Towards a Semantic-based Context-as-a-Service for Internet of Things

Towards a Semantic-based Context-as-a-Service
For Internet of Things
Manar Elkady
Dept. of Computer Science,
Faculty of Computers and
Information, Cairo University
Cairo, Egypt
m.elkady@fci-cu.edu.eg
Abeer M. El-Korany
Cairo, Egypt
a.korani@fci-cu.edu.eg
Reem Bahgat
Cairo, Egypt
r.bahgat@fci-cu.edu.eg
Abstract— there is an obvious trend toward the future Internet
environment, where billions of sensors are connected in a highly
dynamic environment to provide fine-grained information about
the physical world. Applications that utilize such information are
smart and belong to the Internet of Things (IOT) paradigm.
There is a need for IOT systems to have a higher understanding
of the situations in which to provide services or functionalities, to
adapt accordingly. Information context should be collected,
modeled, inferred, and distributed to serve IOT applications with
the required knowledge. Developing context-awareness IOT
applications demands a specialized framework that helps in
applying computations transparently on the cloud and feed the
context-awareness IOT applications with appropriate context
information to prevent the direct access of the context sources.
In this paper, a framework for developing context-aware IOT
applications is proposed. The framework is distributed,
autonomous, and adaptive to fit the IOT applications’
characteristics. The proposed framework utilizes agent’s
technology as well as semantic-based features to provide the basic
functionality required to develop a context-awareness IOT
application. A scenario for elder healthcare is explained to
evaluate this framework.
Keywords— Context-as-a-Service, Internet of Things, Context-
awareness.
I. INTRODUCTION
Nowadays, an obvious trend toward the Future Internet
environment drives different applications to require access to
Internet with the delivery of real time information generated
from a variety of heterogeneous devices. As a result, Internet
of Things (IOT) [1] has appeared as an important step in the
evolution of the Internet. IOT is targeting using such devices
to acquire different types of information, store, and process it,
in order to be able to use it later. Accordingly, several IOT
applications have emerged and then they need specific
characteristics relevant to their users that could be used to
describe the circumstance concerning an entity (which may be
person, place or object). For instance, a context-aware air-
conditioner IOT application could change the air-conditioner's
fan speed or its temperature based on the number of people
present within a room at some time. These characteristics are
normally referred to as application context [2]. We refer to
IOT applications that rely on their context information to react
and adapt its current state as context-awareness IOT
applications. While developing such kind of applications,
supporting both of the IOT applications characteristics in
addition to handling the context life cycle phases are our
target. The main IOT applications characteristics are
distributed, where it require scalability beyond millions of
devices over the Internet, adaptive according to the changes in
the surroundings, and autonomous by applying actions based
on its internal state, goal, and environment parameters.
Previous work targeting context-awareness and IOT
applications failed to prove complete coverage of their
characteristics. Each one of the previous trial is concerning set
of features and neglect the others, which motivate us to build
our own framework. For example, research efforts in [3] [4]
[5] have a shortage in applying semantics modeling and
reasoning to generate emergent context information. In
addition, another research effort concerned with a special
application type and its implementation details rather than
building a general purpose context-as-a-service architecture.
The main idea of the proposed framework is to offer a high
level abstraction to IOT applications’ developers to access the
context information without going into low-level details of
access the context sources. The proposed framework is
International Journal of Computer Science and Information Security (IJCSIS),
Vol. 15, No. 8, August 2017
284 https://sites.google.com/site/ijcsis/
ISSN 1947-5500

supporting context-awareness by understanding the
surrounding context resources, and automatically building
internal data models to process and reason to get knowledge,
which enrich the smart property for the IOT applications and
services. The context management solution lies outside the
application development boundaries. Such solution is
developed and extended independently. Thus, context-
awareness is no longer limited to desktop, web, or mobile
applications. It has already become a service: Context-as-a-
Service [6] [7].
We get a benefit from the well-established technologies;
the Cloud [8]; Agents [9]; and Ontologies. Handling the
distributed execution of the IOT application and offering a
uniform interface to access a massive amount of data
transparently are guaranteed by developing the proposed
framework as a service on the Cloud guarantees. Therefore,
delivering the context information as service through the
Cloud supports usability, reusability, and unification
properties. Moreover, applying the Agent technology inside
the framework design enrich distribution, autonomy, and
adaptability properties. An agent could be thought of as a
software component which not only encapsulates code and
data as in object-oriented programming, but can also be pro-
active; i.e. can take the initiative in interacting with its
environment, or reactive; responding to changes in its
environment. An agent exhibits autonomy in that it controls its
own actions based on its beliefs and surroundings. Adaptation
is another attribute exhibited by agents, where an agent
improves its plan to achieving goals and learns from
experience. Ontology is used as tool to enrich the semantics in
the proposed framework. Ontology is used in Knowledge
representation and emerged as an approach for context
modeling in a higher semantic level to support sharing and
reusing both of data and services.
This paper is organized as follow: Section II presents some
related works with an illustration for their points of strength
and weakness. Section III presents the background concepts of
this work; Internet of Things; and context-awareness
fundamentals. Section IV presents the proposed context-
awareness framework and its main characteristics. A detailed
description for the context-as-a-service layer is shown in
section V. Section VI presents the Ontologies used for context
information modeling. The paper is concluded in Section VII.
II. RELATED WORK
In this section, we briefly discuss the previous research
efforts for developing context-aware IOT applications
illustrating the variety and the shortage in their design
perspectives.
Some of the related approaches for building programming
models that targeting context-awareness applications are
focusing on pervasive and IOT systems [3] [4] [5] . The
Origins Model [3] provides a programming model for the
development of context-aware applications on large-scale
pervasive systems. It consists of the Origin entity which
represents an abstraction for different types of context sources.
It is associated with a type and meta-information. The Origin
Model has a set of elementary operations; filter; infer;
aggregate, which used for achieving context information. It
depends on its predefined properties of replication and
mobility to enable creating distributed and reliable processing
schemes and represented different types of context sources. It
apply the concepts of actors and futures in the implementation
to support scalability. Then a CAPA framework [4] is
designed to build a programming model for developing
context-awareness applications for distributed pervasive
systems based on the Origin model. It utilizes the Origin
model [3] to acquire context information and accessed by its
Entity programming model. The Entity model is used for
developing generic and reusable application components by
decoupling the application layer and the physical layers. It
supports the monitoring and controlling behaviors for the large
scale heterogeneous context sources. PaTRICIA [5] provides a
different construct for its programming model; Intent (explicit
representation of tasks) and IntentScope (group of physical
entities). PaRICIA provides cloud-based IOT application
execution environment and the supporting runtime
mechanisms. It supports two logical views for both of the
domain expert to define the domain libraries and Developer to
use the programming model to write cloud-scale IOT
applications. Although the frameworks structure for these
works originate context information outside of the control of
context-aware applications, it lacks form the clear
identification of the full context life cycle phases and specially
the context modeling. It lacks from the semantics modeling
and reasoning to generate new context information to context-
awareness applications autonomously.
Running context-awareness IOT application on the cloud
guaranteed the execution of IOT main requirements of
handling the distributed and massive amount of data through a
uniform interface. Therefore Applying the context
management in the form of Context-as-a-Service has a wide
acceptance by researchers [10] [7] [11]. Authors of [11]
design a prototype implementation to deliver the context
information forms physical devices to its digital from for IOT
Vol. 15, No. 8, Augus 2017
ISSN 1947-5500

applications. Their design concerning application utilization
for the context-aware services that help older people stay at
their homes safely. The design main components are Event
Processing Unit, which is responsible for detect and aggregate
events acquired from the devices. Data Repository is defined
to record things status to database. Such contextual events can
be directly used by high-level applications or further
processed by other modules. For process management, the
defined Rule Engine and Service container provide flexible
and full spectrum support for process management and
customization. Here, the design is concerned with a special
application type and its implementation details rather than
building a general purpose context-as-a-service architecture.
Another approach [7] targets a design for a general-purpose
architecture and techniques for the contextualization of IOT
data. The design aims to reducing complexity of data analysis,
query, and extract valuable knowledge. Data is represented in
RDF format and processed through three operations; filter;
aggregate; and infer before accessed by the application.
Although the scalability and a distributed solution applied
through the implementation by utilizing the elastic cloud-
based infrastructure, it lacks to guarantee scalability at design
time. In addition it has a limitation in applying complex
reasoning for more autonomous and adaptation.
Applying complex reasoning as a part of the architecture
design can be guaranteed by adding self-organizing and self-
adaptive components that encapsulate such features to the
architecture. Research work such as [12] [13] utilizes the
agent technology in their design for this purpose. Authors of
[12] proposed an agent-based model to generate IOT
applications. It utilized the multi-agents system characteristics
for applying autonomous and adaptive behaviors. Their agent-
model is tightly coupled with the IOT application, which
exceeds the developer effort to be aware of the agent model
details.
Although authors in [13] do not exclude the human
intervention for ontologies matching and alignment, they do
not determine a detailed specification for such intervention.
Moreover, no overall framework illustrating the data flow
from the physical resources till the application that interested
in such data.
III. GENERAL CONCEPTS
A. Internet of Things
“The Internet of Things allows people and things to be
connected anytime, anyplace, with anything and anyone,
ideally using any path/network and Any Service.” is widely
accepted definition for the IOT. It promises to create a world
where all sensors around human are connected to the Internet
and communicate with each other with minimum human
intervention [14].
The IOT comprises a large number of objects (sensors and
actuators) that interact autonomously. Moreover, real-time
data processing is essential. When the number of objects gets
larger, tracking becomes a key requirement for IOT
applications. In addition, Interactions are highly dependent on
their locations, their surroundings, and presence of other
entities (e.g. objects and people). Due to the popularity of
cloud computing [8], consuming resources as a service has
become a mainstream. Everything-as-a-service model is
highly efficient, scalable, and easy to use. These
characteristics need to be considered when developing IOT
context-awareness solutions throughout all the phases from
design, development, implement and evaluation [15].
B. Context-awareness Fundamentals
Context-awareness provides the facility to store context
information which will be used in order to adapt the
application’s current situation; therefore it is essential for
systems developed in areas that have intelligent behaviors [8].
Thus, developing a context-aware framework that is easily
integrated within any IOT application allows the developer to
develop autonomous and adaptive applications.
The Context can be defined as 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 applications themselves [16]. Therefore, not all data
can be considered as a context. Raw (sensor) data is
unprocessed and retrieved directly from the data source, such
as sensors, but the context information is generated by
processing raw sensor data. Further, it is checked for
consistency and metadata is added. Context-aware system uses
context to provide relevant information and/or services to the
user, where relevancy depends on the user's task [16].
Context-awareness frameworks should support the context life
cycle stages; acquisition; modeling; reasoning; and
dissemination [15] [17] survey] as follow:
● Context acquisition: The techniques used to acquire
context can be varied based on responsibility,
frequency, context source, sensor type, and acquisition
process.
● Context Modeling: Acquired data is converted into a
unified format such that the context can be understood
and shared. This can be achieved by a model that
Vol. 15, No. 8, Augus 2017
ISSN 1947-5500

defines, represents and processes the context. The six
most popular context modelling techniques: key-value,
markup schemes, graphical, object based, logic based,
and ontology based modelling. There are many
surveys that describe context modeling techniques in
details such as [8] [17] [18] [19]. According to [19],
using ontology for context modeling would satisfy
both of the normalization and formality of the model.
● Context Reasoning: It is a method of deducing new
knowledge, and understanding better, based on the
available context and by applying reasoning
techniques.
● Context Dissemination: It is the process of context
distribution to the consumers who are interested in that
context such as applications, services, or users.
IV. CONTEXT-AS-A-SERVICE FRAMEWORK
The basic idea of the proposed framework is to offer a
structured solution for developing context-awareness IOT
application. The proposed framework is based on delivering
the context information as a service to the IOT application as
shown in Fig. 1. The framework presents a unified interface
for context-awareness IOT applications to access its context
information through the context-as-service layer.
The proposed framework is defined over three different
Layers; Physical layer; Context-as-a-Services layer; and
Application layer. The Application Layer is the layer that
contains the high level constructs for developing context-
awareness IOT applications based on a domain-based
ontology defined by the application domain experts and
context information delivered by the Context-as-a-Service
Layer. The Context-as-a-Service Layer is responsible for
processing data achieved from data sources that belongs the
physical layer and taking decisions to generate the context
information. The generated context information is the target
knowledge for the context-awareness IOT application. The
physical layer has set of sensors and actuators. The sensors are
used to collect data, which is raw data before the processing
and decisions are made. Finally, actuators perform the decided
actions.
A. Physical Layer
The Physical Layer represents the space of the context
sources. Context sources can be categorized as physical,
logical, or virtual sources [15]. Physical sources are the type
of sensors that generate data by themselves such as the
temperature sensors and light sensors. Virtual sensors retrieve
data from many sources and publish it as sensor data (e.g.
calendar, contact number directory, twitter statuses, email and
chat applications). Logical sensors (also called software
sensors) combine physical sensors and virtual sensors in order
to produce more meaningful information. A web service
dedicated to providing weather information is an example for
a logical sensor.
B. Context-as-a-Service Layer
The Context-as-a-Service Layer encapsulates all the
context life cycle phases starting by context acquisition from
sensors in the physical layer till the context delivery to
context-awareness IOT applications in the application layer.
Massive processing steps are done by Agents for validation,
modeling, and reasoning purpose. The Context-as-a-Service
layer hosts a set of agents organized in an agent model. Each
agent type is responsible to handle a certain phase of the
context life cycle. Agents’ types descriptions are summarized
in Table I. Ontologies are used by agents to solve the
integration issues for different types of domain knowledge as
well as for modeling purpose. More details about agent model
and ontologies used in the framework are described in section
V and section VI.
C. Application Layer
The Application Layer is the high-level developing space
for the context-awareness IOT applications. It includes a
programming model API for context information access, and
application domain ontology for building semantic based
applications.
The programming model for such type of large-scale
applications encapsulates the application behaviors and
increases the abstraction Level. It offers an interface used to
access the context information, which delivered by the
Context-as-a-Service layer. Application developer has the
ability to access the relevant context information without
going into the low level details of neither dealing with context
sources nor programming for context processing. The context
information is a guide for changing the application behavior
during the run time.
Domain ontology is used to enhance the explicit semantic
of the context information based on different domains. It is
built by each domain expert and used by the application
developer to enrich the context information semantics. The
application layer detailed design and API are not our focus in
this paper.
Vol. 15, No. 8, Augus 2017
ISSN 1947-5500

Fig. 1. Context-as-a-Service framework for developing context-awareness Internet of Thing (IOT) applications.
TABLE I AGENTS’ TYPES DESCRIPTION
Goal Behavior Believes
Acquisition
Agent
Obtain a maximum amount of data Pull, Push, Manually_provided,
Derive
Context sources meta- information
Knowledge about the required data. .
Modeling
Agent
Define context model
Store context model in a machine
readable and processable form
Generate_ context_model, Aggregate
Validate_model, Store_context
Context representation structure
Model validation_mechanism
Reasoning
Agent
Efficient Context reasoning and
Validation.
Check for Quality of Context
Clean_data, Derive, Infer,
Validate_context
Reasoning Techniques
Validation rules.
Cleaning data parameters
Dissemination
Agent
Reasoning Techniques
Validation rules.
Cleaning data parameters
Query, Subscribe, Deliver, Execute,
Manual_add
Context delivery mechanism
Vol. 15, No. 8, Augus 2017
ISSN 1947-5500

V. CONTEXT-AS-A-SERVICE LAYER
The Context-as-a-Service Layer is responsible to deliver
the context information to the IOT application. Agents are
playing different roles to accomplish this goal. The agents’
roles are classified as Acquisition Agent, Modeling Agent,
Reasoning Agent, and Dissemination Agent as illustrated in
Fig. 1. The flow of work through this layer starts by an
initiation from the context-awareness IOT application that
asks for specific context information. The Dissemination
Agent receives the application request of context information
and then triggers the Acquisition Agent for fetching the
required information. After finishing the acquisition process,
the Acquisition Agent acknowledges the Dissemination Agent
for finishing its task, and then the later triggers the Modeling
Agent to start its job. The Modeling Agent converts the
unprocessed data into a unified format to enable sharing and
understanding. After that, it stores the context information in
the Context Information repository, and sends a confirmation
message to the Dissemination Agent informing it about
finishing its job. After this phase, the context information is
ready to be delivered to the application if it is required in this
form. Otherwise, a reasoning phase is started for extra
knowledge generation. And then, the Dissemination Agent
initiates the Reasoning Agent to start the reasoning process.
Therefore, new entries are added in the repository that can be
delivered to the application by the Dissemination Agent. Data
Processing and Analysis component is used by agents through
the four phases according to the need of each phase and the
preferences determined through the context information
requests. Rule Engine is the storage for the validation rules
and constraints needed while modeling and reasoning process
or context management rules that used through the four
phases.
An agent has its own belief that could be changed over its
lifetime according to its inference strategy. It also has goals,
behaviors, and access permissions to the available resources.
A behavior specifies the actions that an agent can do through
the interaction with its environment to affect and be affected
by it. It can be either a perception behavior or an effector one.
Beliefs represent the agent’s knowledge about its environment,
other agents, and itself. The agent’s background knowledge is
assigned to it as it is created and becomes part of its mental
state. Then it can be changed during its life due to the agent’s
perception or its communication with other agents or with the
environment. A Goal is a strategic target that an agent aims to
achieve through executing a plan. This is in case the agent is
proactive, but if it is only reactive, so it does not need a goal.
An agent can deal with ontologies defined in different
namespaces.
Acquisition Agent is responsible for acquiring data from
different types of context resources exists in the physical layer
either a physical, virtual, or logical sensors. Data is acquired
from different context sources such as servers, sensors,
middleware, web service, files, or database. In the IOT
paradigm, context can be generated based on two different
event types: instant events and interval events. Open the door,
switch on the light are example to the instant event which
needs to be acquired when the event occurs. Sense and send
data to the software every 20 seconds is an example for the
interval event, which span a certain period of time.
Acquisition agent adapts its behavior to acquire data
accordingly to context sources meta-information. Data can be
sensed from sensors, derived by using computational
operations on the sensed data, or manually provided by users
via predefined settings options such as user preferences.
Accordingly, two main sensed behaviors are assigned to
sensing function; Pull; and Push. In Pull behavior, the
acquisition agent makes a request from any physical device
(such as sensors) periodically or instantly to acquire data.
While in Push behavior, the physical or virtual device pushes
data to the agent’s input buffer data periodically or instantly.
Periodical or instant pushing can be employed to facilitate a
“publish and subscribe” model. Additional effort is needed
here for refining different type of acquired data to extract the
target context information. The acquired information is
buffered to the modeling agent. Applying Agent as an
intelligent component will enrich the context acquisition with
richer context information.
Modeling Agent is responsible for context modeling and
context organization. Regarding the context modeling, the
agent converts the acquired data into a unified format to
enable sharing and understanding. The generated context
model defines, and represents the context information. The
technique used in context modeling affects the possibility and
efficiency of the reasoning process. Semantic-based modeling
using ontology is our choice for context modeling. Ontology is
competitive technique over other models in terms of
interoperability, and reusability [20]. IOT_Context_Ontology
that used in the proposed framework refers to set of ontologies
connected with each other to define both of the IOT paradigm
concepts, and the context types and categories. A detailed
description for the IOT_Context_Ontology is in section VI.
Regarding the context model organization, Modeling Agent
validates the context model for consistency, and stores them in
Vol. 15, No. 8, Augus 2017
ISSN 1947-5500

repository. Applying Agents in context modeling process
enrich dealing with the dynamic contexts [17]. For example, in
smart city application, the smart traffic control has a highly
dynamic context, where the environment is frequently
changed and it is not the case of only having predefined
context information that will be collected and stored.
Reasoning Agent is responsible for selecting and applying
the reasoning technique to deduce knowledge from the
available context information. First, cleaning unnecessary
data, merge the related context is applied. Validation is
applied against validation rules or set of constraints that are
stored in the Rule Engine. Also, sensor data fusion can be
applied by the reasoning agent for combining sensor data from
multiple sensors. Finally, inferring new knowledge by
applying the reasoning technique is occurred. Applying
Agents in context reasoning phase get benefit from agent’s
learning and adaptive characteristics that can enrich the design
of the context reasoning component. As an example from
[15], if we have data represented by a tuple <Who: John,
What: walking:4km/h, Where: ANU, Canberra, When: 2013-
01- 05:9.30am). Inference can generate that PurplePickle cafe
in canberra, PurplePickle cafe in canberra, and may be
inferred as John’s favourite cafe.
Dissemination Agent supports a generic interface to be
accessed by context-aware IOT applications and provides
methods to deliver context information to the applications. A
set of essential behaviors is supported by agent such as: Query
behavior, which is used by the context IOT application to ask
for the required context information. Publish and Deliver
behaviors are used to send adaptive context information to the
application layer periodically or when an event occurs.
Applications ask to apply subscribe behavior to receive
notifications about certain context service, which is used in
real time processing. Dissemination Agent is responsible for
disseminating useful and relevant context information to
corresponding applications.
VI. CONTEXT-IOT ONTOLOGY
Ontology-based modeling is the way towards
standardization and unification of context-awareness concepts.
Although, there are well accepted ontologies used to develop
applications and services for the IOT [20], Ontologies should
be extended by users to meet the evolving context-awareness
IOT application. In this case the extended ontology can be
used by user only to prevent the conflicts. It allows context-
aware applications to further filter relevant information from
irrelevant ones. IOT_Context_Ontology presented in our
framework should interpret both of the IOT paradigm and
context-awareness concepts. Context Categorization should be
presented from both of the operational perspective, and
Conceptual perspective. From Operational perspective
categorize context based on how they were acquired,
modelled, and treated. From Conceptual perspective: context
is defined based on the meaning and conceptual relationship
between the context location, identity, time, and activity.
Thus, the main concepts of Context IOT ontology consist of:
Context, User, location, and time. Each of those concepts
could be mapped with one of the general ontology such as
Friend of a Friend (FOAF) [21] for users, GeoNames [22] for
location, and Time Ontology in OWL [23]. Regarding
defining context concept is an entity that represents the
context information in our model. Context Information has set
of attributes that should be modeled to enrich the context
delivery process:
● Type: is used to describe the context information
provided.
● Context_Quality: It is a measure to define quality
criteria that the context should satisfy. It is important
to assign a sensitive means by which to judge the
reliability of information to support the quality of
context [19] as well as other measurements which
differ based on type of the context. E.g. the quality of
the user’s location is measured by the standard error of
the location system.
● Context_dependence: the association with other related
contexts. This property is important in context-
awareness systems to provide updated and relevant
information when the dependencies are changed.
● Context_Timestamp: the time of the acquired context. It
is useful in the inference to get a conclusion according
to the time of the context.
● Context_AccessRight: this relationship is used to
identify who has the privilege to access this context
information.
● Context_source: this relationship is used to identify
sensed, user_defined, derived.
VII. APPLICATION SCENARIO
Humans, data, communication and devices play key roles
in the IOT ecosystem [24]. IOT paradigm contributes in
monitoring the progress of a point wherein each physical
Vol. 15, No. 8, Augus 2017
ISSN 1947-5500

“Thing” is a digital shadow of the person using its services,
just as the social network era [25]. Healthcare applications
were based on tracking systems for monitoring medical assets,
extracting the location footprints of patients and doctors and
others [24]. Such applications should rely on the advanced
devices and technology such as smartphones and
[24]smartwatches to meet the needs of its users.
Elder Healthcare
Elder Healthcare scenario is inspired from [26]. Elder
people are alone for long time usually. Their relatives want to
know their health indicators. Some scenarios can indicate their
health state. Indicators for the daily performance are steps
walked a day; rooms changed a day, Locations visited outside,
Time the TV set was running. Other situations that trigger the
alarm such as falling alert (resident on the floor for more than
a certain amount of time), Irregular movement pattern (fast
movement between rooms, or standing up and sitting down
repeatedly in a short time), Lack of movement, Active alert
(resident has pushed the alert button), Dramatic decrease of
steps for a longer period. There are some status that
indicate a person may be in danger such as: Room doesn't
change anymore and person is at home, TV is not used at
all and the person is in home, location outside doesn't
change for long and it is night, and location outside and bad
weather (e.g. very cold).
In order to apply the proposed framework in elder
healthcare domain, a set of devices is assigned in the physical
layer such as: watch geer, mobile phone, ibeacons to locate
current room, accelerometer to count steps, GPS to
locate person outside, and Powermeter to detect TV usage.
Flip button is used to send alarm.
A set of application rules is defined to guide through the
context life cycle phases. Such rules are defined by the
application developer and imported to the Context-as-a-
Service layer through the programming model API. Examples
for these rules are as the following:
● if count_flic_pressed >5 within (5 min) call
emergency
● if (step_aggregated < min_Normal) Then
silence_period
● if (silence_period > max_Silence_Level) call
emergency
With the mentioned scenario, the application will have the
following features:
● Check periodically the health status of an elder person
● Generate an alarm in case of danger
● Generate alternative scenarios based on the current
situation
● Alternative reactions in case of no response for the
danger alarm
The context-awareness IOT application subscribes at the
Context-as-a-Service layer to be informed about the elder
health state (the target Context Information). Such
subscription and publication of the elder state is handled by
the Dissemination Agent. During the period from the
subscription till the publication of the elder state, the
Dissemination Agent coordinates with other agents to achieve
a full detail about the elder state. Moreover, Notifications
about dangers and suggested actions are generated through the
reasoning process by the Reasoning Agent. For an example to
get the elder location as a part of his/her state, the Acquisition
Agent applies the pull behavior to fetch the elder coordination.
The Modeling Agent converts this coordination to a standard
format (e.g. OWL, or RDF format). The Reasoning Agent
applies the Drive behavior to the elder coordination to get
his/her location by using the GeoNames ontology. A sample
for agents’ behaviors for elder healthcare application is
illustrated in Table II.
VIII. CONCLUSIONS AND FUTURE WORK
Delivering the real-time context information as a service to
context-awareness IOT applications has an obvious impact on
improving the efficiency of information processing and
delivery of right information to right person at a right time. In
this paper, a framework for context-awareness IOT
applications is presented with a focus on the Context-as-a-
Service layer. The design is light-weight agent-based model
supporting the scalability, reliability, unification, and has an
autonomous, adaptive behavior based on a semantic reasoning
mechanism.
In future work, programming model module to access the
context-as-a-service module will be presented. A complete
case scenario to illustrate this model is going to be
implemented by using Akka’s actor model[27], which
communicate via messaging between actors in distributed
virtual machines and JADE (Java Agent DEvelopment
Framework) [28].
Vol. 15, No. 8, Augus 2017
ISSN 1947-5500

TABLE II SAMPLE FOR AGENTS’ BEHAVIORS FOR ELDER HEALTHCARE
APPLICATION
Function Agent Behavior
Validate indoor
location
Reasoning Agent validate(coordination)
Receive alarm from
Flic button
Acquisition Agent push (alarm)
Aggregate the elder
steps
Modeling Agent aggregate (steps)
Call a contact person Dissemination
Agent
Execute (phone_call)
Publish data about the
current situation
Dissemination
Agent
subscribe (elder_state)
publish (elder_state)
Detect silence period Reasoning Agent infer(silence_period)
Generate suggestion for
the current indicators
Reasoning Agent infer()
Detect GPS and
Beacons location
Acquisition Agent
Modeling Agent
Reasoning Agent
pull (coordination)
store (location)
derive (coordination,
location) - GeoNames
Ontology
infer(elder_state)
REFERENCES
[1] P. Rodrigues, Y.-D. Bromberg, L. Réveillère, and D. Négru, “ZigZag: A
Middleware for Service Discovery in Future Internet,” in Distributed
Applications and Interoperable Systems, vol. 7272, K. M. Göschka and
S. Haridi, Eds. Berlin, Heidelberg: Springer Berlin Heidelberg, 2012, pp.
208–221.
[2] C. Hoareau and I. Satoh, “Modeling and Processing Information for
Context-Aware Computing: A Survey,” New Gener. Comput., vol. 27,
no. 3, pp. 177–196, May 2009.
[3] S. Sehic, F. Li, S. Nastic, and S. Dustdar, “A programming model for
context-aware applications in large-scale pervasive systems,” in 2012
IEEE 8th International Conference on Wireless and Mobile Computing,
Networking and Communications (WiMob), 2012, pp. 142–149.
[4] S. Sehic, S. Nastic, M. Vögler, F. Li, and S. Dustdar, “Entity-adaptation:
A Programming Model for Development of Context-aware
Applications,” in Proceedings of the 29th Annual ACM Symposium on
Applied Computing, Gyeongju, Republic of Korea, 2014, pp. 436–443.
[5] S. Nastic, S. Sehic, M. Vogler, H.-L. Truong, and S. Dustdar, “PatRICIA
-- A Novel Programming Model for IoT Applications on Cloud
Platforms,” in 2013 IEEE 6th International Conference on Service-
Oriented Computing and Applications, Koloa, HI, USA, pp. 53–60.
[6] V. Reynolds, M. Hauswirth, and G. Hynes, “A Context Lifecycle For
Web-Based Context Management Services,” 2009.
[7] A. Yavari, P. P. Jayaraman, D. Georgakopoulos, and S. Nepal,
“ConTaaS: An Approach to Internet-Scale Contextualisation for
Developing Efficient Internet of Things Applications,” in Proceedings of
the 50th Hawaii International Conference on System Sciences, 2017.
[8] S. Patidar, D. Rane, and P. Jain, “A Survey Paper on Cloud Computing,”
in 2012 Second International Conference on Advanced Computing
Communication Technologies, 2012, pp. 394–398.
[9] M. Wooldridge and N. R. Jennings, “Intelligent agents: theory and
practice,” Knowl. Eng. Rev., vol. 10, no. 2, pp. 115–152, Jun. 1995.
[10] P. Moore, F. Xhafa, and L. Barolli, “Context-as-a-Service: A Service
Model for Cloud-Based Systems,” in 2014 Eighth International
Conference on Complex, Intelligent and Software Intensive Systems,
2014, pp. 379–385.
[11] L. Yao, B. Benatallah, X. Wang, N. K. Tran, and Q. Lu, “Context as a
Service: Realizing Internet of Things-Aware Processes for the
Independent Living of the Elderly,” in Service-Oriented Computing, vol.
9936, Q. Z. Sheng, E. Stroulia, S. Tata, and S. Bhiri, Eds. Cham:
Springer International Publishing, 2016, pp. 763–779.
[12] N. M. do Nascimento and C. J. P. de Lucena, “FIoT: An agent-based
framework for self-adaptive and self-organizing applications based on
the Internet of Things,” Inf. Sci. , vol. 378, pp. 161–176, Feb. 2017.
[13] B. Manate, V. I. Munteanu, and T. F. Fortis, “Towards a Smarter Internet
of Things: Semantic Visions,” in 2014 Eighth International Conference
on Complex, Intelligent and Software Intensive Systems, 2014, pp. 582–
587.
[14] D. Le-Phuoc, A. Polleres, M. Hauswirth, G. Tummarello, and C.
Morbidoni, “Rapid prototyping of semantic mash-ups through semantic
web pipes,” in Proceedings of the 18th international conference on
World wide web, 2009, pp. 581–590.
[15] C. Perera, A. Zaslavsky, P. Christen, and D. Georgakopoulos, “Context
Aware Computing for The Internet of Things: A Survey,” IEEE
Communications Surveys Tutorials, vol. 16, no. 1, pp. 414–454, 2014.
[16] G. D. Abowd, A. K. Dey, P. J. Brown, N. Davies, M. Smith, and P.
Steggles, “Towards a Better Understanding of Context and Context-
Awareness,” in Handheld and Ubiquitous Computing, 1999, pp. 304–
307.
[17] X. Li, M. Eckert, J.-F. Martinez, and G. Rubio, “Context Aware
Middleware Architectures: Survey and Challenges,” Sensors , vol. 15,
no. 8, pp. 20570–20607, Aug. 2015.
[18] D. Zhang, H. Huang, C.-F. Lai, X. Liang, Q. Zou, and M. Guo, “Survey
on context-awareness in ubiquitous media,” Multimed. Tools Appl., vol.
67, no. 1, pp. 179–211, Nov. 2013.
[19] C. Bettini et al., “A survey of context modelling and reasoning
techniques,” Pervasive Mob. Comput., vol. 6, no. 2, pp. 161–180, 2010/4.
[20] I. Szilagyi and P. Wira, “Ontologies and Semantic Web for the Internet
of Things - a survey,” in IECON 2016 - 42nd Annual Conference of the
IEEE Industrial Electronics Society, Florence, Italy, pp. 6949–6954.
[21] “FOAF Vocabulary Specification 0.99,” Jan-2014. [Online]. Available:
http://xmlns.com/foaf/spec/.
Vol. 15, No. 8, Augus 2017
ISSN 1947-5500

[22] M. Wick, B. Vatant, and B. Christophe, “Geonames ontology,” URL
http://www. geonames. org/ontology, 2015.
[23] “Time Ontology in OWL,” 2017. [Online]. Available:
https://www.w3.org/TR/owl-time/.
[24] P. Misra, Y. Simmhan, and J. Warrior, “Towards a Practical Architecture
for the Next Generation Internet of Things,” arXiv [cs.CY], 03-Feb-2015.
[25] I. Lee and K. Lee, “The Internet of Things (IoT): Applications,
investments, and challenges for enterprises,” Bus. Horiz., vol. 58, no. 4,
pp. 431–440, 2015/7.
[26] “Healthcare for the Elderly Using the IoT at GOTO Berlin 2015.”
[Online]. Available: https://gotocon.com/berlin-
2015/presentation/Healthcare%20for%20the%20Elderly%20Using%20th
e%20IoT. [Accessed: 09-Apr-2017].
[27] “Akka.” [Online]. Available: http://akka.io/. [Accessed: 09-Apr-2017].
[28] “Jade Site | Java Agent DEvelopment Framework.” [Online]. Available:
http://jade.tilab.com/. [Accessed: 09-Apr-2017].
Vol. 15, No. 8, Augus 2017
ISSN 1947-5500

Towards a Semantic-based Context-as-a-Service for Internet of Things

More Related Content

What's hot (17)

Similar to Towards a Semantic-based Context-as-a-Service for Internet of Things (20)

Recently uploaded (20)

Towards a Semantic-based Context-as-a-Service for Internet of Things