Evaluation of a threshold-based tri-axial accelerometer fall detection algorithm.

A K Bourke; J V O'Brien; G M Lyons

doi:10.1016/j.gaitpost.2006.09.012

Home/Archive/PubMed

PubMed · 17101272

Evaluation of a threshold-based tri-axial accelerometer fall detection algorithm.

A K Bourke·J V O'Brien·G M Lyons

View original source ↗Export metadata ↓

Abstract

Using simulated falls performed under supervised conditions and activities of daily living (ADL) performed by elderly subjects, the ability to discriminate between falls and ADL was investigated using tri-axial accelerometer sensors, mounted on the trunk and thigh. Data analysis was performed using MATLAB to determine the peak accelerations recorded during eight different types of falls. These included; forward falls, backward falls and lateral falls left and right, performed with legs straight and flexed. Falls detection algorithms were devised using thresholding techniques. Falls could be distinguished from ADL for a total data set from 480 movements. This was accomplished using a single threshold determined by the fall-event data-set, applied to the resultant-magnitude acceleration signal from a tri-axial accelerometer located at the trunk.

Explore related subjects

Acceleration Accidental Falls Activities of Daily Living Aged Aged, 80 and over Biomechanical Phenomena Female Humans Male Monitoring, Ambulatory Movement Sensitivity and Specificity Evaluation Study Journal Article

Keep this discovery

Explore connections, maps & timelines

BibTeX RIS

A K Bourke, J V O'Brien, G M Lyons. 2006-11-13. Evaluation of a threshold-based tri-axial accelerometer fall detection algorithm.. https://doi.org/10.1016/j.gaitpost.2006.09.012

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related citations

PubMed2007

The role of attention in horizontal curves: a comparison of advance warning, delineation, and road marking treatments.

Horizontal curves have been recognized as a significant safety issue for many years, a more important factor than road width or sight distance. The research literature suggests that driver errors associated with horizontal curves result from three inter-related problems: failures of driver attention, misperceptions of speed and curvature, and poor lane positioning. To examine the roles of attentional, perceptual, and lane placement factors in driver behaviour at curves, two groups of curve treatments were identified for testing with a driving simulator. The first group of treatments consisted of four combinations of warning signs designed to alert drivers to the presence of curves and produce a reduction in curve approach speeds. The second group was comprised of several types of road markings designed to affect drivers' speed and lane position as they drove through curves. The results indicated that advance warning signs by themselves were not as effective at reducing speeds as when they were used in conjunction with chevron sight boards and/or repeater arrows. Of the road marking treatments only rumble strips produced any appreciable reductions in speed. A herringbones road marking was found to produce significant improvements in drivers' lane positions, effectively flattening the drivers' paths through the curves. A treatment combining the herringbones treatment with chevron and repeater arrow signs produced both a reliable reduction in speed as well as improved lane positions. The results are interpreted as evidence that treatments that highlight perceptual cues are the most effective means of moderating drivers' curve speeds.

Acceleration↗

PubMed2007

Uppsalator's acceleration.

Semiquantitative prediction of the Uppsalator (electroosmotical oscillator developed in Uppsala) under constant voltage is confirmed via numerical solution of exact nonlinear integro-differential equations in partial derivatives (PDE). Critical voltage and characteristic features of pressure and liquid flow velocity as functions of time are reproduced with high precision. In particular, the number of inflection points per period of oscillating velocity increases with growing voltage from two to six. A new feature is demonstrated. For sufficiently high voltages the Uppsalator shows normal deceleration, i.e., the oscillation period grows with voltage. However, in some supercritical range of voltages, acceleration is observed, as expressed by decrease of the oscillation period with growing voltage. Thus, the Uppsalator at constant voltage considered earlier as an impossible phenomenon represents a great challenge for experimental verification.

Acceleration↗

PubMed2007

Photodissociation at 193 nm of some singly protonated peptides and proteins with m/z 2000-9000 using a tandem time-of-flight mass spectrometer equipped with a second source for delayed extraction/post-acceleration of product ions.

A tandem time-of-flight mass spectrometer was built for photodissociation (PD) of singly protonated peptides and small proteins generated by matrix-assisted laser desorption/ionization. PD was performed in a second source after deceleration of precursor ions. The delayed extraction/post-acceleration scheme was used for the product ions. For the PD at 193 nm of small singly protonated peptides, the present instrument showed much better sensitivity and resolution for product ions than the previous one (Moon JH, Yoon SH, Kim MS, Bull. Korean Chem. Soc. 2005; 26: 763) even though the overall spectral patterns obtained with the two instruments were similar. The present instrument was inferior in precursor ion selection and background noise level. PD was achieved for precursor ions as large as the singly protonated ubiquitin (m/z 8560.63), indicating that the photoexcitation is capable of supplying a sufficient amount of internal energy to dissociate large singly protonated proteins. As the precursor ion m/z increased, however, product ion signals deteriorated rather rapidly. As in the PD of small peptide ions with m/z around 1000, the types of the product ions generated from singly protonated peptides with m/z in the range 2000-4000 were mostly determined by the positions of arginine residues. Namely, a(n) and d(n) ions dominated when an arginine residue(s) was near the N-terminus while v(n), w(n), x(n) and y(n) dominated when the same residue(s) was near the C-terminus. In addition, d(n), v(n) and w(n) ions were generated according to the correlation rules previously observed in the collisionally activated dissociation. Isoleucine and leucine isomers could be easily distinguished based on the w(n) and d(n) ions.

Acceleration↗