See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/221719887

Future Challenges in Context Aware Computing

Conference Paper · October 2007

CITATIONS READS

22 2,560

3 authors, including:

Muhammad Younus Javed

HITEC University, Taxila, Pakistan
326 PUBLICATIONS   3,400 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Organizing Contextual Data in Context Aware Systems: View project

Energy Aware Framework for Smart Environments (EAFSE) View project

All content following this page was uploaded by Umar Mahmud on 16 May 2014.

The user has requested enhancement of the downloaded file.

ISBN: 978-972-8924-44-7 © 2007 IADIS

FUTURE CHALLENGES IN CONTEXT-AWARE

COMPUTING

Nazir Malik
IS Department, Military College of Signals
4 – Humayun Road, Rawalpindi
nazirmalik@mcs.edu.pk

Umar Mahmud
CS Department, Military College of Signals
4 – Humayun Road, Rawalpindi
umar_mahmud@mcs.edu.pk

Younus Javed
Department of CE, College of Electrical and Mechanical Engineering
Peshawar Road, Rawalpindi
myjaved@ceme.edu.pk

ABSTRACT
Context-awareness is the ability of computing systems to acquire and reason about the situational context and adapt
application accordingly. Context-aware system start with gathering of raw, low-level contextual data, interpret the raw
contextual data into high-level interpreted context, reason the interpreted context to derive implications and adapt the
application behaviour on the basis of the implications. This article identifies the research challenges in context-aware
computing and elaborates the context-aware process with the issues in each stage. Some of the current application
domains of context-aware computing are also listed.

KEYWORDS
Context-Awareness, Future Challenges, Technology Strategies.

1. CONTEXT AND CONTEXT-AWARENESS

The ability of smart devices and smart applications to identify current operating conditions or context and
adapt their behaviour on the basis of the context is termed as context awareness. The context information is a
set of parameters defining the surrounding situation that emerge in the moment. The entities involved in a
context setting include the persons, the objects and the computing resources present in the environment (Dey,
2001). Humans being a very complex architecture are quite successful at conveying their thoughts to each
other in an implicit natural way. This is because humans communicate through rich languages as well as
gestures and expressions. Modern ubiquitous computer systems lack an automatic mechanism of inferring
information like humans do. By improving the ability of computers to gather context and infer it, the richness
of communication in human-computer interaction increases that results in a more powerful and more useful
computational environment.

2. RELATED RESEARCH
The pioneer context-aware systems like Active Badge (Want, 1992) and Xerox PARC (Schilit, 2002) are call
forwarding systems that use only location as their context information. Tour guide systems like GUIDE
(Chaverst, 1998) and Cyberguide (Abowd, et al, 1997) evolved the concept of context by adding temporal

306
 IADIS International Conference WWW/Internet 2007

information in addition to spatial information. These systems are primarily context-aware application
designed to provide better and customized services to their users. The Context Toolkit (Dey and Abowd,
2000) provides an Application Programming Interface (API) to develop context-aware applications but is
limited to tightly coupled ‘Widgets’ that access the hardware context sensors.
The recent context-aware systems are context-aware frameworks that facilitate smart service discovery,
delivery and adaptation through rich ontology based context representation. This rich context ontology
considers all parameters relevant to an interaction as the context. Gaia (Roman, et al, 2002) ( a CORBA
based distributed operating system) and CAMUS (Riaz, et al, 2005) a JINI based service oriented
framework) provides context-aware service delivery limited only to context-aware applications. CoBrA
(Chen, et al, 2003) is a mobile agent based framework that dispatches mobile agents to gather context
information from the sensors in the environment. CAPEUS (Samulowitz, et al, 2001) uses a document based
approach that exchanges context-aware packets that describe service requests. All these systems are either
context-based applications or context-aware frameworks that facilitate context adaptation processes. The
context interpretation is carried out as per the requirements of the adaptation process.

3. RESEARCH CHALLENGES OF CONTEXT-AWARENESS

Some of the future research challenges (Winograd, 2001) of context-aware computing are listed in Table 1..
Table 1. Research Challenges

Research Challenge Description

Defining Context The context is a vast concept that encompasses all possible parameters identifying a situation.
The applications and frameworks must define context and identify relevant parameters
limited by their scope.
Context-Aware The context-aware architectures are still in their infancy. Most of the models and architectures
Models are task specific. Standards and support tools are yet to be developed.
Sensing Context Context sensing devices are undergoing development process. Wearable computing is a sub-
Data field of context-aware computing that realizes the design and development of personalized
context sensors.
Predicting Context Lack of sensors requires prediction of context data on the basis of history. Probabilistic
Data techniques like Bayesian models can be used to predict context.
Representing and The context representation scheme should facilitate the context interpretation and context
Storing Context sharing processes and follow a standardized structure.
Information
Inferring Context The interpretation of context and adaptation of service behaviour is a prime challenge of
and Adapting context-awareness. The interpretation process is adaptation oriented.
System Behaviour
Evaluation of Evaluation criteria need to be defined for verification of context-aware systems. Measures for
Context Aware quality control and end-user satisfaction of context-aware products need to be outlined.
Systems
Privacy Control The contextual data of the participating entities is private information and must be secure
from exposure to malicious entities while in transit or at storage device.

4. COMPONENTS OF CONTEXT-AWARE SYSTEM

Figure 1 shows a context-aware system in illustrated form. A context-aware system has the following distinct
components that perform specific tasks.

4.1 Context Acquisition

Any context aware system starts with the sensing of contextual data and subsequent gathering of the
contextual data. Both of these phases may include the exchange of identities of the participating entities and
establishing of secret keys for future communication. The context is sensed through the sensors that are

307
ISBN: 978-972-8924-44-7 © 2007 IADIS

locally built into the mobile devices as well as present in the environment and is gathered by the acquisition
modules of the context-aware systems. The data generated by the sensors is dynamic and is generated in high
quantity.

Figure 1. Components of context-aware system

4.2 Context Representation

The gathered contextual data requires a representation scheme that provides efficient structuring and
retrieval. This gathered data is represented in a standard format so that it assists the sharing process. The
structure of the shared context must be the same across the smart space although it can have different internal
structure. The context space is a hyper space of context data where a single point in the context space is a set
of all attributes of the context data that identifies a situation. The issues in context representation are
discussed is Table 2.
Table 2. Context Representation Issues

Issue Description
Standard The representation scheme should preferably be an industry standard, all-across the smart
Representation space so that it can easily be shared among context-aware systems. Extensible Markup
Schemes language (XML), Resource Description Framework (RDF) and Web Ontology Language
(OWL) are tag-based accepted standard formats for representing contextual data.
Context In case of absence of a standard representation scheme context transformation mechanisms
Transformation should be designed to translate contextual data between the standard formats.
Authentication Identities of the context-aware systems involved in inter-communication need to be
authenticated before starting exchange of data.
Confidentiality Exchange of data needs to be secure from the malicious objects requiring confidentiality.
Communication can be established using either public or private cryptosystems.
Fuzzy Variables The contextual parameters are a set of fuzzy variables and require the use of fuzzy logic.

4.3 Context Storage

The gathered and represented context data is stored on local storage. The storage process allows the system to
maintain a history of context that is used to identify preferences of the entities. The constraints in storage
processes are discussed in Table 3.
Table 3. Constraints in Storage Processes

Constraint Description
Storage The context data is large and requires a sufficient memory and fast retrieval mechanisms to
Requirements achieve memory related efficiency. The dynamic nature of context requires that the context
data be consistent throughout the smart space.
Purging Outdated The outdated and old context data must be purged but its effect should be maintained e.g., a
Context Data month long history of food preference of Alice should not be completely stored although the

308
 IADIS International Conference WWW/Internet 2007

preference should be stored as most recent to beyond a fortnight. Artificial Neural Networks
(ANN) provides a smart way to preserve the preferences as weights.
Predicting Context In the absence of sensor data the context data is predicted through history. Probabilistic
techniques are used to approximate the contextual data of the missing variables.
Encrypted Data Contextual data is stored in encrypted form at local storage areas to avoid exposure to
unauthorized and malicious entities. Encrypted data is shared amongst the participants in
areas of high risk.

4.4 Context Interpretation

Interpreting raw, low-level gathered context to meaningful, high-level interpreted context is the first phase of
the interpretation process. The interpreted context provides us a high-level view of the context data and is
used for adaptation process. The interpreted context consists of subsets of the gathered low-level context
data. Each entity has a number of subsets that identify the context of that entity. The set of interpreted
contexts of all entities in an interaction identify the context of the interaction. In the second phase
implications are then reasoned from the interpreted context. These implications provide instructions for
adaptation process and applications. The interpreted contexts are categorized as ‘What’, ‘Who’, ‘Where’ and
‘When’ contexts. These contexts are the subsets of the gathered contextual data and provide identification,
activity, spatial and temporal information. Some of the research issues in interpretation process are:
approximate interpretation of context; multiple context hypotheses; searching in context space.
The extent of adaptation provides a guideline for the type of interpretation we are interested in. For
example, if the task is to search for the best service available then the interpreted context must be based on
identifying the contexts of the services and the user present in the smart space while it is not feasible to
gather and interpret the context of the entities present in adjacent environments. On the other hand, if the task
is to identify a best route then the context of the intermediate environments on the route should be
interpreted. The issues in interpretation process are discussed in Table 4.
Table 4. Issues in Interpretation Process

Issue Description
User Intention and The user intention and biases can modify the interpretation mechanism. The interpretation
Biases should take into account the user mood and preferences before adaptation. The intention and
preference is predicted through history.
Scope The context-aware systems interpret the context on the basis of their scope. A context-aware
phone needs not to interpret the context of irrelevant parties like all people walking down the
street. Whereas, a context-aware laboratory needs to gather contexts of all people working on
different experiments.
Context Reasoning Interpreted context describes the situation at a high-level of interpretation. This interpretation
Context can be reasoned to infer implications. For example, if the context of a scholar engaged in
research activities is busy for the past two hours then it may imply that the scholar needs a
break and would like to have some coffee with sandwiches. This reasoning provides a basis
for the discovery and the consequent adaptation of services.

4.5 Context Adaptation

The adaptation process starts as soon as the interpretation and reasoning of the contextual data is completed.
The adaptation facilitates the users in different everyday tasks by employing decision making techniques.
Context adaptation can be categorized into the following directions:
4.5.1 Service Discovery
The contextual data is used to discover services for the user. The interpreted context provides the basis for
selecting the appropriate service among available services. The selection process also selects the most
appropriate service in case multiple services of the same type are present in an environment. Services
requiring other services for information, etc. discover the appropriate service on the basis of load and cost as
part of their context preference.

309
 ISBN: 978-972-8924-44-7 © 2007 IADIS

4.5.2 Service Delivery

The selected service is then delivered to the user. The mobile devices mediate directly with the service and
negotiate parameters that include session, cost, paths, protocols, policies and lease times. The delivery
process selects a uniform interface that is compatible with the participating devices.
4.5.3 Service Adaptation
The contextual data is used to adapt the service behaviour as well as the device behaviour. Context adaptation
can be event-based or event-less. Event-based adaptation is carried out when an external event occurs. Event-
less adaptation is carried out in the absence of external event when internal reasoning suggests access and
adaptation of services.

5. CONTEXT AWARE APPLICATION DOMAINS

Context-Aware computing promises to provide a better future in everyday computation. The application
areas are primarily smart spaces where users can interact with each other and traditional systems using their
mobile devices. Some of the context-aware application domains are homes, offices, hotels, community areas,
healthcare, campuses and military applications.

6. CONCLUSION
Context-awareness is an emerging field in the realm of pervasive computing. Potential beneficiaries of
context-awareness range from homes to workplaces and from civil domains to military domains. Context-
aware systems sense, gather, represent and interpret situational context. The interpreted context is used to
provide smart service discovery and delivery and to adapt application behaviour. Context-aware systems
utilize the implicit perception of context found in humans and will revolutionize the way entities interact with
each other.

REFERENCES
Abowd, G.D. et al, 1997, Cyberguide: a mobile Context-Aware Tour Guide. ACM Wireless Networks. Vol. 3, No. 5 pp
421-433.
Chaverst, K. et al, 1998, Design of an Object Model for a Context Sensitive Tourist GUIDE. Proceedings of the
International Workshop on Interactive Applications of Mobile Computing.(IMC 98), Rostock, Germany, pp 883-891.
Chen, H. et al, 2003, An Intelligent Broker for Context-Aware Systems. In Proceedings of UbiComp, Seattle, pp 183-
184. pp. 183-184.
Dey, A.K. and Abowd, G.D., 2000, The Context Toolkit: Aiding the Development of Context-Aware Applications.
Workshop on Software Engineering for Wearable and Pervasive Computing. Pittsburgh, pp. 431-441.
Dey, A.K. 2001, Understanding and Using Context. Personal and Ubiquitous Computing, Vol. 5, No. 1, pp 4 – 7.
Riaz, M. et al, 2005, Service Delivery in Context Aware Environments: Lookup and Access Control Issues. Proceedings
of the 11th IEEE International Conference on Embedded and Real-Time Computing Systems and Applications
(RTCSA'05), Hong Kong, pp 455-458.
Roman, M. et al, 2002, Gaia: A Middleware to Enable Active Spaces. IEEE Pervasive Computing, Vol. 1,pp. 74-83.
Samulowitz, M. et al, 2001, CAPEUS: An Architecture for Context-Aware Selection and Execution of Services. in New
developments in distributed applications and interoperable systems, Netherlands, pp. 23–40.
Schilit, B.N. et al, 2002, Context-Aware Communication. IEEE Wireless Communication, Vol. 5, No. 5, pp 46-54.
Want, R. et al, 1992, The Active Badge Location System. ACM Transactions on Information Systems, Vol. 10, pp 9-102.
Winograd, T., 2001, Architectures for Context. Human-Computer Interaction Journal, pp 401-419.

310

View publication stats