Step detection using multi- versus single tri-axial accelerometer-based systems

E. Fortune, Va Lugade, Shreyasee Amin, Kenton R Kaufman

Research output: Contribution to journalArticle

11 Scopus citations

Abstract

Multiple sensors are often considered necessary for increased step count accuracy. However, subject adherence to device-wear increases using a minimal number of activity monitors (AMs). The study aims were to determine and compare the validity of using multiple AMs versus a single AM to detect steps by comparison to video using a modification of an algorithm previously developed for a four-accelerometer AM system capable, unlike other algorithms, of accurate step detection for gait velocities as low as 0.1 m s-1. Twelve healthy adults wore ankle, thigh and waist AMs while performing walking/jogging trials at gait velocities from 0.1-4.8 m s-1 and a simulated free-living dynamic activities protocol. Nineteen older adults wore ankle and waist AMs while walking at velocities from 0.5-2.0 m s-1. As little as one AM (thigh or waist) accurately detected steps for velocities >0.5 m s-1. A single ankle AM accurately detected steps for velocities 0.1 m s-1. Only the thigh AM could not accurately detect steps during the dynamic activities. Only the thigh-ankle combination or single waist AM could accurately distinguish between walking and jogging steps. These laboratory-based results suggest that the presented algorithm can accurately detect steps in a free-living environment using only one ankle or waist AM.

Original languageEnglish (US)
Pages (from-to)2519-2535
Number of pages17
JournalPhysiological Measurement
Volume36
Issue number12
DOIs
StatePublished - Nov 23 2015

Keywords

  • accelerometer
  • body-worn sensors
  • movement analysis
  • sensor location
  • step detection

ASJC Scopus subject areas

  • Biophysics
  • Physiology
  • Physiology (medical)

Fingerprint Dive into the research topics of 'Step detection using multi- versus single tri-axial accelerometer-based systems'. Together they form a unique fingerprint.

  • Cite this