Offre de stage : Developing ontology approach for elderly people habits monitoring for frailty detection

Sujet : Developing ontology approach for elderly people habits monitoring for frailty
detection

Laboratoire/équipe : UMR CNRS 7338 Bio­mé­canique et Bio­ingénierie – Equipe C2MUST

Mots clés : Ontolo­gies, ADL, frailty detec­tion, data fusion, uncer­tain­ty

More and more elder­ly peo­ple pre­fer to live at home but secure­ly; while liv­ing at home some sig­nals of frailty may start but are typ­i­cal­ly detect­ed late (e.g., after a dis­tress­ful sit­u­a­tion). By mon­i­tor­ing the person’s habits, we are able to iden­ti­fy the base­line lev­el and to detect slight changes. At the same time, we have more and more sen­sor at home that are used for domotics appli­ca­tions (i.e., light and heat­ing automa­tion); the infor­ma­tion from these sen­sors can be an impor­tant source of infor­ma­tion for mon­i­tor­ing the activ­i­ties of the elder­ly with­out need­ing to install new mon­i­tor­ing sys­tems. The aim of this research is to pro­pose an adap­tive phys­i­o­log­i­cal state mon­i­tor­ing sys­tem for elder­ly peo­ple liv­ing at home using existin­gand new sen­sors at home in order to pro­vide a sus­tain­able sys­tem. All infor­ma­tion from sen­sors will be rep­re­sent­ed by a spe­cif­ic ontol­ogy and used by AI sys­tem in order to detect elder­ly frailty ear­ly but to also detect dis­tress. The ontol­ogy rep­re­sen­ta­tion will allow the easy addi­tion or replace­ment of sen­sors and pro­vide data fusion. The AI sys­tem will be com­posed of spe­cif­ic machine learn­ing approach­es in order to detect dis­tress and frailty, but also help with deci­sion-mak­ing by send­ing syn­thet­ic data to the users, fam­i­ly of the users and med­ical staff. This research work is in the frame­work of an exist­ing col­lab­o­ra­tion between lab­o­ra­to­ries from Sor­bonne Uni­ver­sité – SU (BMBI, GEEPS, LIP6) and Water­loo Uni­ver­si­ty (Ubi­lab, CBB). Two exper­i­men­tal data bases are avail­able: the first one is on the eBio­Med plat­form (UTC), a ful­ly equipped lab­o­ra­to­ry with dif­fer­ent sen­sors allow­ing for sce­nar­ios test­ing; the sec­ond one is at the Research Insti­tu­tion for Aging (RIA) Water­loo, which has a ful­ly equipped apart­ment and uses elder­ly vol­un­teers and who can live there for a spec­i­fied period.