Home > Research > Publications & Outputs > Towards future situation-Awareness

Associated organisational unit

Links

Text available via DOI:

View graph of relations

Towards future situation-Awareness: A conceptual middleware framework for opportunistic situation identification

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNConference contribution/Paperpeer-review

Published
Close
Publication date13/11/2016
Host publicationQ2SWinet 2016 - Proceedings of the 12th ACM Symposium on QoS and Security for Wireless and Mobile Networks, co-located with MSWiM 2016
Place of PublicationNew York
PublisherAssociation for Computing Machinery, Inc
Pages95-101
Number of pages7
ISBN (electronic)9781450345040
<mark>Original language</mark>English
Event12th ACM Symposium on QoS and Security for Wireless and Mobile Networks, Q2SWinet 2016 - Malta, Malta
Duration: 13/11/201617/11/2016

Conference

Conference12th ACM Symposium on QoS and Security for Wireless and Mobile Networks, Q2SWinet 2016
Country/TerritoryMalta
CityMalta
Period13/11/1617/11/16

Publication series

NameQ2SWinet 2016 - Proceedings of the 12th ACM Symposium on QoS and Security for Wireless and Mobile Networks, co-located with MSWiM 2016

Conference

Conference12th ACM Symposium on QoS and Security for Wireless and Mobile Networks, Q2SWinet 2016
Country/TerritoryMalta
CityMalta
Period13/11/1617/11/16

Abstract

Opportunistic Situation Identification (OSI) is new paradigms for situation-Aware systems, in which contexts for situation identification are sensed through sensors that happen to be available rather than pre-deployed and application-specific ones. OSI extends the application usage scale and reduces system costs. However, designing and implementing OSI module of situation-Aware systems encounters several challenges, including the uncertainty of context availability, vulnerable network connectivity and privacy threat. This paper proposes a novel middleware framework to tackle such challenges, and its intuition is that it facilitates performing the situation reasoning locally on a smartphone without needing to rely on the cloud, thus reducing the dependency on the network and being more privacy-preserving. To realize such intuitions, we propose a hybrid learning approach to maximize the reasoning accuracy using limited phone's storage space, with the combination of two the-state-The-Art techniques. Specifically, this paper provides a genetic algorithm based optimization approach to determine which pre-computed models will be selected for storage under the storage constraints. Validation of the approach based on an open dataset indicates that the proposed approach achieves higher accuracy with comparatively small storage cost. Further, the proposed utility function for model selection performs better than three baseline utility functions.