Object (B)logging: a Decentralized Cognitive Paradigm for the Industrial Internet of Things

This paper proposes object (b)togging, a general framework for managing object networks in the Semantic Web of Things. By applying standard supervised machine learning algorithms on raw sensor data along with non-standard semantic-based reasoning services, this approach enables resource-constrained components to produce rich and meaningful representations of real-world environments and events as in a log. The acquired knowledge is progressively enriched during object’s lifetime and exposed to the outside world as in a blog, in order to trigger complex interactions through an advanced resource discovery. A case study of predictive maintenance and early results support the feasibility of the proposal to Industrial Internet of Things settings.