Behaviour modelling for detecting the onset of dementia using ambient intelligence

Nidal Al-Beiruti, Khalid Al-Begain

Research output: Contribution to journalArticlepeer-review


Using technology to provide sustainable proactive care for elderly people is an emerging and evolving field of research. Ambient Intelligence (AmI) solutions can be used to collect data from different sensor modalities installed inside homes. The collected data can help to support decision making for clinicians and carers. In addition, it can provide family members with reassurance about their relative’s wellbeing. Behavioural models can be inferred from the collected data. Several currently available solutions try to detect activities before building behavioural models. Our approach is using Hidden Markov Models to build behavioural models directly from raw sensor data without the need of detecting activities first. The model depends on using simple and non-invasive binary sensors. The model is designed to detect abnormalities, sudden or gradual, in daily behaviour which may be considered an indicator of dementia. The model is tested using a real data set and evaluated using generated data.

Original languageEnglish
Pages (from-to)1-10
Number of pages10
JournalInternational Journal of Simulation, Systems, Science and Technology
Issue number5
Publication statusPublished - 1 Oct 2014


  • Ambient Intelligence
  • Behavioural Modelling
  • Hidden Markov Models


Dive into the research topics of 'Behaviour modelling for detecting the onset of dementia using ambient intelligence'. Together they form a unique fingerprint.

Cite this