Real-time elderly activity monitoring system based on a tri-axial accelerometer

In this study, we developed the automatic human movement classification system for the elderly using only single waist-mounted tri-axial accelerometer. The system can distinguish several activities such as fall, walking, running, standing, lying and sitting and transition between each movement in real-time. To evaluate proposed algorithm which utilizes the acceleration and tilt information from the sensor module, experiments were performed on ten healthy subjects with several activities such as falls, walking, running, sit to stand, stand to sit, stand to lie, lie to stand, etc. The successful human movement detection rate of the system was 96.1%. For further improvement of the system, it is necessary to include more detailed classification algorithm to distinguish several daily activities and to carry out actual experiments with the elderly.