Wireless Signal Based Elderly Fall Detection Using XGboost Algorithm

With the rapid population ageing and increase of the elderly who live alone, there is a growing demand for intelligent monitoring, especially fall detection systems. In this paper, based on received signal strength (RSS) and machine learning algorithm, a fall detection method is proposed. It using multi-domain features, including time-domain and wavelet-domain, and Boost algorithm trains a model to discriminate fall and other actions, such as, sit, stand and squat. The experimental results show that the proposed method can identify falls well.