Activity Context Modeling in Context-Aware

International Journal of Computer Applications Technology and Research
Volume 5–Issue 11, 687-692,2016,ISSN:-2319–8656
www.ijcat.com 687
Activity Context Modeling in Context-Aware
Environment
Samuel King Opoku
Computer Science Department,
Technical University of Kumasi
Kumasi, Ghana
Abstract: The explosion of mobile devices has fuelled the advancement of pervasive computing to provide personal assistance in this
information-driven world. Pervasive computing takes advantage of context-aware computing to track, use and adapt to contextual
information. The context that has attracted the attention of many researchers is the activity context. There are six major techniques that
are used to model activity context. These techniques are key-value, logic-based, ontology-based, object-oriented, mark-up schemes and
graphical. This paper analyses these techniques in detail by describing how each technique is implemented while reviewing their pros
and cons. The paper ends with a hybrid modeling method that fits heterogeneous environment while considering the entire of modeling
through data acquisition and utilization stages. The modeling stages of activity context are data sensation, data abstraction and
reasoning and planning. The work revealed that mark-up schemes and object-oriented are best applicable at the data sensation stage.
Key-value and object-oriented techniques fairly support data abstraction stage whereas the logic-based and ontology-based techniques
are the ideal techniques for reasoning and planning stage. In a distributed system, mark-up schemes are very useful in data
communication over a network and graphical technique should be used when saving context data into database.
Keywords: context; context modeling; context-aware computing; hybrid context model; mobile computing; pervasive computing
1. INTRODUCTION
With the proliferation of mobile devices, one looks to
pervasive computing to provide personal assistance in this
digital world. Pervasive computing refers to obtaining
available data any time at any place. One critical aspect of
pervasive computing is context-aware computing in which
applications are made to track, use and adapt to contextual
information. The definition of context by Dey and Abowd [1]
which is also adopted in this work is, “Any information that
can be used to characterize the situation of an entity. An entity
is a person, place or object that is considered relevant to the
interaction between a user and an application including the
user and application themselves”. The relatively important
contexts [2], [3] are location, time, identity and activity. The
first three contexts are obtained through well-known
mechanisms. The location context can be determined when
the device is outdoor through Global Positioning System
(GPS) [4] or indoor through radio frequency technologies like
Bluetooth [5]. Concerning the time of the day, the clock
system of a mobile device can be used in connection with
external authenticated system [6]. However, many techniques
are employed in tracking activity context. This paper reviews
the prominent techniques stating the pros and cons of the
various mechanisms. The paper ends with the best approach
for activity modeling through the various stages of data
acquisition and interpretation.
Context modeling is the process of identifying the appropriate
contextual information of interest and establishing
relationships and reactions among the pieces of contextual
information. Schmohl and Baumgarten in [7] suggested that
context modelling requires two phases: the first phase is to
determine the conceptual abstraction of real world
characteristics. The second phase is to map the concepts on a
context model that represent the information. To sense and
interpret contextual information, many components are
required. The core components are sensors, middleware,
context repository, context reasoning and communication
interface. Context sensor can be hardware or software that
fetches raw data for contextual information. Because the raw
data cannot be interpreted by high level languages [8], there is
the need for a middleware that refines the raw data from
sensors into the required data structures [9]. The context
repository stores the current context based on the data
structures generated by the middleware. Context reasoning
infers new context based on the current context information
obtained from the middleware. The context reasoning relies
on a set of rules [10] to infer the new context. Communication
interface is required in distributed system. Reasoning
mechanism may commit contextual updates into a network or
may request contextual information through a network. The
context reasoning after inference passes the message to the
context-aware application which uses the information to adapt
itself through the execution or termination of certain
programs. A conceptual view of the context utilization
mechanism is illustrated below:
Figure 1 Context Acquisition and Utilization Framework
There are six mechanisms or models that can be used to sense
and interpret a context within the various components stated
above. These are key-value, mark-up scheme, graphical,
object-oriented, logic-based and ontology-based models

www.ijcat.com 688
2. MODELING TECHNIQUES
The six modeling techniques that can be used to model
activity context are discussed in details below. To illustrate
how each model is implemented, the following case study is
model during the discussion of the various techniques. “A
computer science book (BK) is read by computer scientist
(CS). BK has Title, ISBN; Author, Barcode as its attributes.
The title and author are string data type whereas ISBN and
Barcode are integer data type. The author is usually computer
scientist who of course is a human. Humans are characterized
by name and age being string and integer data types
respectively.”
2.1 Key-Value Models
They require simple data structures to associate context
attributes with specific values of conceptual information [7].
Schilit et al in [11] used key-value pairs to model the context
by assigning the value of context information to an
environment variable in an application. From the case study
above, the characteristics of the entities – BK and CS – which
form the keywords for the modeling are:
 BK: title, ISBN, author, barcode
 CS: name, age
Since human and CS are the same in terms of characteristics,
human is ignored in the modeling. The characteristics that can
be used to identify BK are ISBN and barcode. Both features
are of integer data types, though barcode are sometimes
alphanumeric. This work restricts itself to only numeric
barcodes. It is therefore not applicable to alphanumeric
barcodes. There is other integer feature called age that belongs
to CS. However, the length of age should not be more than
three. Barcode has varying length but usually more than three.
Similarly, the length of ISBN is more than three. Hence, every
set of data consisting of numeric characters with length
greater than 3, is associated with BK. The other features of
BK are author and title. They cannot be used in our model
since author and title consist of set of characters similar to the
features of CS. Thus to avoid ambiguity, author and title are
ignored. Similarly, the name feature of CS is also ignored.
Regarding CS, the characteristics the can be used to identify
CA is the age. Thus if the data is integer with length less than
or equal to three, then CS is being referenced. The figure
below gives the key-value model for the above case study
Figure 2 Key-Value Sample Model
There is no relationship established between BK and CS as
discussed above. The characteristics provided should be
followed closely so that entities can be identified and
reasoning inferred. It can be deduced that key-value model is
flexible and easy to manage in a small system. However, the
model limits the amount of data. The model is also application
dependent and is not adaptive. There is no validation support
and relationship modeling.
2.2 Mark-up scheme models
They used hierarchical data structures based on mark-up tags
with attributes and comments. It is based on serialization of
Standard Generic Mark-up Language (SGML), the superclass
of all mark-up languages. Among the commonly used mark-
up schemes are eXensible Mark-up Language (XML) and
Resource Description Framework (RDF). XML is used to
package data or information and RDF is used for conceptual
description or modelling in web resources using a data
serialization format. The above case study is designed using
XML. XML file cannot contain multiple root elements. Thus
each root element – BK, CS and Human – are contained in
separate files. In the figure below, only section of BK is
implemented.
Figure 3 Mark-up Sample Model
Some researchers extended RDF capabilities to Composite
Capabilities/Preferences Profile (CC/PP) [12] and User Agent
Profile (UAProf) [13] which allow the definition and
preferences in context delivery. For instance, in [14] and [15],
the researchers designed Comprehensive Structured Context
Profile (CSCP) and CC/PP Context Extension respectively to
handle the limitation of CC/PP. The limitation of CC/PP is
that it allows specific values only. Another context modelling
approach which does not follow CC/PP is the pervasive
Profile Description Language (PPDL) [16]. This is an XML-
based language for designing interaction patterns on limited
scales. In order to reduce bandwidth consumption, XML can
be combined with JavaScript Object Notation (JSON) [17]
which reduces the transmission time of context-aware
applications. JSON is lightweight data exchange format that
allows computers to generate and parse data easily and faster.
Mark-up schemes are flexible and structured. There are
available tools for processing and very useful when data is
travelled along a network or communication link.
Unfortunately, mark-up schemes depend on the application
and information is hard to extract.
2.3 Graphical models
They represent contextual entities and their relationships
graphically using tools like Uniform Modelling Language
(UML) and Entity Relationships (ER) schemas. UML is very

www.ijcat.com 689
expressive since it uses directed graph to indicate the
relationships between concepts. However, it is difficult to
work with UML [18] due to its increasing complexity. As of
2016, there are more than fourteen different diagrams making
it virtually impossible for developers to recognize. ER on the
other hand describes conceptual pieces as entities and the
interactions as relationships. The major limitation of ER is the
lack of semantics between the entities due to limited
representation of relationships, data manipulation, constraints
and specifications between entities. The above case study is
modeled using ER. The entities in the case study are CS, BK
and Human
Figure 4 Sample Entity Relations Graphical Model
To handle the limitations of ER, extensions are provided by
researchers through the additions of extra features in context
modeling. Bauer in [19] used UML extensions to model air
traffic management system. Another type of graphical model
is the Object-Role Model (ORM) [20] which is a conceptual
level modelling method to handle the semantics of data and its
interrelationship among the data. ORM models concepts as
facts [21]. The work in [22] extended ORM by employing
contextual classification and description properties which
introduced history fact to cover time aspect of the context and
fact dependencies. Thus a change in one fact automatically
leads to a change in another fact. According to Mohan and
Singh in [21], formal semantics of ORM and Context
Modeling Language (CML) can be supplemented to provide
integration with other implementations. Graphical modeling
technique as its strength provides relationship modeling,
flexible implementation and it is very useful for data storage
and historic context store [23]. However, the limitations of
graphical modelling technique include its complexity to
retrieve information which requires obligatory configuration.
It does not support interoperability between heterogeneous
implementations.
2.4 Object-Oriented Models
They employ the use of object-oriented characteristics such as
encapsulation and reusability. Encapsulation hides the
implementation of objects whiles reusability allows models to
be reused. The Active Object Model of the GUIDE project in
[24] was based on object-oriented model. In object-oriented
model, entities are modeled using classes which are
implemented as objects. The attributes are implemented as
data fields or variable in object-oriented model whereas entity
behaviors are modeled as methods (in Java) or functions (in
C++). The above case study is modeled using Java in the
figure below.
Figure 5 Object-Oriented Sample Model
The object-oriented model provides relationship modeling.
The tools for processing are available and it allows every
integration whiles supporting data transformation over
network or communication link. It is however, limited by its
complexity to retrieve information.
2.5 Logic-based Models
They use rules and expressions to define a context. Logics are
used to define conditions for formulating expressions or facts.
It is best used at the reasoning section of a context-aware
application. The first logic based context modelling approach
occurred in 1993 by McCarthy [25] and refined in 1997 [26].
The refined work introduced abstract mathematical entities
complemented with useful properties in artificial intelligence
which allowed simple axioms to be used in common sense
phenomena. Akram and Surav in [27] tried to give theoretical
semantic model of natural language in a formal logic system.
Many researchers had focused on first order logic in their
implementations. Gray and Salber in [28] used it to represent
contextual propositions and relations. In [29], first order logic
was used in connection with Boolean algebra to design
middleware infrastructure called Gaia which allowed various
rules to describe context information. It was used in [30] to
describe context information properties and structure and the
kinds of operations they can perform. Another application is
to use logic based model with other modelling techniques. Gu
et al in [31] proposed Service-Oriented Context-Aware
Middleware (SOCAM) architecture for building context-
aware services. In their model, first-order predicate calculus
was used to model a context and used different modeling
techniques, ontology based model written in Web Ontology
Language (OWL) as a collection of RDF triples, to describe
the context predicate. Using the first order logic, the following
model can be obtained from the above case study:
, (is-isbn(x) ˅ is-barcode(x) => BK (x))
, ( (is-name(x) ˄ is-age(y) => CS (x,y)))

www.ijcat.com 690
, (is-CS(x) => is-Human (x))
, (is-BK(x) => ¬ is-CS (x))
, (is-BK(x) => ¬ is-Human (x))
, ( ((is-author(x) ˄ is-title(x)) ˅ ¬ is-age (y)) =>
BK(x,y))).
The above first order logic can also be described using
fuzzy logic. The figure below illustrates the fuzzy logic
representation.
Figure 6 Fuzzy Logic Sample Model
The strengths of logic-based model are that it generates high-
level context based on low-level context. It is simple to use
and simple to model. Its limitation is the complication of its
applicability. It is also difficult to maintain due to partial
validation
2.6 Ontology-based Models
Ontologies are used to represent concepts and their
interrelationships. This model enables contextual knowledge
sharing and reusability [7]. Ontologies are the most expressive
context representation models [8] that support dynamic
aspects of context awareness. However, they require ontology
engines which have high requirements on resources producing
negative performance impact on local context processing
where resource-constrained devices are employed. Among the
prominent proposal on ontology based modeling techniques
are SOUPA [33] for pervasive environment and CONON [34]
for smart home environment. The figure below illustrates the
ontology model of the above case study.
Figure 5: Ontology Model Sample
The ellipses define the basic classes and every arrow defines a
relation between these entities. The rectangles are added to
show data type values for completion.
OWL-DL is the choice for model context [32] or its
variations since it is supported by a number of reasoning
services. Web Ontology Language – Description Logic
(OWL-DL) allows definition of classes, individuals,
characteristics of individuals and relations between various
individuals. The figure below illustrates the OWL – DL
version for some part of the model.
Figure 6: OWL-DL Model Sample
Generally, ontology based models support semantic
reasoning, provide an easier representation of context and
provide support by standardization. It is however, complex to
retrieved data. Also, it can be deduced from the OWL-DL
model in Figure 6 that it is inadequate in defining complex
context descriptions and thus affects context ontological
reasoning.
3. BEST CHOICE TECHNIQUE
What is the best technique(s) to use to model activity context?
The works of [21] and [35] introduced hybrid approach by
combining different modelling techniques to improve
performance. Perera et al in [36] surveyed high-level context
modelling techniques and concluded that diversification of
modelling techniques is the best way to provide efficient
results which will lessen each other’s weaknesses. Finally,
[23] discussed the context modelling techniques and used
logic based technique to model spatiotemporal context. They
finally concluded that no modelling technique is ideal to be
used in a standalone manner. From the above discussion, it is
therefore laudable to combine modeling techniques to produce
efficient results. Thus the best technique from context
acquisition to utilization is hybrid
There are many stages involved in sensing and interpreting
context. At each stage, it is better to study the strengths and
limitations of each modelling technique and decide the better
technique. For example, at the middleware layer, if the
abstracted context is to be sent via network, it will not be
appropriate to use logic-based modeling, graphical modeling
and key-value modeling techniques. However, mark-up and
object-based modelling techniques can be used depending on
hardware, software and architectural heterogeneity and
interoperability between the communicating devices.
Similarly, it will be ideal to model context reasoning using the
logic-based modeling technique when the application chooses
to use the usual if-then rule set mechanisms. Hence the
dynamic behaviour of the context-aware system cannot be
overlooked when choosing context modeling techniques. The

www.ijcat.com 691
figure below illustrates the appropriate modeling techniques
that can be used at the various modeling stages from which
the desired method can be selected.
Figure 7 Modeling Implementation of Activity Context Life
4. REFERENCES
[1] Dey, A. K., and Abowd, G. D. Towards a better
understanding of context and context-awareness.
Workshop on the What, Who, Where and How of
Context-Awareness, affiliated with the 2000 ACM
Conference on Human Factors in Computing Systems
(CHI 2000), 2000
[2] Dey, A. K. Context-aware computing. Ubiquitous
Computing Fundamentals, 321-352, 2010
[3] Castillejo, E., Almeida, A. & Lopez-de-Ipina, D.
Modelling users, context and devices for adaptive user
interface systems. International journal of pervasive
computing and communications, 10(1), 69-91, 2014
[4] Bajaj, R., Ranaweera S. L., & Agrawal, D. P., GPS:
Location-Tracking Technology. IEEE 35(4), 92–94,
2002
[5] Opoku, S. K., An indoor tracking system based on
Bluetooth technology. Cyber Journals: Multidisciplinary
Journals in Science and Technology, Journal of Selected
Areas in Telecommunications (JSAT), 2(12), 1-8, 2011
[6] Opoku, S. K., An automated biometric attendance
management system with dual authentication mechanism
based on Bluetooth and NFC technologies. International
Journal of Computer Science and Mobile Computing
2(3), 18-25, 2013
[7] Schmohl, R. & Baumgarten, U., Context-aware
computing: a survey preparing for generalized approach.
Proceedings of the international multi-conference of
engineers and computer scientists, UMECS, 1(3), 19-21,
2008
[8] Strang, T. & Linnhoff-Popien, C (2004). A context
modeling survey. In Proceedings of the Workshop on
Advanced Context Modelling, Reasoning and
Management associated with the 6th International
Conference on Ubiquitous Computing (UbiComp),
Nottingham.
[9] Gellersen, H. W., Schmidt, A. & Beigl, M., Multi-sensor
context-awareness in mobile devices and smart artifacts.
Mobile Networks Applications, 7(5):341–351, 2002
[10] Jacob, C., Linner, D, Radusch, I. & Steglich, S. (2007).
Loosely coupled and context-aware service provision
incorporating the quality of rules. In ICOMP 07:
Proceedings of the 2007International Conference on
Internet Computing. CSREA Press, 2007
[11] Schilit, B, Adams, N., R. Want, et al. (1994). Context-
aware Computing Applications, Xerox Corp., Palo Alto
Research Center.
[12] W3C. Composite Capabilities / Preferences Profile
(CC/PP). http://www.w3.org/Mo-bile/CCPP.
[13] Wapforum. User Agent Profile (UAProf).
http://www.wapforum.org.
[14] Held, A., Buchholz, S. & Sachill, A. (2002) Modeling of
context information for pervasive computing
applications, In Proceedings of SCI 2002/ISAS 2002
[15] Indulska, J. & Robinsona, R. (2003). Experiences in
using cc/pp in context-aware systems, In LNCS 2574:
Proceedings of the 4th International Conference on
Mobile Data Management (MDM2003)
(Melbourne/Australia, January 2003), M.S. Chen,P. K.
Chrysanthis, M. Sloman, and A. Zaslavsky, Eds.,Lecture
Notes in Computer Science (LNCS), Springer, pp. 247–
261.
[16] Chtcherbina, E. & Franz, M. (2003). Peer-to-peer
coordination framework (p2pc): Enabler of mobile ad-
hoc networking for medicine, business, and
entertainment, In Proceedings of International
Conference on Advances in Infrastructure for Electronic
Business, Education, Science, Medicine, and Mobile
Technologies on the Internet (SSGRR2003w)
(L’Aquila/Italy, January 2003).
[17] Erfianto B, Mahmood, A. K. & Rahman, A. S. A. (2007,
December 11-12). Modelling context and exchange
format for context-aware computing. Proceedings of 5th
student conference on research and development.
Malaysia
[18] Henricksen, K., Indulska, J., Modeling Context
Information in Pervasive Computing Systems. 167–180,
2002
[19] Bauer, J. (2003, March). Identification and Modeling of
Contexts for Different Information Scenarios in Air
Traffic. Diplomarbeit
[20] Halpin, T. A., (2001). Information Modeling and
Relational Databases: From Conceptual Analysis to
Logical Design, Morgan Kaufman Publishers, San
Francisco.
[21] Mohan, P. & Singh, M. (2013). Formal models for
context-aware computing. International Journal of
Computer Applications Technology and Research
(IJCATR), 2(1), 53-58
[22] Henricksen, K., Indulska, J., and Rakotonirainy, A.
(2003). Generating Context Management Infrastructure
from High-Level Context Models‖, In Industrial Track
Proceedings of the 4th International Conference on
Mobile Data Management (MDM2003)
(Melbourne/Australia, January 2003), pp. 1–6.
[23] Ameyed, D., Miraouli, M. & Tadj, C. (2016).
Spatiotemoral context modelling in pervasive context-
aware computing environment: A logic perspective.

www.ijcat.com 692
International Journal of Advanced Computer Science and
Applications, 7(4), 407-414
[24] Cheverst, K., Mitchell, K. & Davies, N. (1999). Design
of an object model for a context sensitive tourist GUIDE,
Computers and Graphics 23(6), 883–891
[25] McCarthy, J. (1993). Notes on formalizing contexts, In
Proceedings of the Thirteenth International Joint
Conference on Artificial Intelligence (San Mateo,
California, 1993), R. Bajcsy, Ed., Morgan Kaufmann, pp.
555–560.
[26] Mccarthy, J. & Buva. (1997). Formalizing context. In
Working Papers of the AAAI Fall Symposium on
Context in Knowledge Representation and Natural
Language (Menlo Park, California, 1997), American
Association for Artificial Intelligence, pp. 99–135.
[27] Akram V. & Surav, M. (1997). The use of situation
theory in context modelling. Computational intelligence,
13, 427-438
[28] Gray, P. & Salber, D. (2001). Modelling and using
sensed context information in the design of interactive
applications, in Engineering for Human-Computer
Interaction, ed: Springer, 317-335.
[29] Roman, M., Hess, C., Cerqueira, R., Ranganathan, A.,
Campbell, R. H., & Nahrstedt, k. (2002). A middleware
infrastructure for active space. IEEE pervasive
computing, 1(4), 74-83
[30] Ranganathan, A. & Campbell, R. H. (2003). An
infrastructure for context-awareness based on first order
logic. Personal and Ubiquitous Computing, 7, 353-364
[31] Gu, T., Pung, H. K. & Zhang, D. Q. (2005). A service-
oriented middleware for building context-aware services.
Journal of Network and computer applications, 28, 1-18
[32] Horrocks, I., Patel-Schneider, P.F., F. van Harmelen,
(2003). From SHIQ and RDF to OWL: The making of a
web ontology language, Journal of Web Semantics, 1 (1),
7–26.
[33] Chen, H., Perich, F., Finin, T. W., Joshi, A. (2004).
SOUPA: Standard Ontology for Ubiquitous and
Pervasive Applications, in: 1st Annual International
Conference on Mobile and Ubiquitous Systems
(MobiQuitous 2004), IEEE Computer Society
[34] Zhang, D., Gu, T., Wang, X. (2005). Enabling Context-
aware Smart Home with Semantic Technology,
International Journal of Human-friendly Welfare Robotic
Systems, 6 (4), 12–20
[35] Bettini, C., Brdiczka, O., Henricksen, K., Indulska, J.,
Nicklas, D., Ranganathan, A. et al. (2010). A survey of
context modelling and reasoning techniques. Pervasive
and Mobile Computing. 6, 161-180
[36] Perera, C., Zaslavsky, A., Christen, P. and
Georgakopoulos, D. (2014). Context aware computing
for the internet of things: A survey. Communications
Surveys & Tutorials, IEEE, 16, 414-454

Activity Context Modeling in Context-Aware

More Related Content

Activity Context Modeling in Context-Aware