230
B. Di Martino and A. Esposito
plexity of the networks which can be built upon such technology, issues are bound
to arise. One of the main problems which needs to be addressed regards the ability of sensors and devices, coming from different manufacturers, to immediately
interoperate with each other. Whereas the adoption of open standards for devices’
interconnection such as OPC-UA [12] or OMG Data-Distribution Service for RealTime Systems (DDS) [17] could solve, or at least ease, the interoperability problem,
often manufacturers prefer to resolve to proprietary solutions, or they simply introduce variations to differentiate their products. Another aspect to be considered is
the possibility to automatically build and manage complex sensor topologies, given
a set of specific requirements and a description of the sensors’ functionalities and
interfaces. Semantic technologies have proved in the past to be particularly suitable
for the representation of heterogeneous data, and their consequent homogenization,
and have also been adopted in the representation of complex systems and services [8,
15]. The application of semantic technologies to the representation of sensors’ API
can lead to interesting results and lead to the resolution of possible interoperability
issues which may arise when trying to compose complex sensor networks. Also,
semantic matchmaking can represent the base for advanced sensor composition and
the management and orchestration of services based on smart devices. This paper
aims at describing a semantic based framework which, starting from the description
of Rest-full interfaces used to communicate with smart sensors, is able to derive a
semantic representation of the sensors’ functionalities and input/output parameters.
Such a representation is then used to match different sensors and identify possible
combinations thereof, having a specific goal in mind, and to match their input/output
data. The presented tool is an improvement of the framework presented in [9]. The
remainder of this paper is organized as follows: Sect. 2 reports a State of the Art
in the field of sensors’ interoperability; Sect. 3 introduces the semantic technologies
used to implement the tool; Sect. 4 provides a description of the methodology applied
to build the tool; Sect. 5 focuses on the creation of the Graph behind the semantic
representation of the sensors’ API; Sect. 6 provides information on the WSD and
OWL-S generation; Sect. 7 closes the paper with final considerations.
2 State of the Art
Interoperability among services and sensors provided by different vendors is considered a hot topic, and related problems have been addressed at various levels, generally
involving the device, network and data levels. To tackle such issues, one of the most
desirable solutions to possibly enable complete interoperability is to adopt shareable
and open (when available) standards. This could also represent a good opportunity
for consortiums and organizations of difference sizes to engage in the development
of new technologies, imposing de-facto standards in an ever changing and chaotic
market.
The OPC Foundation developed the OPC Unified Architecture (OPC-UA) [14],
a communication technology completely based on TCP/IP, belonging to the family
of Machine to Machine (M2M) communication protocols, that allows sensors and
B. Di Martino and A. Esposito
plexity of the networks which can be built upon such technology, issues are bound
to arise. One of the main problems which needs to be addressed regards the ability of sensors and devices, coming from different manufacturers, to immediately
interoperate with each other. Whereas the adoption of open standards for devices’
interconnection such as OPC-UA [12] or OMG Data-Distribution Service for RealTime Systems (DDS) [17] could solve, or at least ease, the interoperability problem,
often manufacturers prefer to resolve to proprietary solutions, or they simply introduce variations to differentiate their products. Another aspect to be considered is
the possibility to automatically build and manage complex sensor topologies, given
a set of specific requirements and a description of the sensors’ functionalities and
interfaces. Semantic technologies have proved in the past to be particularly suitable
for the representation of heterogeneous data, and their consequent homogenization,
and have also been adopted in the representation of complex systems and services [8,
15]. The application of semantic technologies to the representation of sensors’ API
can lead to interesting results and lead to the resolution of possible interoperability
issues which may arise when trying to compose complex sensor networks. Also,
semantic matchmaking can represent the base for advanced sensor composition and
the management and orchestration of services based on smart devices. This paper
aims at describing a semantic based framework which, starting from the description
of Rest-full interfaces used to communicate with smart sensors, is able to derive a
semantic representation of the sensors’ functionalities and input/output parameters.
Such a representation is then used to match different sensors and identify possible
combinations thereof, having a specific goal in mind, and to match their input/output
data. The presented tool is an improvement of the framework presented in [9]. The
remainder of this paper is organized as follows: Sect. 2 reports a State of the Art
in the field of sensors’ interoperability; Sect. 3 introduces the semantic technologies
used to implement the tool; Sect. 4 provides a description of the methodology applied
to build the tool; Sect. 5 focuses on the creation of the Graph behind the semantic
representation of the sensors’ API; Sect. 6 provides information on the WSD and
OWL-S generation; Sect. 7 closes the paper with final considerations.
2 State of the Art
Interoperability among services and sensors provided by different vendors is considered a hot topic, and related problems have been addressed at various levels, generally
involving the device, network and data levels. To tackle such issues, one of the most
desirable solutions to possibly enable complete interoperability is to adopt shareable
and open (when available) standards. This could also represent a good opportunity
for consortiums and organizations of difference sizes to engage in the development
of new technologies, imposing de-facto standards in an ever changing and chaotic
market.
The OPC Foundation developed the OPC Unified Architecture (OPC-UA) [14],
a communication technology completely based on TCP/IP, belonging to the family
of Machine to Machine (M2M) communication protocols, that allows sensors and