User Tools

Site Tools


projects:accelerometryfalldetection

Accelerometry for Fall Detection

Jelle van den Putte, Indra Schoonbrood, Kamal Mokhtar, Lars Feijen, Steven van Klinken, David Heurtaux, and Joey van Mol
Coach: Geert Langereis
November 2014 - January 2015

Martijn Tuijten, Kanyu Tang, Nader Al Zadjali, Ali Al Tubi, Steven van Klinken, Tero Turkki
Coach: Geert Langereis
February 2015 - April 2015

Jan Janssen, Fabiane Rodrigues Cavalcante, Maurice Appelhof, Kevin Russel, Marco Deurloo, Freek Rutten
Coach: Pavel Samalik
August 2015 - November 2015

Jan Janssen, Bart Kerkers, Rob Ekelmans
Coach: Nico van der Aa
November 2015 - January 2016

Jan Janssen, Bart Kerkers, Rob Ekelmans
Coach: Nico van der Aa
November 2015 - January 2016

Martin Vladimirov, Artjoms Belovs, Mustafa Omid, Michelle van Bussel, Jeffrey de Jong, Martijn van de Wijdeven
Coach: Geert Langereis, Dennis Cornuijt
February 2016 - April 2016

The result of this project is described on the Accelerometry page.

The first two groups have worked on the algorithms and hardware to measure “fall” from an acceleration signal. figure 1 shows how the experiments were done.

Fig. 1: Experiments to find acceleration patterns in fall behaviour

One of the problems with the hardware is that accelerometers are only available in SMD packages. As a result, for prototyping we are dependent on breakout boards that are normally expensive. Because we had no SMD facilities available, the prototyping method of figure 2 was used.

Fig. 2: A method to mount a 3D accelerometer in an SMD package on a prototyping board

For fall detection, normally three indicators are used:

  • Free fall detection (measured acceleration magnitude is zero)
  • No motion (measured acceleration magnitude is constant $1g$)
  • Heavy impact (measured acceleration maginute above a certain threshold)

A combination of these in a certain sequence is the easiest way to detect a fall condition.

Fig. 3: The 3D acceleration signal and the meaning of the patterns

When the project was continued in Q1/Q2 2015-2016 the focus was especially on the printed circuit board, like described on the mounting microcontrollers page.

File Version Date Description
final_report_fall_detection_2014-2015_q2.pdf 1.0 January 13, 2015 Final report first group
final_report_fall_detection_2014-2015_q3.pdf 1.0 April 13, 2015 Final report second group +mdd
expo18_2015-2016_q1_accelerometer.pdf 1.0 November, 2015 Final report third group
February, 2016 Final report fourth group
projects/accelerometryfalldetection.txt · Last modified: 2016/03/15 16:19 by geert