Rowan University
ALGORITHMS FOR RAPID TRIP DETECTION IN HUMAN WALKING USING INERTIAL SENSORS
Abstract
dc:description.abstractTrips and falls are one of the most common causes for injuries among elderly. The existing trip-and-falls studies primarily focus on the proactive fall prevention approaches, while active prevention strategies remain largely unexplored. This thesis aims to provide first steps towards active trip-and-fall prevention by developing various algorithms capable of detecting trip in human walking faster than the human voluntary reactions (~200 ms). The measurements of human kinematics are used as the inputs in the algorithms. The proposed algorithms include three threshold-based detection methods, an optimized elastic time-series alignment tool called dynamic time warping (DTW) that overcomes problems of time dependency and signal phasing, and an optimized signal-frequency decomposition method known as continuous wavelet transform (CWT) that allows signal analysis in time-frequency domain. In addition, a planar covariation law (PCL), which is an intersegmental co-ordination kinematic law that relates lower limb angles during walking, is explored for trip gaits for the first time. A quantification of trip perturbation as a function of perturbation duration was developed to provide an insight about trip evolution and severity. The developed rapid trip detection algorithms integrated and combined with assistive technologies have the potential to serve as an enabling tools of future trip-and-fall prevention strategies.
Degree
thesis:*- Name thesis:degree_name
- M.S. Mechanical Engineering
- Level thesis:degree_level
- Thesis
- Discipline thesis:degree_discipline
- Mechanical Engineering
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ellouzi, Chadi
- Contributors dc:contributor
-
- Behrad Koohbor, Ph.D.
- Chen Shen, Ph.D.
Subjects
dc:subject × 8Identifiers
dc:identifier.*- Repository record dc:identifier
- https://rdw.rowan.edu/etd/2962