Fall Detection for Independently Living Older People Utilizing Machine Learning

Abstract

The population of the developed nations is ageing. Several studies have been conducted, identifying different issues, each highlighting the need to enhance our current care systems for older people. A major threat faced by independently living older people is falling accidents. Robust, reliable and unobtrusive fall-detection is needed to counter the threat. This paper presents a fall-detection scheme utilizing bio-inspired asynchronous temporal contrast sensors and support vector machine (SVM), a type of machine learning to realize such a system. We focus on the assembly of the training and test data, the training process of the SVM and on presenting and interpreting the obtained results.