Reconnaissance des comportements d'une personne âgée vivant seule dans un habitat intelligent pour la santé
Xuan Hoa Binh Le · HAL (Le Centre pour la Communication Scientifique Directe) · 2008
In France, in 2008, the number of people aged 60 and over is approximately 13.5 million. It is estimated at 22.3 million in 2050. Most elders live at home (96% of people aged 60 and over in 1999) and many live alone at home (27% in 1999). Elders living alone at home with aging have some risks at long-term (bad feeding, insufficient hygiene, loss of autonomy ...). To ensure a minimum safety for elders who wish to live at home as long as possible, we can install sensors in their home to obtain a Health Smart Home, also called “Habitat Intelligent pour la Santé” (HIS). This installation allows a remote monitoring of the occupant. Through an automatic recognition of daily activities performed by the occupant, our objective is to detect a possible loss of autonomy. We are interested in HIS which are equipped only with non-invasive, non-intrusive sensors (infra-red sensors and/or magnetic contacts). This type of HIS, simple and respectful of the privacy of the person, functions in a passive way, i.e. without needing the participation of the occupant. The work of the thesis proposes a method of processing data provided by passive infrared sensors installed in a HIS, in order to recognize the activities of daily living (or ADL) performed by the elder in a day, and monitor the evolution of his level of autonomy. This method has been written in Matlab and applied to real data coming from HIS occupied by elders living alone.