Today’s wide range of smart personal devices and online services generate a constant stream of information that can indicate a person’s plans, activities, and situations.
By treating these devices as part of a personal sensor network
and analyzing the generated information collectively, valuable context information can be gathered and interpreted in
an endless number of scenarios. Efforts by both private industry and research communities have produced a wide variety
of similar context-aware techniques and services. However,
this technology remains strictly tied to a native application,
system, or device, thus limiting its reuse and integration to
address new scenarios.
To address the above problem, we propose a semantic
infrastructure for context-aware systems, based on machine-processable ontologies that cater for personal information
scattered across multiple personal sources. Moreover, the
semantic infrastructure also covers context-related information obtained from the personal sensor network. Thus, context information created by numerous disconnected platforms, including social networking services (for example,
check-ins, tags), physical data sensors (for example, GPS),
and legacy tools and applications (for example, calendar),
can all be integrated and processed centrally. Domain models covering some of these personal information domains
Copyright © 2015, Association for the Advancement of Artificial Intelligence. All rights reserved. ISSN 0738-4602
A Semantic Infrastructure
Simon Scerri, Jeremy Debattista,
Judie Attard, Ismael Rivera
; Although a number of initiatives provide personalized context-aware guidance for niche use cases, a standard
framework for context awareness
remains lacking. This article explains
how semantic technology has been
exploited to generate a centralized
repository of personal activity context.
This data drives advanced features such
as personal situation recognition and
customizable rules for the context-sensitive management of personal devices
and data sharing. As a proof of concept,
we demonstrate how an innovative context-aware system has successfully
adopted such an infrastructure.