Fall detection with body-worn sensors
- PDF / 853,350 Bytes
- 14 Pages / 595.276 x 790.866 pts Page_size
- 26 Downloads / 244 Views
ickert1 · C. Becker1 · U. Lindemann1 · C. Maréchal1 · A. Bourke2 · L. Chiari3 · J.L. Helbostad4, 5 · W. Zijlstra6 · K. Aminian2 · C. Todd7 · S. Bandinelli8 · J. Klenk1, 9 · for the FARSEEING Consortium and the FARSEEING Meta Database Consensus Group 1 Department of Clinical Gerontology, Robert-Bosch Hospital, Stuttgart 2 Laboratory of Movement Analysis and Measurement, Ecole Polytechnique Fédérale de Lausanne (EPFL) 3 Department of Electrical, Electronic, and Information Engineering, University of Bologna 4 Department of Neuroscience, NTNU, Trondheim 5 Department of Geriatrics, St. Olav University Hospital, Trondheim 6 Institute of Movement and Sport Gerontology, German Sport University Cologne 7 School of Nursing, Midwifery & Social Work, The University of Manchester 8 ASF - Azienda Sanitaria di Firenze, Florence 9 Institute of Epidemiology and Medical Biometry, Ulm University
Fall detection with body-worn sensors A systematic review
Falls in older people remain a major public health challenge [1]. Most of the current knowledge on risk factors is derived from epidemiological studies [2], interviews [3] and intervention studies [4]. To date, the contribution of objective measurements using inertial sensor-based technology to improve the understanding of the underlying mechanisms and kinematics of falls is modest. Recently, a Canadian research group demonstrated that video footage can fill in some of the knowledge gaps pertaining to the contextual factors of falls and generate new hypotheses for the design of possible prevention strategies [5]. However, it is clear that video footage also has its limitations and that other technological approaches such as ambient sensing, mobile health devices [6] and body-worn sensors will be needed to generate a clearer picture [7]. Falls are known to be multifactorial incidents caused by a complex interaction of intrinsic problems, extrinsic factors and exposures [8]. Falls occur during different activities, including lying, sitting, standing, walking or running [5, 9]. Fall direction differs widely and might actually change during the fall due to a defen-
706 |
sive response, e.g. to avoid head contact. Responses to falling, such as stepping or grasping, may lead to a “broken fall”, i.e. multiple impacts during a single fall. Up until now, the published technological approaches to fall detection appear to have been very heterogeneous, employing a variety of devices and algorithms [10]. For example, many studies have not used an explicit fall definition [11] or specific fall model [12, 13, 14], nor have they carefully specified the technology being used. The motivation for this review is embedded in the FARSEEING project (FAll Repository for the design of Smart and sElf‐adaptive Environments prolonging INdependent livinG, EU Grant agreement no.: 288940; http://farseeingresearch.eu/). One goal of this project is to generate a
large metadatabase of real-world fall signals. Among the objectives of the database is the development of new algorithmic approaches for the prediction,
Data Loading...
