This When the context seems to be invalid

This
section contains number of research publications in the perspective of context
awareness and IoT.

Selecting the most suitable sensor to
get the desired form of output is one of the most difficult and important task
in the IoT environment. Perera et al 2 proposed a novel mechanism that is
sensing-as-a-service platform to select an appropriate sensor. This paper propose
the Context Awareness for Internet of Things (CA4IOT) architecture to help
users by automating the task of selecting the sensors according to the problems/tasks
at hand. This paper also focus on automated configuration of filtering, fusion
and reasoning mechanisms that can be applied to the collected sensor data
streams using selected sensors.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

 

Bellavista et al 15 present a
unified architectural model and a new taxonomy for context data distribution,
by considering and comparing a large number of solutions.

 

Context is a broad term, it is not
limited to location only. Schmidt
el al 16 explains context to be much more the location only. Previously
context was considered only to be location.

 

To model the context E. Elnahrawy 17 has used an
object oriented approach to model context using XML. There are three sections in
the proposed context model: behavioral information (e.g. whether the context
attribute has a constant or variable value), context-specific abstractions
(e.g. contextual events and queries) and structural information (e.g.
attributes and dependencies among context types).

G. Castelli
18 uses a W4 (who, what, where, when) based context model to structure data
in order to extract high-level information from location data. All these four
parameters defines the modeling of the context. After this a new modeling
approach was proposed that contains 5W (who, what, where, when and Why).

 

H. Chang
at el 19 defines the six states of context in an IOT infrastructure (ready,
running, suspended, resumed, expired and terminated).

Ready:
Every context is in the ready state at the initial stage.

Running:
When the context is in running state.

Suspended:
When the context seems to be invalid temporally.

Resumed:
When the context becomes valid from being suspended.

Expired:
When the context has expired and further information is not available.

Terminated:
When the context is no longer valid and further information is not available

 

Selecting the appropriate raw context
for reasoning is critical to infer high-level context with high accuracy. Guan
et al. 20 has proved that using more context will not necessarily improve the
accuracy of the inference in a considerable manner. They have used two reasoning
models in their research: back-propagation neural networks and k-nearest neighbors.
According to the results, 93% accuracy has been achieved by using ten raw
context. Adding 30 more raw context to the reasoning model has increased the
accuracy by only 1.63%. So using appropriate context is important.

 

Aura 21 is an
application that shows the requirement of having IoT middleware running over
many platforms and devices under different resource limitations where different
versions with different capabilities would fit on different devices. People
move from one situation to another and IoT solutions need to track user
movements and facilitate context-aware functionalities over different forms of
devices. Mobility is an important aspect to be considerd in the IoT.

 

End-users in the IoT paradigm are
more interested in high level information compared to low-level raw sensor data
22. The following examples explain the difference between high level information
and low-level raw sensor data. It is raining (high-level information) can be
derived from humidity is 80% (low-level sensor data). Further, high-level
sensor data can be explained as semantic information as it provides more
meaning to the end users.

 

Ontology based
approach can be used to manage user privacy via policies which allow it to monitor
and access contextual control context 23. Ontologies are getting popular and
adopted in web related developments, such practice will makes IoT development
much easier.

 

In the IoT, fusion is extremely
important as there will be billions of sensors available so a large number of
alternative sources will exist to provide the same information. Sensor data
fusion is a method of combining sensor data from multiple sensors to produce
more accurate, more complete, and more dependable information that could not be
achieve through a single sensor 24.

 

Dynamic
composition plays a vital role in IoT. This is applicable where possible
interactions cannot be identified at the design and development stage. COSMOS
25, Solar 26 and CMF (MAGNET) 27 promote the provision of dynamic
composition.

 

In the IoT, there
will be many sensors that makes data available from many sources providing
similar type of data/information. CARISMA 28 shows how conflict resolution
can be done using profiles and rules where it stress the importance of making
decisions to optimize the return for every party involved. This can be used to
derive the same knowledge where conflict resolution will help to make accurate
actions.

Content Sharing is
one of the main concerns while implementing IoT. Park et al. 29 mentions the
significance of context sharing using mobile devices. This provides more accuracy
in reasoning, comprehensiveness and high level context recognition.

 

Dynamic organizing is critical in
environments like the IoT, because it is impossible to identify or plan
possible interaction at the development stage. Software solutions should be
able to understand the requirements and demands on each situation, then
organize and structure its internal components according to them. IoT solutions
must have a programming model that allows dynamic composition without requiring
the developer or user to identify specific sensors and devices 26.

 

SOCAM 30 shows
how knowledge can be separated among different levels of ontologies. Upper
ontology models general purpose data while domain specific ontologies model
domain specific data. Data models need to be extensible on demand Due to
unpredictability and broadness of IoT. IoT solutions may need to be expand its
knowledge-base towards different domains.

 

Dynamic
Configuration is essential for the system of IoT. SALES 31 shows various different
devices with different resource constraints in the hardware level can be
managed using distributed architecture.

 

Energy is one of
the main concerns while implementing IoT infrastructure. PROCON 32 proposes
simple optimization techniques that can save significant amount of energy in an
environment such as the IoT where billions of objects communicate with each
other.

 

Embedded devices
to applications is one of the early efforts at building IoT middleware 33. It
shows how the context modelling needs to be done in order to model device
information. Where domains and required knowledge cannot be predicted during
the development stage pluggable rules are used that allow insertions when necessary
as it is a major requirement in IoT middleware applications.

 

Automated
configuration has great significance in an Iot environment. UPnP FRAMEWORK 34
is relying on the vision of the IoT where machine-to-machine communication play
a significant role. This approach is applicable to devices such as cameras, web
cams, and microwaves. UPnP approach is a key technology that enables automated
configuration.