Fall event analysis based on sensor fusion (análisis de eventos de caída basado en la fusión de sensores)

Resumen

Fall events represent a major problem in today’s time for the society, since efficient methods need to be researched to reliably detect these events in order to minimize the resulting risk of injuries. These events can be caused by a variety of reasons, such as health-related or event-related issues, that if not timely intervened can result in serious injuries or, in the worst case, to death. This risk is present in every day life and depending on the daily circumstances some people are exposed to a higher risk of suffering fall events. Especially elderly people living alone are at risk, but also occupational groups due to their hazardous working conditions. Considering the situation in nursing homes where not enough nursing personnel is available, providing fast assistance in case fall events occur is problematic due to the under-resourced facilities. Another aspect of falls is not only the detection, but additionally to automatically trigger the emergency. This is an essential feature that needs to be considered, because in case the person suffers a severe fall leading to unconsciousness, the person will not be able to interact with the system for calling help. To solve the fall phenomena, a solution has to be provided which takes into account the abovementioned aspects. It has to be differentiated between a location-dependent and location-independent solution, preferring the second solution since it provides a wide range usage by using wearable sensors. To facilitate a location-independent, efficient and fast assistance, a portable sensor-based solution has been developed during this dissertation which is intended to serve for different use cases, like in nursing homes and for firefighters to provide fast fall detection assistance or for medical specialists as a research instrument to examine the causes and dynamics of these fall events. For ensuring an efficient and real time-based analysis the proposed fall detection system incorporates a distributed system architecture.