Gatica: Linked Sensed Data Enrichment and Analytics Middleware for IoT Gateways

Abstract

Raw sensed data lacks semantics. This poses a challenge to apply analytics directly to raw IoT sensor data. Such operational data requires an intensive enrichment processes to drive value. Pragmatic use of naming conventions and taxonomies can increase the quality of data and make it more interpretable. In this paper, we incorporate semantic and linked data technologies and offer a middleware called Gatica, to dynamically inject semantics to make the raw streaming data of an IoT gateway "Rich" on the device layer. Gatica collects the real-time sensor data, enriches them using annotations then transforms and exposes them in RDF triples, and finally streams RDF objects to the analytic endpoint for querying the linked sensor streaming data. Various analytic applications can utilize our middleware by sending SPARQL requests over the sensor network to our query interface and retrieving the results. Our middleware offers the ability to discover hidden patterns of mutually correlated variables and uncover actionable information within raw data for more utility. This paper details Gatica's architecture together with its implementation.