A falls detection system for the elderly based on a WSN

Accidental falls are one of the main causes of deaths and severe injuries of people over 65 years old. For this reason, the development of fall detection systems for the elderly has been an important research topic. In this paper, a non-invasive fall detection system for older people, based on the use of a wireless sensor network (WSN), is proposed. It uses the acoustic signal sensed by a node of the WSN, as well as signal processing and pattern recognition techniques to detect a fall. The model uses a signal-processing algorithm based on the use of cross-correlation to measure the similarity between the sampled signal and a reference template signal, which represents a fall event. If these two signals are similar, then the Mel-frequency cepstral coefficients (MFCC) of the fall sound are extracted. Afterwards, the dynamic time warping (DTW) method is used for pattern recognition. The evaluation of the proposed system showed a very good detection rate.