Wearable pendant device monitoring using new wavelet-based methods shows daily life and laboratory gaits are different
- 1.1k Downloads
Morbidity and falls are problematic for older people. Wearable devices are increasingly used to monitor daily activities. However, sensors often require rigid attachment to specific locations and shuffling or quiet standing may be confused with walking. Furthermore, it is unclear whether clinical gait assessments are correlated with how older people usually walk during daily life. Wavelet transformations of accelerometer and barometer data from a pendant device worn inside or outside clothing were used to identify walking (excluding shuffling or standing) by 51 older people (83 ± 4 years) during 25 min of ‘free-living’ activities. Accuracy was validated against annotated video. Training and testing were separated. Activities were only loosely structured including noisy data preceding pendant wearing. An electronic walkway was used for laboratory comparisons. Walking was classified (accuracy ≥97 %) with low false-positive errors (≤1.9 %, κ ≥ 0.90). Median free-living cadence was lower than laboratory-assessed cadence (101 vs. 110 steps/min, p < 0.001) but correlated (r = 0.69). Free-living step time variability was significantly higher and uncorrelated with laboratory-assessed variability unless detrended. Remote gait impairment monitoring using wearable devices is feasible providing new ways to investigate morbidity and falls risk. Laboratory-assessed gait performances are correlated with free-living walks, but likely reflect the individual’s ‘best’ performance.
KeywordsGait analysis Wavelet Remote Wearable devices Falls
We gratefully acknowledge support which made this research possible. Devices were lent from Philips Research Europe, Netherlands. M.B., S.L., and K.D. were NHMRC Fellows, Australia. Y.G. was supported by the Margarete and Walter Lichtenstein Foundation, Switzerland.
- 2.Ayrulu-Erdem B, Barshan B (2011) Leg motion classification with artificial neural networks using wavelet-based features of gyroscope signals. Sensors (Basel, Switzerland) 11(2):1721–1743. doi: 10.3390/s110201721
- 10.Delbaere K, Sherrington C, Lord SR (2013) Chapter 70—falls prevention interventions. In: Marcus R, Feldman D, Dempster DW, Luckey M, Cauley JA (eds) Osteoporosis (4th edn). Academic Press, San Diego, pp 1649–1666. doi: 10.1016/B978-0-12-415853-5.00070-4
- 16.Kaymak U, Ben-David A, Potharst R (2012) The AUK: a simple alternative to the AUC. Eng Appl Artif Intell. doi: 10.1016/j.engappai.2012.02.012
- 18.Kirtley C (2006) Clinical gait analysis. Elsevier, LondonGoogle Scholar
- 21.Mannini A et al (2011) Healthcare and accelerometry: applications for activity monitoring, recognition, and functional assessment. In: Lai DTH, Palaniswami M, Begg R (eds) Healthcare sensor networks: challenges toward practical implementation. CRC Press, New York, pp 21–46Google Scholar
- 26.Rispens SM, van Schooten KS, Pijnappels M, Daffertshofer A, Beek PJ, van Dieen JH (2015) Identification of fall risk predictors in daily life measurements: gait characteristics’ reliability and association with self-reported fall history. Neurorehabilitation Neural Repair 29(1):54–61. doi: 10.1177/1545968314532031 CrossRefPubMedGoogle Scholar
- 28.Sachdev PS, Brodaty H, Reppermund S, Kochan NA, Trollor JN, Draper B, Slavin MJ, Crawford J, Kang K, Broe GA, Mather KA, Lux O, Memory, Ageing Study T (2010) The Sydney memory and ageing study (MAS): methodology and baseline medical and neuropsychiatric characteristics of an elderly epidemiological non-demented cohort of Australians aged 70–90 years. Int Psychogeriatr 22(8):1248–1264. doi: 10.1017/S1041610210001067
- 34.Wang K, Lovell NH, Del Rosario MB, Ying L, Jingjing W, Narayanan MR, Brodie MAD, Delbaere K, Menant J, Lord SR, Redmond SJ (2014) Inertial measurements of free-living activities: assessing mobility to predict falls. In: Engineering in medicine and biology society (EMBC), 36th annual international conference of the IEEE, 26–30 Aug 2014, pp 6892–6895. doi: 10.1109/embc.2014.6945212