Towards Sensor Failure Detection in Ambient Assisted Living: Sensors Correlations

Naney E. Elhady, Julien Provost

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

Ambient Assisted Living promotes healthy independent ageing of the elderly at their homes by monitoring their behaviour, and support medical assistance whenever needed. For privacy and acceptance issues, non-intrusive sensors are preferably used. However, such sensors are more prone to produce false positive or negative data. Faulty sensor data could be automatically detected if correlations between sensors can be identified. This paper aims to propose the use of association rule mining to find correlations between binary event-driven sensors installed for monitoring purposes in an apartment. A case study was carried out to validate the approach and investigate the effect of different data mining parameters on the quality of obtained association rules. The results show that correlations could be successfully deduced from unlabelled datasets with no prior expert knowledge on the sensors topology.

Original languageEnglish
Title of host publication2018 IEEE 14th International Conference on Automation Science and Engineering, CASE 2018
PublisherIEEE Computer Society
Pages378-383
Number of pages6
ISBN (Electronic)9781538635933
DOIs
StatePublished - 4 Dec 2018
Event14th IEEE International Conference on Automation Science and Engineering, CASE 2018 - Munich, Germany
Duration: 20 Aug 201824 Aug 2018

Publication series

NameIEEE International Conference on Automation Science and Engineering
Volume2018-August
ISSN (Print)2161-8070
ISSN (Electronic)2161-8089

Conference

Conference14th IEEE International Conference on Automation Science and Engineering, CASE 2018
Country/TerritoryGermany
CityMunich
Period20/08/1824/08/18

Fingerprint

Dive into the research topics of 'Towards Sensor Failure Detection in Ambient Assisted Living: Sensors Correlations'. Together they form a unique fingerprint.

Cite this