An Overview of Context-Aware

Information Systems

Contents

Introduction 2

Background 2

Literature Review 3

Emerging Concepts 5

A Critical Evaluation 6

Bibliography 8

Introduction

The ever-growing incorporation of information technology in day-to-day applications

presents new opportunities to develop computer systems that can be aware of the

context in which they are operating. Such computer-systems can be inherently more

responsive to the expectations of their users. Context-aware systems offer

developers and programmers exciting new prospects to gather contextual data and

adapt the behaviour of their dynamic systems according to user expectations. In

conjunction with mobile devices, such mechanisms can be extremely valuable in

increasing the usability of information systems. However, it is now accepted widely

that the efforts to adapt the usability and capability of the desktop PC in to the mobile

environment are limited in their scope. The debate in present literature seems to

focus in particular on the trade-offs and compromises between the performance of

such systems in theoretical or laboratory environments, and the actual results when

tested in the field. This essay will aim to critically evaluate the success of making

context-aware information systems into a feasible reality.

Background

Context-aware computing is a general term used to refer to a class of mobile

computer-systems that are able to sense the context they are being used in- that is

to say, the physical environment in which they are being used in- and can adapt their

behaviour in accordance to this context. The concept came forth from the model of

`ubiquitous computing′ and was first proposed by Schilit et al at Xerox PARC in 1994.

The three main factors in context are:

(1) User location vis-a-vis the physical environment

(2) Position relative to other users

(3) Resources available at that position

Context-aware systems include other persons, devices, and network/power resource

levels etc that exist within that general physical environment (Schilit et al, 1994).

They are also able to adapt their operations to the current context without any

2

explicit inputs from the user - it is desirable in the field of mobile computing that

programs and devices are able to gauge current location, time and other

environmental attributes and react accordingly to the changing circumstances as

context data may change rapidly. The context information that is needed may be

gathered in a variety of ways, including applying sensors, network information,

device status, browsing user profiles and other such sources (Baldauf et al, 2007).

Of course, in designing any context-aware system, it becomes necessary to define

context, for which we turn to Dey & Abowd (2000)- "any information that can be used

to characterize the situation of entities (i.e., whether a person, place or object) that

are considered relevant to the interaction

between a user and an application, including the user and the application

themselves."

The history of context-aware systems started when Want et al. (1992) introduced

their Active Badge Location System which is considered to be one of the first

context-aware applications. Schilit and his team, arguably the pioneers of this field

followed up in 1994 with the PARCTAB, a small hand-held that used infrared-based

cellular networks for communication. The device itself functioned mainly as an

imaging terminal, with most of the computing done by remote hosts.

Literature Review

According to (Baldauf et al, 2007), there are two important considerations when

constructing a context-aware system. These are the architecture and the context

models. Chen (2004) cites three different ways to implement context-aware

architecture, dependent on the specific requirements of the system, the location of

the sensors, the number of users, available system resources, scalability

considerations, etc. The first of these is `direct sensor access′, whereby client

software gathers information from locally-built sensors without an additional layer for

this task. However, since the sensor′s drivers are directly embedded onto the

application it is not suitable for distributed systems. The second approach is termed

`middleware infrastructure′ which introduces a layered architecture to hide low-level

3