Automatic Fall Detection Based on Kinematic Characteristics during the Pre-impact Phase of Falls

Advanced age is often associated with an increased risk of accidental falls. Injuries such as hip fractures as a result of accidental falls in the elderly often lead to functional impairment, declined health, and loss of independence [1].

Fall-related hip fractures may be prevented by hip protection pads [2]. However, currently available hip pads are not widely accepted by elders because of cosmetic and comfort reasons [3]. Alternatively, an active protection apparatus - an airbag-based smart hip pad - would provide both an effective protection against the large impact when a fall occurs and the attractive cosmetic look when not sustaining a fall [4, 5]. One of the critical elements of the smart hip pad is the automated fall detection mechanism.

The goals of this study were to (1) examine and distinguish the kinematic characteristics of impending falls from those of routine physical activities, and (2) develop a single sensor system to reliably detect falls before the impact occurs. It was hypothesized that the inertial frame vertical velocity near the body center of gravity would be significantly higher during the pre-impact phase of falls than regular daily activities, and that a single inertial sensor would be able to distinguish an in-progress and unrecoverable fall from non-falling activities.