The aim of this research project is to automatically analyse human behaviour in indoor environments using vibration sensors attached to the floor. Vibration sensors are used in an absolute passive manner so the monitored persons do not have to wear any sensors. Furthermore this technology preserves the private sphere of monitored persons. In order to extract geometric gait features like motion trajectories out of the vibrational signals, the main part of this work focuses on methods for localization of seismic sources on the two-dimensional floor surface. Starting from a conventional TDOA (Time Difference of Arrival) based technique using uniaxial acceleration sensors we show how to minimize installation cost by reducing the number of sensor entities via tri-axial sensor technology. We describe the main challenges of wave propagation in solids, namely dispersion and multipath propagation and discuss their implications on robustness and accuracy of localization results. To conclude our work we introduce potential application scenarios.
[1]
Jung Wook Park,et al.
Estimation of Indoor Physical Activity Level Based on Footstep Vibration Signal Measured by MEMS Accelerometer in Smart Home Environments
,
2009,
MELT.
[2]
Marco Fabiani.
Development of a tangible human-machine interface exploiting in-solid vibrational signals acquired by multiple sensors
,
2006
.
[3]
J. Chanussot,et al.
Dispersion estimation from linear array data in the time-frequency plane
,
2005,
IEEE Transactions on Signal Processing.
[4]
M. Marschollek,et al.
Sensor-based Fall Risk Assessment – an Expert ‘to go’
,
2011,
Methods of Information in Medicine.
[5]
J. Sabatier,et al.
Vibration and sound signatures of human footsteps in buildings.
,
2006,
The Journal of the Acoustical Society of America.