Object (b)logging: Semantically rich context mining and annotation in pervasive environments

Abstract

Object (b)logging is proposed as a novel general framework for the Semantic Web of Things, based on an evolution of conventional Web of Things paradigms. The advanced performance and the miniaturization of sensors allow to acquire several environmental parameters for event and phenomenon detection in many operational contexts. By leveraging the integration of standard supervised machine learning techniques with non-standard semantic-based reasoning services, smart objects annotate in a fully automatic way the context they are in, continuously enriching their descriptive core based on events they detect. Finally they expose them to the outside world as in a blog. The feasibility of the proposed framework is supported by a case study and an early experimental campaign.