The University of New Brunswick

Published purpose

Falls and accidents are the leading cause for injury-related hospitalizations among Canadian seniors, and often become a main obstacle to independent living. Falls are often caused by a combination of multiple factors. For example, a decline in vision in combination with limited mobility due to arthritis might contribute to a stumble over a carpet edge. To prevent falls, the person-specific risk factors need to be addressed as early as possible. Commonly, movement and activity assessments for fall predictions are performed in clinics or emergency departments and therefore provide only single snapshots of the person’s movement pattern. This method has limited ability to detect small changes in mobility and movement patterns. The aim of this project is to investigate fully automatic methods to detect multifactorial falls in an in-home environment. Data from different in-home sensors, such as cameras or movement detectors, will be used to analyze movement and activity patterns for each inhabitant of the household to detect abnormalities related to an increased risk of falling or accidents. Output of these methods can support early targeted interventions, with a possible decrease of the number of falls of seniors living in community dwellings. Therefore, the project has the potential to reduce fall-related premature nursing home admission, which is in alignment with the overall objective of the AiP Challenge Program.