JUCS - Journal of Universal Computer Science 19(8): 1105-1122, doi: 10.3217/jucs-019-08-1105
Automatic Detection of Falls and Fainting
expand article infoJuan E. Garrido, Victor M.R. Penichet, María D. Lozano, José A. F. Valls
‡ University of Castilla-La Mancha, Albacete, Spain
Open Access
Abstract
Healthcare environments have always been considered an important scenario in which to apply new technologies to improve residents and employees conditions, solve problems and facilitate the performance of tasks. In this way, the use of sensors based on user movement interaction allows solving complicated situations that should be immediately addressed, such as controlling falls and fainting spells in residential care homes. However, ensuring that all the residents are always visually controlled by at least one employee is quite complicated. In this paper, we present a ubiquitous and context-aware system focused on geriatrics and residential care homes, but it could be applied to any other healthcare centre. This system has been designed to automatically detect falls and fainting spells, alerting the most appropriate employees to address the emergency. To that end, the system is based on movement interaction through a set of Kinect devices that allows the identification of the position of a person. These devices imply some development problems that authors have had to deal with, including camera location, the detection of head movements and people in horizontal position. The proposed system allows controlling each resident posture through a notification and warning procedure. When an anomalous situation is detected, the system analyses the resident posture and, if necessary, the most adequate employee will be warned to react urgently. Ubiquity and context-awareness are essential features since the proposed system has to be able to know where any employee is and what they are doing at any time. Finally, we present the outcomes of an evaluation based on the ISO 9126-4 about the usability of the system.
Keywords
collaboration, ubiquity, context-awareness, healthcare, Movement, Interaction, Kinect