Software simulation of unobtrusive falls detection at night-time using passive infrared and pressure mat sensors

摘要: Falls and their related injuries are a major challenge facing elderly people. One serious issue to falls among the living at home is ‘long-lie’ scenario, which inability get up from floor after fall, followed by lying on for 60 minutes, or more. Several studies of accelerometer gyroscope-based wearable detection devices have been cited in literature. However, when subject moves around night-time, such as making trip bedroom toilet, it unlikely that they will remember even feel an inclination wear device. This research investigate potential usefulness unobtrusive fall system, based use passive infrared sensors (PIRs) pressure mats (PMs), detect automatically recognizing unusual activity sequences environment; hence, decreasing number subjects suffering scenario fall. A Java-based wireless sensor network (WSN) simulation was developed. reads room coordinates residential map, path-finding algorithm (A*) simulates subject's movement through environment, PIR PM respond binary manner movement. The tested four scenarios; one including activities daily (ADL) three scenarios simulating falls. simulator generates movements ten people (5 female 5 male; age: 50–70 years; body mass index: 25.85–26.77 kg/m2). decision tree heuristic classification model used analyze data differentiate events normal activities. sensitivity, specificity accuracy 100%, 66.67% 90.91%, respectively, across all scenarios.