environment observation system based on semantics in the ...€¦ · and this system is collecting...
TRANSCRIPT
Environment Observation System based on
Semantics in the Internet of Things
Liang Hu, Jingyan Jiang, Jin Zhou, and Kuo Zhao* College of Computer Science and Technology, Jilin University, Changchun 130012, China
Email: [email protected], [email protected]
*Corresponding author
Liang Chen College of Software, Jilin University, Changchun 130012, China
Huimin Lu College of Software, Changchun University of Technology, Changchun 130012, China
Abstract—The Internet of Things is rapidly developing in
recent years. A number of devices connect with the Internet.
Hence, the interoperability, which is the access and
interpretation of unambiguous data, is strongly needed by
distributed and heterogeneous devices. The semantics
promotes the interoperability in the Internet of Things using
ontology to provide precise definition of concepts and
relations. In this paper, we demonstrate the importance of
the semantics in three aspects: firstly, the semantics means
that the machines could understand and respond to the
human command. Secondly, the semantics is mainly
reflected in the ontology of the physical world. Thirdly, the
semantics is important for interoperability, data integration
and reasoning. Then, we introduce a semantic approach to
construct an environment observation system in the Internet
of Things. The environment observation system contains
three major components, the first is the ontology of the
environment observation system, the second is the semantic
map of the environment observation system, and the last is
exposing the observation data. Finally, the system publishes
the environment observation data on the Web successfully.
The environment observation system with semantics
provides a better service in the Internet of Things.
Index Terms—Internet of Things; Semantics; Ontology;
Semantic Sensor Network; Environment Observation
System
I. INTRODUCTION
The Internet of Things (IoT) is an expansion of the
current Internet, which refers to interconnecting “Things”
has attracted lots of interest in recent years. It is the potentialities of the IoT that makes it possible for the
development of a huge number of applications [1], such
as transportation and logistics domain, healthcare domain,
smart environment (home, office, plant) domain.
Reference [2] proposed a smart hospital model using IoT
technology in healthcare domain. Reference [3] proposed
an intelligent traffic monitoring system based on IoT
technology in transportation and logistics domain. In smart environment domain, reference [4] puts forward a
smart home system based on the Internet of Things. The essential requirements of IoT are the interoperability,
automation and data analytics, in which case, the
semantics is a significant perspective and solution.
With the help of the semantic technology, we could
effectively help promote the interoperability of the IoT.
So, a number of works has been developed.
The ontology is a modeling tool for describing
knowledge in a semantic way. The ontologies in IoT have a great improvement. The Sensor Web Enablement (SWE)
of OGC defines a suite of interfaces and services to
connect heterogeneous sensors. It contains several
specifications: Observations & Measurement (O&M)
defines standard models and XML Schema for encoding
observations and measurements, which are archived and
real-time; Sensor Model Language (SensorML) defines
standard models and XML Schema for describing sensors systems and processes refer to sensor observations;
Sensor Observations Service (SOS) is a standard web
service interface for a client to request, filter, and retrieve;
Transducer Model Language (TransducerML or TML),
Sensor Planning Service (SPS), Sensor Alert Service
(SAS) and Web Notification Services (WNS) are also
part and parcel of SWE. Recently, an ontology called
SSN for describing sensors and sensor network resources has been developed by the W3C Semantic Sensor
Networks Incubator Group. The SSN ontology provides a
better support for abstraction, categorization, and
reasoning offered by semantic technologies [5]. Since the
SSN ontology defines high level schema of deployments,
systems, the operating restrictions, processes, devices and
data, it answers the need for interoperating with other
ontologies. Linked data plays an important role in effective
reasoning, processing mechanisms and better
interoperability. It is based on the HTTP and URIs and
enables not only the people but also the machines to
explore data on the Web. The resources are identified by
URI and described in RDF, URIs link to other URIs, so
that data from different sources are connected [6]. The
JOURNAL OF NETWORKS, VOL. 8, NO. 12, DECEMBER 2013 2721
© 2013 ACADEMY PUBLISHERdoi:10.4304/jnw.8.12.2721-2727
linked open data (LOD) integrates numerous distributed
data in form of RDF across the Web [7]. Many
applications use linked data principles to annotate IoT
data: Kno. e. sis linked sensor data describes the weather
data using linked data principles and links the Geonames
for searching by geographical location [8]. Sense2Web
provides a GUI to describe IoT data [9]. Page et al. propose an API to publish data of observatory using
linked data principles and REST [10].
In this paper, we just focus on the environment
observation, which is in a higher level of the Internet of
Things. The lower level is the part of data sensing and
data collection. We deployed many sensors in different
rooms of our lab, and these sensors sensed the
environment and collected real data (e. g. temperature, humidity and illumination) into a relational database. In
the higher level, we introduce an approach of semantics
to publish the real sensor data on the Web, it contains
three parts: ontology, semantic mapping and exposing.
The ontology is a data abstract of the existing sensor data,
the real things are mapped to concept and properties. In
our environment observation system, the nodes and some
environmental indicators (e. g. temperature) are described as classes with properties. The goal of semantic mapping
is to map the relational database schema to RDF or
ontology using D2R mapping language, in this process,
the data integration can be applied for context and
situation awareness. Semantic interpretation is a crucial
issue for exploring the semantic descriptions. In this work,
the D2R server is applied to implement the publishment.
The rest of this paper is structured as follows. In section 2, we demonstrate the importance of the
semantics in three aspects. In section 3, we introduce the
architecture of the environment observation system in the
Internet of Things. In section 4, we general evaluate the
environment observation system. Section 5 is a summary
of this paper.
II. SEMANTICS IN THE INTERNET OF THINGS
The semantics makes a great contribution in many domains and draws a lot interests. Reference [11] extends
the FOAF vocabulary for social network domain.
Reference [12] and reference [13] propose the novel
video retrieval approach based on semantics in the muti-
media domain. The semantics is important for the Internet
of Things as well. In order to shed light on the semantics
in the Internet of Things, three issues should be discussed.
Firstly, we will explain the meaning of the semantics. The semantics is the interpretation of an expression. M.
Uschold divided semantics into four levels: implicit,
explicit and informal, explicit and formal for human
processing, and explicit and formal for machine
processing. The implicit semantics is a kind of shared
human consensus such as meaning tags. The explicit and
informal semantics is an informally expressed notation or
language which is mainly for humans. For example, the meaning terms in the Dublin core. Explicit and formal for
human processing semantics is a formal documentation
just for humans. For example, the modal logic for the
semantic definition of ontological categories. The explicit
and formal for machine processing semantics is the kind
of semantics which is used at runtime for automated
inference for machines. In Semantic Web, the semantics
means that the machines could understand and respond to
the human command. It is explicit and formal for
machine processing as M. Uschold explained in 2003
[14]. For example, the machines understand the “temperature” by meaning instead of a character string.
And then the machines will have a reaction relate to
“temperature”, for instance, the machines will give a
command or instruction to a device to open the air
condition. In this process, semantically structured
information (Explicit and formal for human processing
semantics) should change into the machines processing
semantics level for automated inference. Secondly, we will set forth the formal description of
the semantics. The “understanding” ability of machine
requires the information which is semantically structured
derived from the Internet of Things. Hence, the IoT data
should be modeled as ontology. The ontology is an
explicit, formal and shared conceptual model which is
structured using standard ontology language such as
OWL and RDF, and it provides the precise definition of concepts and relations. The precise definition is the key
of knowledge representation, information sharing and
semantic reasoning. Therefore, the semantics is mainly
reflected in the ontology. In the Internet of Things, the
scenario could be described as ontology. The nodes and
observation attributes are the classes and relations.
Thirdly, we will underline importance of the semantics
in the Internet of Things. First, the semantics is important for interoperability. The Internet of Things requires a
stunning number of devices connecting with Internet. For
this reason, the interoperability, which is the access and
interpretation of unambiguous data, is intensively needed
by distributed and heterogeneous devices and is provided
by the unambiguous data representation. The
semantically structured data precisely answers the need to
unambiguity. Therefore, the semantics is essential or even indispensable in interoperability. Second, the semantics
plays a vital role in data integration. The goal of the
Internet of Things is to sense the environment and react
intelligently. The IoT data can combine with other
domain information and integrate to infer new
information. Consequently, the unambiguous data
representation is an approach for heterogeneous data
integration and semantic descriptions. Third, the semantics is a key foundation in reasoning. The reasoning
enables the inference from existing rules and information.
The semantically structured data is the premise of the
reasoning. In this way, the Internet of Things could react
intelligently.
III. THE ARCHITECTURE OF THE ENVIRONMENT
OBSERVATION SYSTEM
This section introduces the architecture of the environment observation system. The environment
observation system is based on semantics, using standard
ontology languages such as RDF or OWL to represent the
data which is stored in the RDB (relational database).
2722 JOURNAL OF NETWORKS, VOL. 8, NO. 12, DECEMBER 2013
© 2013 ACADEMY PUBLISHER
And this system is collecting real world data (i. e.
humidity and temperature) through the sensors deployed
in the Internet of Things platform. Consequently, this
system aims to publish the real observational
environment data in a semantic way and integrate more
information in specific scenario.
The environment observation system comprises three major components as illustrated in Fig. 1, the first one is
the ontology of the environment observation system, the
second one is the semantic map of the environment
observation system, and the last one is exposing the
observation data. And the system is in the top level of
Internet of Things platform. Obviously, the initial data is
provided by sensors and stored in the database in the
lower level. The ontology of the environment observation system
describes the observation data in sensor network domain,
and exploits the Semantic Sensor Network (SSN)
ontology which is the standard ontology for describing
sensors and sensor network by the W3C Semantic Sensor
Networks Incubator Group. In this paper, we just focus
on the observation data of the sensors so that only a
subset of the SSN ontology concepts is applied to the system. The ontology is modeled through the actual IoT
scene to describe the environment of two rooms in our
labs.
The semantic map of the environment observation
system is the core component of this system. The RDFS
vocabularies and OWL ontologies could be published on
the Web, thus, the observation data which stored in the
relational database (i. e. MySQL) could be mapped into the RDFS vocabularies and OWL ontologies. In order to
describe the relation between a relational database
schema and RDFS vocabularies or OWL ontologies, the
D2RQ platform is applied in the system. The observation
table is mapped into an observation class, and the
columns in the table are mapped into the properties. The
classes and properties provide precise descriptions of the
environment observation scene and the semantization of data. In addition, the semantic map could provide more
information by integrating other domain knowledge such
as the location. The data in relational database is flexibly
reassembled and linked as well.
Data Provider
Data Collection
semantic mapontology
Exposing
Environment observation system
Database
Sensors
Internet of Things
Figure 1. The levels of the environment observation system.
The exposing the observation data process publishes
data on the Web. In order to access the content of the
database as linked data over the Web, the D2R server is
an appropriate choice. D2R server provides a linked data
view, a HTML view and a SPARQL Protocol endpoint
over the database by HTTP protocol [15]. Hence, the
linked data could be browsed and searched using web
browsers and SPARQL easily.
sensor
RDB
sensor
ontology
mapping
Virtual RDF Graphs
HTML Browser
data
information
knowledge
Lo
we
r
Figure 2. The data flow diagram
Figure 3. The database schema
In order to illustrate the data flow of the environment observation system, a data flow diagram is shown in Fig.
2. The data transformation process is mainly divided into
three stages. First, the raw sensory data is perceived by
the sensors, and collected by a relational database directly.
The data is structured in the form of database table
without any semantics. Then the data is described as a
sensory ontology based on the content of database table.
All the observation items are described as classes, and each class has several properties. The mapping file is
created on the basis of the ontology and database schema
using D2R mapping language. Finally, the data from
relational database is mapped into a virtual RDF graphs
via a mapping file. It is worth noting that a dotted line
points to virtual RDF graphs from RDB. In fact, the
transformation from RDB to virtual RDF graph is direct
under the guidance of mapping file. Since the mapping file guides how the instances of virtual RDF graph are
created, it is crucial in the process. On the right of the
figure, a “knowledge hierarchy” corresponding to the
three transformation stages are shown. The data layer is a
lower layer refers to numberous raw data collected by the
IoT devices. The information layer is in the middle, the
interoperability is achieved by structuring with semantics
from the raw data. The knowledge layer provide machine-readable and machine-comprehensible data,
which could be browsed and searched on the Web. It is
obvious that the data becomes more structured, integrated
and semantic through the environment observation
system, which answers the requirement of the Internet of
Things.
JOURNAL OF NETWORKS, VOL. 8, NO. 12, DECEMBER 2013 2723
© 2013 ACADEMY PUBLISHER
nodeIdCls
time
airHumidityCls_unit
airTemperatureCls
LightCls
nodeIdCls_parent
nodeIdCls_board_id
nodeIdCls_roomOnessn:observedBy
ssn:observationRusultTime
ssn:observedProperty
ssn:observedProperty
ssn:observedProperty
eos:unit
ssn:hasValue
eos:hasParent
eos:hasBoard
eos:roomBelong
airHumidityCls
airTemperatureCls_unit
LightCls_light
airHumidityCls_humidity
airTemperatureCls_temperature
LightCls_uniteos:unit
ssn:hasValue
eos:unit
ssn:hasValue
nodeIdCls_roomTwo
eos:roomBelong
Observation_sht11_5w
Figure 4. The ontology of the environment observation system
A. The Ontology of the Environment Observation System
In general, ontology is a specification of a
conceptualization [16], it is a modeling tool for
describing knowledge in a semantic way. In this paper,
the ontology is applied to modeling the concept of the environment observation system based on the actual IoT
scene and take advantage of the SSN ontology.
TABLE I. THE PROPERTIES OF ONTOLOGY
Property name explanation
ssn:
observationRusult
Time
Points to the time when the observation
result became available.
ssn: observedBy Points to a sensor which produced the
observation data.
ssn:
observedProperty
Points to the specific quality of the feature
of interest which was observed.
eos: hasParent Points to the parent nodes.
eos: hasBoard Points to the sensor board ID.
ssn: hasValue Points to the actual value of the
observation data.
eos: roomBelong Points to the room ID which the nodes
belong to.
eos: unit Points to the measuring unit of the
observation value.
The observation data is obtained from the sensors
which are deployed in our lab, and is stored in the
database. The database schema is shown in Fig. 3.
Idsht11_5w is the primary key, it is an ID of all these
data. Nodeid is the ID of the nodes which are placed in
the lab. The parent records the node ID which represents
routing path of the data. Board_id is applied to manage sensor boards by recording the ID of sensor boards which
is attached to the nodes. The airHumidity is the real air
humidity of the room, and measured by the relative
humidity. The airTemperature is the real temperature of
the room, and its unit is centigrade. The light is the real
illumination intensity of the room, its unit of
measurement is lux. The time is the date and time, and
the pattern demonstrates like this “yyyy-mm-dd hh: mm: ss”.
The ontology of the system is proposed on the base of
database schema as illustrated in Fig. 4, The ellipse
means classes and the arrows means properties. Explicitly,
the ontology provides more information, the measuring
unit and the room ID that deployed nodes is the
additional information. The data is coordinated by the location information and the time information which
makes the observation model more concrete and precise.
Also the ontology reassembles the data in a hierarchical
way. The properties are shown in the Table I.
B. The Semantic Map of the Environment Observation
System
The semantic map plays a very important role in the
system. It maps the schema of a relational database to
RDF and defines a virtual RDF graph which contains
information from the database. D2RQ platform provides
a mapping language to accomplish the map process. The
main goal of the mapping language is to map a RDF from the RDB without changing the existing database schema.
The complexity is reduced greatly by joining the SQL
statement into the mapping process. The mapping process
is mainly divided into four steps as follows:
(1). A record set is selected for each class by SQL
statement.
(2). Grouping the record set according to the specific
ClassMap. (3). Creating the instances of each class and assigning
the URI.
(4). Datatype and property bridges are applied to each
instance property.
A virtual RDF graph is created after the four steps
above. The virtual RDF structure for the system is shown
in Fig. 5.
The database is mapped to RDF terms through d2rq: ClassMap and d2rq: PropertyBridges, shown on the right.
The class map is the most important concept. The class
map represents a class or a group of similar classes of the
ontology. A class map specifies how URIs are generated
for the instances of the classes [17]. In this semantic map,
the map: observation_sht11_5w_ClassMap is the major
class map, and all the observation items are included in it.
In order to expand the observation model, we structure additional class map, nodeidCls_classMap provides more
information of the nodes, such as serial numbers, board
IDs and the locations. It is noteworthy that the nodeid
property of observation_sht11_5w_ClassMap doesn’t
have the concrete value, instead, it references another
d2rq: ClassMap which creates the instances which are
used as the values of nodeidCls_classMap by d2rq:
2724 JOURNAL OF NETWORKS, VOL. 8, NO. 12, DECEMBER 2013
© 2013 ACADEMY PUBLISHER
Map:database
Map:nodeIdCls_classMap
Map:id_PropertyBridge
Map:parent_PropertyBridge
Map:board_id_PropertyBridge
Map:room_one_PropertyBridge
Map:room_two_PropertyBridge
eos:id
eos:hasParent
eos:hasBoard
eos:roomBelong
eos:roomBelong
refersToClassMap map:nodeIdCls
belo
ngsTo
ClassM
ap
uriPattern "nodeIdCls/@@observation_sht11_5w.ideos_5w@@"
column "observation_sht11_5w.nodeid"
column "observation_sht11_5w.parent"
column "observation_sht11_5w.board_id"
constantValue eos:roomOne
constantValue eos:roomTwo;
uriPattern "airHumidityCls/@@observation_sht11_5w.ideos_5w@@"
constantValue eos:Relative_Humidity
property
property
property
property
property
Map:Observation_eos_5w_classMap
Map:ideos_5w_PropertyBridge
Map:Nodeid_PropertyBridge
Map:airHumidity_PropertyBridge
Map:time_PropertyBridge
column "observation_sht11_5w.ideos_5w"
vocab:observation_eos_5w_idsht11_5w
ssn:observationResultTime
ssn:observedProperty
ssn:observedBy
column "observation_sht11_5w.time"
uriPattern "observation_eos_5w/@@observation_sht11_5w.ideos_5w@@"
refersToClassMap map:airHumidityCls
belo
ngsTo
ClassM
ap
property
property
property
property
Map:airHumidityCls_classMap
Map:humidity_PropertyBridge
Map:unit_PropertyBridge
ssn:hasValue
eos:uint
column "observation_sht11_5w.airHumidity"
Property
property
ssn:Sensor
class
class
class
vocab:observation_sht11_5w
vocab:airHumidityCls
:dataStorage
:dataStorage
:dataStorage
…
…
…
…
belo
ngsTo
ClassM
ap
Figure 5. The virtual RDF structure for the system
referenceClassMap. It is just like the concept of the
foreign key. Along the same lines,
airHumidityCls_classMap, airTemperatureCls_classMap
and lightCls_classMap are structured for the same
purpose.
Each class map has many property bridges, which
specifies how the properties of an instance are created [17]. Therefore, property bridges map a database column
to an RDF property. The d2rq: column points the
properties with literal values which are the table column
names in the relational database. For example, d2rq:
column "observation_sht11_5w. time" means the value of
map: time_PropertyBridge is observation_sht11_5w. time.
The d2rq: constantValue points that the property has the
same constant value on all the instances of the class map.
For example, d2rq: constantValue eos: roomTwo is the
constant value of the map: room_two_PropertyBridge,
which implies that the location for some nodes. Taking
advantage of the d2rq: constantValue, we could provide some constant value inexistent in the database. The d2rq:
referenceClassMap is also a property of d2rq:
PropertyBridge and we have discussed it before.
JOURNAL OF NETWORKS, VOL. 8, NO. 12, DECEMBER 2013 2725
© 2013 ACADEMY PUBLISHER
Figure 6. The query interface
Figure 7. The classes of the system
Figure 8. The properties of the system
C. Exposing the Observation Data
In order to publish the RDF that we have created before, the D2R server is applied. D2R Server is a tool
for publishing the content of relational databases on the
Semantic Web. D2R Server uses a D2RQ mapping to
map relational database schemas into RDF, and allows
the RDF data to be browsed and searched on the
Semantic Web. To start D2R Server, we need the
command as follows: “d2r-server –p 8080 mappingName.
ttl”, then open the URL “http: //localhost: 8080/snorql/” in a web browser. A query interface is shown in the web
just as the Fig 6, Then we could write the query into the
text of the SPARQL section and push the "Go!" button to
obtain the results of the query in the SPARQL results
section. The classes and properties of the system is shown
in the Fig. 7 and Fig. 8. An example of a query for air
temperature resources is shown in the Fig. 9. This
example searches all the useful information about the air temperature of No. 997 through the data links such as
time, unit, value and observed nodeid. The result is
shown in the Fig. 10. The information of air temperature
of No. 997 has been assigned a uniform resource
identifier (URI), http: //localhost:
8080/resource/airTemperatureCls/997. The SPARQL
result is a triple in a pattern as “subject-verb-object”. The
subject is the information of air temperature of No. 997,
in other words, the URI mentioned above. The verb is the
property column listed in the left such as eos: unit and rdf:
type. The object is the hasValue column and isValueOf
column. A underlying sentence is made by connecting the
subject-verb-object. For example, the unit of air
temperature of No. 997 is centigrade. The underlying sentence demonstrates the characteristics of semantic.
WHERE { { <http://localhost:8080/resource/airTemperatureCls/997> ?property ?hasValue } UNION { ?isValueOf ?property <http://localhost:8080/resource/airTemperatureCls/997> } UNION { <http://localhost:8080/resource/observation_sht11_5w/997> ?property ?isValueOf FILTER (?isValueOf != <http://localhost:8080/resource/LightCls/997> ). FILTER (?isValueOf != <http://localhost:8080/resource/airHumidityCls/997>) . FILTER (?isValueOf != <http://localhost:8080/resource/airTemperatureCls/997> ). FILTER (?property != rdf:type) }}ORDER BY (!BOUND(?hasValue)) ?property ?hasValue ?isValueOf
Figure 9. The SPARQL query
IV. EVALUATION
The Internet of Things is rapidly developing in recent
years. Many applications based on the Internet of Things
are developed in various domains. Reference [4] puts
forward a smart home system based on the Internet of
Things. The monitoring system has three levels, the sensing layer, network layer and application layer. The
data is collected in the sensing layer, transmitted in the
networks layer and processed in application layer.
Obviously, the whole system is without the semantics.
In this paper, we propose an environment observation
system architecture which exposed the observation data
on the Semantic Web. Besides the three layers mentioned
above, environment observation system has a semantics layer. Hence, the observation data which is provided by
the sensors could be browsed and searched in a semantic
way. The semantics is derived from the given ontology
model. The ontology is modeled through the actual IoT
scene to describe the environment of two rooms in our
labs. The ontology provides precise definition of concepts
and relations which reflects semantics. However, the
smart home system could not be searched in the semantic way and different devices do not share the same concepts
which lead to the lower interoperability.
The D2R mapping language perfectly maps the
relational database schemas to the RDF. The mapping
process mixes the data from the database and integrates
more information as specific scenario. However, in smart
home system, the data is structured in relation database
without semantics. Hence, the integration of the data is only implemented in application layer and the degree of
integration is lower.
In general, to publish environment observation system
in a semantic way is a non-trivial task. In the Internet of
Things, it is not enough to network the devices, but the
semantization of data, which promote data understanding
and semantic reasoning, so that it makes the Internet of
Things smarter.
2726 JOURNAL OF NETWORKS, VOL. 8, NO. 12, DECEMBER 2013
© 2013 ACADEMY PUBLISHER
Figure 10. An example
V. CONCLUSIONS AND FUTURE WORK
In this paper, we first illustrate the importance of the semantics in the Internet of Things by discussing three
issues: What is the semantics? Where is semantics in the
Internet of Things? Why is semantics important? Then,
we introduce a semantic approach to the Internet of
Things through constructing an environment observation
system. The environment observation system comprises
three major components, the first is the ontology of the
environment observation system, the second is the semantic map of the environment observation system,
and the last is exposing the observation data. The system
publishes the environment observation data on the
Semantic Web successfully. In the future, we could add
our semantic environment observation data in the Internet
of Things into the Linked Open Data (LOD) for sharing
and interoperability.
ACKNOWLEDGEMENT
This work was supported in part by the National High
Technology Research and Development Program of
China under Grant No. 2011AA010101, the National
Natural Science Foundation of China under Grant No.
61103197, 61073009 and 61240029, the Science and
Technology Key Project of Jilin Province under Grant No.
2011ZDGG007, the Youth Foundation of Jilin Province
of China under Grant No. 201101035, and the Fundamental Research Funds for the Central Universities
of China under Grant NO. 200903179, the China
Postdoctoral Science Foundation under Grant No.
2011M500611, he 2011 Industrial Technology Research
and Development Special Project of Jilin Province under
Grant No. 20110069, the 2012 National College Students'
Innovative Training Program of China.
REFERENCES
[1] L. Atzori, A. Iera, and G. Morabito, "The Internet of
Things: A survey, " Computer Networks, vol. 54, pp. 2787-2805, Oct 28 2010.
[2] L. Yu, Y. Lu, and X. Zhu, "Smart hospital based on internet of things, " Journal of Networks, vol. 7, pp. 1654-1661, 2012.
[3] L. Xiao and Z. Wang, "Internet of things: A new application for intelligent traffic monitoring system, " Journal of Networks, vol. 6, pp. 887-894, 2011.
[4] K. Bing, L. Fu, Y. Zhuo, and L. Yanlei, "Design of an Internet of things-based smart home system, " in 2nd International Conference on Intelligent Control and Information Processing, ICICIP 2011, July 25, 2011 - July
28, 2011, Harbin, China, 2011, pp. 921-924. [5] Semantic Sensor Network XG Final Report, 2011 http:
//www. w3. org/2005/Incubator/ssn/XGR-ssn-20110628 [6] http://www.w3.org/DesignIssues/LinkedData. html [7] C. Bizer, T. Heath, and T. Berners-Lee, "Linked data - The
story so far, " International Journal on Semantic Web and Information Systems, vol. 5, pp. 1-22, 2009.
[8] H. Patni, C. Henson, and A. Sheth, "Linked sensor data, " in 2010 International Symposium on Collaborative Technologies and Systems, CTS 2010, May 17, 2010 - May 21, 2010, Chicago, IL, United states, 2010, pp. 362-370.
[9] S. De, T. Elsaleh, P. Barnaghi, Meissner, and Stefan, "An internet of things platform for real-world and digital objects, " Scalable Computing, vol. 13, pp. 45-57, 2012.
[10] Page, K. , et al. , "Linked Sensor Data: RESTfully serving RDF and GML " in 2nd International Workshop on Semantic Sensor Networks, 2009.
[11] B. Zhou and C. Wu, "Semantic model for social networking federation, " Advances in Information Sciences and Service Sciences, vol. 3, 2011
[12] J. Sang and C. Xu, "Faceted subtopic retrieval: Exploiting the topic hierarchy via a multi-modal framework, " Journal of Multimedia, vol. 7, pp. 9-20, 2012
[13] M. H. Kolekar, K. Palaniappan, S. Sengupta, and G. Seetharaman, "Semantic concept mining based on hierarchical event detection for soccer video indexing, " Journal of Multimedia, vol. 4, pp. 298-312, 2009.
[14] M. Uschold, "Where are the semantics in the semantic web?, " Ai Magazine, vol. 24, pp. 25-36, Fal 2003.
[15] http: //d2rq. org/ [16] T. R. Gruber, "Toward principles for the design of
ontologies used for knowledge sharing, " International Journal of Human Computer Studies, vol. 43, pp. 907-907, 1995.
[17] http: //d2rq. org/d2rq-language
JOURNAL OF NETWORKS, VOL. 8, NO. 12, DECEMBER 2013 2727
© 2013 ACADEMY PUBLISHER