Mobile-phone based immobility tracking system for elderly care

Over the last few decades, the world population structure has changed dramatically where the number of elderly people increases and that of young people decreases. The number of elderly people has now quadrupled from its number in 1950. It is projected to reach around the 2 billion by 2050. These statistics strongly indicate the fact that we are entering a super-aged society. Elderly people are usually encountered several problems resulting from age-related diseases. One of the major problems in the elderly is a decline in physical ability, generally called Immobility. Immobility consequently leads to a decline in ability to carry out Activities of Daily Living (ADLs), increased falls and other medical problems. It has significant effects on both physiological and psychological conditions of elderly people. To support medical staffs and caregivers, we propose an immobility tracking system that can analyze physical activities and identify immobility behavior of the elderly in timely fashion. Our system exploits an activity recognition service as a stream component. It utilizes acceleration force and angular velocity data derived from mobile-phone worn on the waist of the elderly to remotely monitor mobility-related activities of the elderly in real-time. The experimental results show that our proposed system achieves a high activity recognition rate - about 93.5% for overall accuracy with reasonable processing overheads.