2nd International ICST Conference on Pervasive Computing Technologies for Healthcare

Research Article

Automatic Self-Calibration of Body Worn Triaxial-Accelerometers for Application in Healthcare

Download746 downloads
  • @INPROCEEDINGS{10.4108/ICST.PERVASIVEHEALTH2008.2537,
        author={Matthias Gietzelt and Klaus-Hendrik Wolf and Michael Marschollek and Reinhold Haux},
        title={Automatic Self-Calibration of Body Worn Triaxial-Accelerometers for Application in Healthcare},
        proceedings={2nd International ICST Conference on Pervasive Computing Technologies for Healthcare},
        publisher={IEEE},
        proceedings_a={PERVASIVEHEALTH},
        year={2008},
        month={7},
        keywords={Acceleration Calibration Pattern recognition Wearable sensor},
        doi={10.4108/ICST.PERVASIVEHEALTH2008.2537}
    }
    
  • Matthias Gietzelt
    Klaus-Hendrik Wolf
    Michael Marschollek
    Reinhold Haux
    Year: 2008
    Automatic Self-Calibration of Body Worn Triaxial-Accelerometers for Application in Healthcare
    PERVASIVEHEALTH
    ICST
    DOI: 10.4108/ICST.PERVASIVEHEALTH2008.2537
Matthias Gietzelt1,*, Klaus-Hendrik Wolf, Michael Marschollek, Reinhold Haux
  • 1: Institute of Medical Informatics of the Technical University Carolo-Wilhelmina, Braunschweig, Germany
*Contact email: m.gietzelt@tu-bs.de

Abstract

In healthcare acceleration sensors have various applications in monitoring movements for prevention, diagnostics, therapy and rehabilitation. Unfortunately they react sensitive against changes of environmental conditions and therefore it is necessary to calibrate them from time to time. This paper introduces a technique for automatic self-calibration of body worn triaxial-accelerometers that do neither need any user interaction nor a special calibration procedure. The data are calibrated automatically during operation of the sensor. The method does not require any knowledge about the actual orientation of the sensor and uses the activity of the subject to calibrate, because the activity is the region of interest for evaluation. The error of the automatic self-calibration will be as small as the error of common techniques for calibration that need an additional procedure, whereby an important error source will be excluded. The resulting algorithm is so small and compact, that it can even be transferred onto a sensor system and executed by its microprocessor.