Real-Time Fall Detection Using Autoregressive Integrated Moving Average(Arima)

Over the last few years, fall detection has emerged as an essential field of study, given the implications it has on the protection of the health of older adults and other populations susceptible to uninformed falls. This research focuses on a new, real time fall detection system that employs the Autoregressive Integrated Moving Average model (ARIMA) in the detection of falls among many people and many cameras. The onward movement covers the application of ARIMA in temporal data analysis. The proposed system deals with capturing the models of movement and the relevant anomalies which occur during the period of a fall. Our approach is commendable since it is highly scalable and flexible to different surveillance systems and therefore effective monitoring can be done in different places without intrusive means. The results of the experiments carried out show that the model has a high ability and efficacy in distinguishing falls from other activities, thus making the model feasible for real time applications where the processing speed is fast, and the response is instantaneous. This research proves that ARIMA based fall detection can be useful in multi-person scenarios paving the way to risk-free systems in the real world.