Smart Sound Sensor for distress situation and number of presented person

In order to allow older adults to active and healthy ageing with comfort and security we have already de veloped a smart audio sensor being able to recognize everyday life sounds in order to detect activities of daily living (ADL) and distres s situations. In this paper, we propose to add a new functionality by ana lyzing the speech flow in order to detect the number of person in a room. The proposed algorithms are based on speaker diarizatio n methods. This information is useful in order to better detect act ivities of daily life but also to know when the person is home alone. This functionality can also offer more comfort through light, heating and air conditioning adaptation to the number of persons.