Vibroakustisches Monitoring in Smart Homes

Project Type: Master/Diploma Thesis
Student: Klaus Dobbler

Starting from the application of vibroacoustic technology in the field of Ambient Assisted Living,
a concept for a vibroacoustic monitoring-system is introduced, which is able to realize several
potential application scenarios in smart homes. Within the scope of this work, the two main
functionalities of the system, namely classification and localization of seismic events are
developed for the aim of human gait analysis. The classification stage has the task to extract
footsteps out of the vibrational signals whose positions are determined in the subsequent
localization stage. The extraction is realized through the detection of the first phase of a gait
cycle by the computation of MFCCs. In order to localize the extracted signals, two tri-axial
sensor stations are used, which are both determining the angle of arrival of the propagating
surface wave. Appropriate floor conditions and enough energy in the excitation assumed, it is
possible to determine the position of the footstep by triangulation. The developed algorithms
are implemented in a real-time system which is demonstrated by two possible applications as
Tangible Acoustic Interface and Gait Analysis-Instrument.