Dissertation/Thesis Abstract

Real-Time Detection and Tracking of Vital Signs with an Ambulatory Subject Using Millimeter-Wave Interferometry
by Mikhelson, Ilya V., Ph.D., Northwestern University, 2013, 146; 3595683
Abstract (Summary)

Finding a subject's heart rate from a distance without any contact is a difficult and very practical problem. This kind of technology would allow more comfortable patient monitoring in hospitals or in home settings. It would also allow another level of security screening, as a person's heart rate increases in stressful situations, such as when lying or hiding malicious intent. In addition, the fact that the heart rate is obtained remotely means that the subject would not have to know he/she is being monitored at all, adding to the efficacy of the measurement.

Using millimeter-wave interferometry, a signal can be obtained that contains composite chest wall motion made up of component motions due to cardiac activity, respiration, and interference. To be of use, these components have to be separated from each other by signal processing. To do this, the quadrature and in-phase components of the received signal are analyzed to get a displacement waveform. After that, processing can be done on that waveform in either the time or frequency domains to find the individual heartbeats. The first method is to find the power spectrum of the displacement waveform and to look for peaks corresponding to heartbeats and respiration. Another approach is to examine the signal in the time domain using wavelets for multiresolution analysis. One more method involves studying the statistics of the wavelet-processed signal. The final method uses a heartbeat model along with probabilistic processing to find heartbeats.

For any of the above methods to work, the millimeter-wave sensor has to be accurately pointed at the subject's chest. However, even small subject motions can render the rest of the gathered data useless as the antenna may have lost its aim. To combat this, a color and a depth camera are used with a servo-pan/tilt base. My program finds a face in the image and subsequently tracks that face through upcoming frames. The pan/tilt base adjusts the aim of the antenna depending on the subject's position. This makes the entire system self-aiming and also allows the subject to move to a new location and to have data acquisition continue.

Indexing (document details)
Advisor: Sahakian, Alan V.
Commitee: Katsaggelos, Aggelos K., Swiryn, Steven, Taflove, Allen
School: Northwestern University
Department: Electrical Engineering
School Location: United States -- Illinois
Source: DAI-B 75/01(E), Dissertation Abstracts International
Subjects: Biomedical engineering, Electrical engineering, Remote sensing
Keywords: Ambulatory monitoring, Heart rate, Heartbeat detection, Millimeter-wave, Remote sensing
Publication Number: 3595683
ISBN: 978-1-303-41564-7
Copyright © 2021 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy