In order to promote an effective and personalized interaction, smart environments should be endowed with the capability of understanding what the user is doing. To this aim we developed a system called WoMan that, using a process mining approach, is able to incrementally learn user’s activities and daily routines as workflow models. In order to test its efficacy in a real-world setting, we set up a smart office environment, SOFiA, equipped with a sensor network based on Arduino. Then we collected an annotated dataset of 45 days and from this dataset we learned the workflow models of the user daily routines and of the activities performed in the office. Then we performed some experiments that show how our approach perform in learning and recognizing activities and routines. In particular, we achieve in average the accuracy of 82% for tasks and the accuracy of 98% for the transitions among tasks. Moreover we test the real-time performance of the approach with sensor data coming from the SOFiA sensors and the system started to make a correct prediction since the fourth execution in 82% of the cases.

Learning and recognizing routines and activities in SOFiA

De Carolis B.;Ferilli S.;Mallardi G.
2014-01-01

Abstract

In order to promote an effective and personalized interaction, smart environments should be endowed with the capability of understanding what the user is doing. To this aim we developed a system called WoMan that, using a process mining approach, is able to incrementally learn user’s activities and daily routines as workflow models. In order to test its efficacy in a real-world setting, we set up a smart office environment, SOFiA, equipped with a sensor network based on Arduino. Then we collected an annotated dataset of 45 days and from this dataset we learned the workflow models of the user daily routines and of the activities performed in the office. Then we performed some experiments that show how our approach perform in learning and recognizing activities and routines. In particular, we achieve in average the accuracy of 82% for tasks and the accuracy of 98% for the transitions among tasks. Moreover we test the real-time performance of the approach with sensor data coming from the SOFiA sensors and the system started to make a correct prediction since the fourth execution in 82% of the cases.
2014
978-3-319-14111-4
978-3-319-14112-1
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11586/304228
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 3
  • ???jsp.display-item.citation.isi??? 3
social impact