234
B. Di Martino and A. Esposito
Fig. 1 The applied methodology
published standards such as SSN and SensorML,
8 which describes both smartphones
and sensors from different points of view, including the specific platforms they are
supported by, their measurements capabilities and observations, taking also in consideration their context model.
3 The Methodology
The approach applied in the development of the tool is composed of a series of consecutive steps, which are graphically summarized in Fig. 1. The first fundamental
step is represented by the analysis of the sensors’ API, which represents the base to
populate an ontology of the sensors’ functionalities, inputs and outputs. This step is
essential, as without a shared machine-readable knowledge base, it would be impossible to actually match sensors and try to build new topologies. The analysis must take
in consideration both input and output parameters of the sensors’ API, as they will
represent the “glue” between different types of devices. The result of the analysis is
represented by an ontological representation, in OWL, which contains the information extracted from the API in a machine-readable format, and organized according
to a structured graph, the so called API Graph. The information contained in the
graph are complemented with a WSDL representation of the API, which is used as
the grounding of an OWL-S representation. Both the API ontology and WSDL are
obtained via the API analysis, whereas the OWL-S representation is created afterwards, and only if the representation of API requires dynamic information regarding,
for an instance, the flow of steps needed to operate the sensors. The OWL-S representation becomes essential when sensors are composed, in order to manage their
orchestration.
8 Sensor Model Language (SensorML)—https://www.ogc.org/standards/sensorml.
B. Di Martino and A. Esposito
Fig. 1 The applied methodology
published standards such as SSN and SensorML,
8 which describes both smartphones
and sensors from different points of view, including the specific platforms they are
supported by, their measurements capabilities and observations, taking also in consideration their context model.
3 The Methodology
The approach applied in the development of the tool is composed of a series of consecutive steps, which are graphically summarized in Fig. 1. The first fundamental
step is represented by the analysis of the sensors’ API, which represents the base to
populate an ontology of the sensors’ functionalities, inputs and outputs. This step is
essential, as without a shared machine-readable knowledge base, it would be impossible to actually match sensors and try to build new topologies. The analysis must take
in consideration both input and output parameters of the sensors’ API, as they will
represent the “glue” between different types of devices. The result of the analysis is
represented by an ontological representation, in OWL, which contains the information extracted from the API in a machine-readable format, and organized according
to a structured graph, the so called API Graph. The information contained in the
graph are complemented with a WSDL representation of the API, which is used as
the grounding of an OWL-S representation. Both the API ontology and WSDL are
obtained via the API analysis, whereas the OWL-S representation is created afterwards, and only if the representation of API requires dynamic information regarding,
for an instance, the flow of steps needed to operate the sensors. The OWL-S representation becomes essential when sensors are composed, in order to manage their
orchestration.
8 Sensor Model Language (SensorML)—https://www.ogc.org/standards/sensorml.