Elderly people mostly stay at home due to illness and
other conditions and have to live independently. The number
of elderly people is growing every year, as World Health
Organization has forecasted that the elder population will
reach a total of 2 billion by 2050 [1]. Systems that provide
elder care focus on detecting distress situations, where a
person cannot move or reach the remote. To achieve this,
video, sound, fall and motion detection techniques have been
used. In our work, we focus only on sound for detection of
abnormal situations that may occur inside home. A popular
trend is to use multiple sensors and combine their outputs for
detection [2]. An opposite approach is to use the output of
only one sensor. Our aim is to reduce the cost of equipment
involved in elder care and still provide the best results. So,
sound is the most natural choice of channels to achieve this
goal. It carries a huge amount of information about the
elderly people and their surroundings. Aged persons in
distress situation can make different sounds like crying or
glass breaking etc. that can be detected and classified through
our system.