate

Context(<ContextType>, <Subject>, <Relater>, <Object>)

Context Storing

Context File System

Context Consumer

Application service that consumes context information

Application Framework





3rd Generation Context-Aware Systems

Characteristics

Period : 2003 ~

Ontology Context Model

– Formal, expressive, flexible context model

– Suitable for knowledge Sharing

– CoBrA, SOCAM, CASS, CoCA

Focus on privacy protection

– Context Fabric

Performance & scalability

– Context Fabric, CoCA

Support for easy implementation of context-aware application

– CORTEX



OWL (Web Ontology Language)

W3C announced the OWL (Web Ontology Language) in 2003

A semantic markup language

Designed for the Semantic Web

Formal, flexible and expressive

Has many predefined classes and properties

Suitable for knowledge sharing




CoBrA

Architecture Style

An agent-based context-aware system

A centered resource-rich con-text broker maintains and manages the shared context data

Context Model

Uses OWL to model ontological context & privacy policy

Privacy Protection

Rei : Allows users to define pri-vacy policy so that it can pro-tect privacy of users

Context knowledge API - assert, delete, modify, query context data



Context Knowledge Base

Privacy Management Module

Context Inference Engine

Context Acquisition Module

Context Storing

Privacy Protection

Context Abstraction & Reasoning

Context Acquiring


SOCAM

A centralized architecture

Context Model

A common context model that can be shared by all devices and services

Upper/Domain-specific Ontology : Reduce ontology processing cost

A set of services that perform context acquisition, context discovery, context interpretation and context dissemination

Context Providers

• Acquire context from sources

Context Interpreter

• Provides logic reasoning

Context Database

• Stores context ontologies

Context-aware Services

• Adapt the way they behave

Service Locating Service

• provides a mechanism where Context Providers and the Context Interpreter can advertise their presence




SOCAM 3rd Generation Context-Aware Systems



The Context Fabric

Primarily concerned with privacy rather than with context sensing and processing

provides an architecture for pri-vacy-sensitive systems, as well as a set of privacy mechanisms that can be used by application devel-opers

Share the right information, with the right people and services, in the right situation

Downside: It does not address tradi-tional distributed systems require-ments such as mobility, scalability, component failures and deployment/configurations

XML Context Tuple




Copyright 2008 by CEBTIDS Lab. 32

The CoCA Service Platform

Interface Manager

Manages a UI and interface between the CoCA platform and other modules

Data Source

Responsible to provide necessary data to the core service (GCoM)

Core Service

Responsible to provide the core context aware service after reasoning on the com-ponents

Supplementary Service

Knowledge discovery & Collaboration service

The Platform aims at acquiring and utilizing context information to provide appropriate services

E.g) A cell phone is always set to vibrating mode when its holder is in the libraryIt filters and sends use-

ful contexts to the con-text repository

It Keeps the rules in the rule repository

It consists of domain depen-dent/independent ontology

Reasoning -> Decision & ActionInterpretation, Aggregation




The Enhanced CoCA Service Platform

Enhanced version of CoCA Combine the best of the relational approach and ontology approach

Selective feature of loading only relevant context data into the rea-soner using heuristics




ComparisonSystem Name

Architecture Style

Context Acquiring

Context Model

Context Repository

Context Abstraction

Handling Dynamicity

Privacy Protection

Performance &

Scalability

1st generation

The Active Badge

Location System

Stand-alone Architecture

Active Badge

- - - - - -

CyberguideStand-alone Architecture

GPS, IR - - - - - -

2nd generation

Context Toolkit


Context Widget

Attribute-Value

-

Context Interpreter,

Context Aggregator

Central Discoverer

Context Ownership

-

HydrogenDistributed Architecture

Adaptor Layer

Object-oriented Model

- - - - -

GaiaDistributed Architecture

Context Provider

First Order Logic 4-ary predicate

Context File System

Synthesizer Context Provider

Presence Service

- -

3rd generation

CoBrACentralized Architecture

Context Acquisition

Module

Ontology Model

Context Knowledge

Base

Context Inference Engine

Context Broker

Rei : privacy policy

language-

SOCAMCentralized Architecture

Context Provider

Ontology Model

Context Database

Context Interpreter

Context Service Locating Service

-

Upper/Domain-specific

Ontology

The Context Fabric


InfospaceXML

Context Tuple

Infospace - Infospace Privacy Tag -

CoCADistributed Architecture

Data Source Module

Ontology Model

Data Modeling &

Storage Layer

Pre-processing

Layer

Collaboration Layer

-

Hybrid Approach – Ontology &

RDBMS

Comparison


Summary & Conclusions

We explained characteristics & requirements of context-aware systems and surveyed several existing context-aware systems

Summarized Trend

Summary & Conclusion

Context-Aware Application

1st generation context-aware systems


Context-Aware Framework

2nd generation context-aware systems

3rd generation context-aware systems


Context-Aware Framework

Knowledge

Domain Specific

Limited Context Types

Domain Independent

Various Context Types

Simple Context Models

Domain Independent


Sophisticated Context Mod-

els

Simple Methods

for Privacy Protection

Acceptable for non-time crit-

ical context-aware applica-

tions

Context Types

Domain

Context Models

PrivacyProtection

Performance& Scalability

Architecture


Future Direction We still need to work on “Privacy Protection” & “Performance & Scalability”

Should be ready for ‘time-critical applications

Sophisticated privacy handling is required

Balancing Performance, Scalability & Intelligence is important

Interface Standardization is required

Heterogeneous devices

Too many and similar systems

Future context-aware systems should become more service centric.

What is a Service?

– A set of interaction between applications & entities(devices, context data, places, people, and so on)

Existing context-aware framework support context-aware application development

Service-level historical log should be handled Interface Standardization

Reliable Privacy Protection

Balancing Performance, Scalability & Intelligence

Managing Service-Level Historical data

Summary & Conclusion

Extend Service Layer to Context-Aware Framework

Context-Aware Ap-plication

1st generation context-aware sys-tems

Context-Aware Appli-

cationContext-Aware

Framework2nd generation context-aware sys-

tems3rd generation context-aware sys-

tems

Context-Aware Appli-

cationContext-Aware

FrameworkKnowledge

Domain Specific

Limited Context Types

Domain Independent


Simple Context Models

Domain Independent


Sophisticated Context Models

Simple Methods

for Privacy Protection

Acceptable for non-time critical

context-aware applications

Context Types

Domain

Context Models

PrivacyProtection

Performance& Scalability

Architecture


End of Presentation

Thank You!

A Survey on Context-Aware Systems Center for E-Business Technology Seoul National University Seoul,...

Documents

Transcript of A Survey on Context-Aware Systems Center for E-Business Technology Seoul National University Seoul,...