Thesis info

Portable Device for Algorithmic Analysis and Feedback according to the Stress Level of Humans
Master Thesis

Philipp Emmanuel B¨nhof o ETH Zurich, Switzerland pboenhof@student.ethz.ch

June 14, 2009 - December 14, 2009

Supervised by: Florian Negele Prof. Dr. J¨rg Gutknecht u

Abstract A high stress level over a long time affects the health and causes a decrement of the quality of liveas stress symptoms are often not noticed or neglected. Therefore it would be important to feed back human stress level parameters, e.g. the heart rate variability (HRV), directly to the stressed person. The aim of this Master Thesis is to present a portable device measuring the HRV and give the feedback directly to the person wearing this device. Immediate knowledge about the stress level allows tolocate momentary stress sources and to cope with it. Being able to analyse stress directly from the HRV allows also to relate it to other measurements taken close to that point in time which make this feedback useful in human medicine e.g. to prevent heart attacks. A first programme code for such a device was already available at the beginning of the thesis. It then was improved and madecompatible to show the heart beats from the Electrocardiogram (ECG). To achieve this the beat interval was cleaned up and then interpolated. Moreover the frequency energy was calculated with wavelet frequency analysis. The gathered values of the energies in the LF and HF HRV frequency bands were then visualised with the HrvStressMon application. Also an ECG recorder and according software for visualisingand analysing the ECG signal were developed. Interaction with the ECG is now robust as well as the detection of the beats is now initialised properly. With the moving median filter for cleaning up the detected beat intervals and the generic wavelet decomposition tree it is now rather easy to do online processing of the ECG signal to retrieve information about the HRV frequency spectrum.Acknowledgements I would like to thank the following people for their support and help with my master thesis. • Prof. Dr. J¨rg Gutknecht u – providing the opportunity of such an interdisciplinary thesis • Florian Negele – discussing problems – supervising the thesis • Dr. Daniel Keller – starting the project – basic know how – connection with Barbora Blaha, Alexey Morozov and Dr.med. Markus Fritzsche •Dr.med. Markus Fritzsche – collaboration – providing location and effort for recording ECG data and counting leukocytes • Larisa Fritzsche – verification of similarity of automatic and manual ectopic beat removal • Alexey Morozov – code for viewing ECG data – code for detecting ECG beats – code to perform wave-let analysis – explanations how wave-let analysis works • Barbora Blaha – prototype code forcommunicating with the ECG device – coefficients for the wave-let Analysis – prototype code for working with wave-let analysis

Contents
1 Introduction 2 Background 2.1 Glossary . . . . . 2.2 Stress detection . 2.3 Bio-Feedback . . 2.4 Existing solutions 5 7 7 7 8 8 9 9 9 9 15 15 15 19 20 21

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

.. . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

. . . .

3 Proposed Solution 3.1 ECG signal acquisition . . . . . . . . . . . . . . . . 3.2 Heart beat detection . . . . . . . . . . . . . . . . . 3.3 Beat interval cleanup . . . . . . . . . . . . . . . . . 3.4Beat interval interpolation . . . . . . . . . . . . . . 3.5 HRV energy calculation . . . . . . . . . . . . . . . 3.5.1 Frequency analysis methods . . . . . . . . . 3.5.2 Used method to calculate the required HRV 3.6 Feedback of gathered parameters . . . . . . . . . . 4 Contribution

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . energies . . . . .

. . . . . . . .

. . . . . . ....