SIGHTED: A Framework for Semantic Integration of Heterogeneous Sensor Data on the Internet of Things

Sensors are embedded nowadays in a growing number of everyday life objects. Smartphones, wearables, and sensor networks together play an important role in bridging the gap between physical and cyber worlds, a fundamental aspect of the Internet of Things vision. The ability to reuse sensor data integrated from multiple heterogeneous sources is a step towards building innovative applications and services. In this paper SIGHTED, a sensor data integration framework, is proposed exploiting semantic web technologies and linked data principles. It provides a layered structure as a guideline for integrating sensor data from various sources supporting accessibility and usability. DotThing, a demo platform, is implemented as an instantiation of SIGHTED framework and evaluated. Smartphones and sensor nodes are connected to DotThing showing the ability to query and reuse integrated sensor data from multiple sources to create more flexible horizontal applications. DotThing implementation also demonstrates the need for adding a semantic layer to existing IoT cloud-based platforms, like Xively, that generally lack such layer resulting in proprietary vertical solutions with limited data integration and discovery capabilities. DotThing makes use of vocabularies from existing ontologies on the linked data cloud providing a unified model to annotate data and link it to existing resources on the web. (C) 2016 The Authors. Published by Elsevier B.V.