Lightweight context-aware activity recognition

Document Type

Conference Proceeding

Source of Publication

Lecture Notes in Electrical Engineering

Publication Date

1-1-2016

Abstract

© Springer-Verlag Berlin Heidelberg 2016. In ubiquitous environments, it is important to recognize the situation and deliver services accordingly. In addition, it is equally important to have a fast response time. The existing context-aware activity recognition engines have good recognition rates; however, they consume lots of time to produce feasible results. Our focus in this research is to reduce the time required by eliminating the need for ontology matching (in context-aware activity manipulation engine) and extend the rules. In addition, we incorporate the sliding time window concept to retain activities for a longer duration and maintain their relevance using ontological data for a better accuracy. The proposed scheme has increased the overall accuracy against the existing system by 12.6 % for individual activities relevance and 6 % for high level activities.

ISBN

9783662478943

ISSN

1876-1100

Publisher

Springer Verlag

Volume

354

First Page

367

Last Page

373

Disciplines

Computer Sciences

Keywords

Activity recognition, Knowledgebase, Ontology

Scopus ID

84947246965

Indexed in Scopus

yes

Open Access

no

Share

COinS