Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Electrical Engineering

Major Professor

Salvatore D. Morgera, Ph.D.

Committee Member

Stephen E. Saddow, Ph.D.

Committee Member

Ashwin Parthasarathy, Ph.D.

Committee Member

Mile Krajcevski, Ph.D.

Committee Member

Mark Jaroszeski, Ph.D.


Volatile Organic Compounds, COVID-19, Viral and Bacterial Pathogens, Gas sensor, SARS-CoV-2


Detection of specific target biomarkers is a critical step for electronic nose technologies in breath analysis. Electronic nose technologies have been successfully demonstrated in the air quality, meat processing, and the citrus industries, as well as cancer clinical trials. The novel design, assembly, and implementation of a test bed prototype of an electronic nose system is introduced in this Dissertation. To verify the feasibility of a test bed prototype design, a Corona Virus Disease 2019 (COVID-19) breath simulation mixture of alcohol, acetone, and carbon monoxide was presented to the test bed prototype in the laboratory. The COVID-19 breath simulation mixture was developed from a literature review and from preliminary gas chromatography ion mobility spectrometry data obtained from parallel COVID-19 breath studies based in Edinburgh and Dortmund. As a result, the test bed prototype was fit with metal oxide semiconductor (MOS) MQ-2 and MQ-135 gas sensors due to their availability, affordability, high sensitivity, high selectivity, and low recovery time properties related to their respective target gasses. The sensors were configured to show a detection range for alcohol, acetone, and carbon monoxide in part per million (ppm) level.The problem with the use of MOS based sensors in electronic nose technologies is system instability, due to a result of gradual drifts in the output responses of the MQ-2 and MQ-135 sensors. This instability directly pertains to the baseline output response of the MQ-2 and MQ-135 sensors not being accurately obtained. A main focus of this Dissertation analyzes a signal processing approach involving the manipulation of the gas sensor output response baseline using Orthogonal Signal Correction (OSC) in an attempt to decrease the negative influence of gradual drifts associated with the output response of the MQ-2 and MQ-135 sensors. The proposed signal processing approach of applying baseline manipulation with OSC was performed to eliminate and/or minimize the drift effects of the MQ-2 and MQ-135 gas sensors responding to varying concentrations of the COVID-19 breath simulated mixture of methyl alcohol, acetone, and carbon monoxide (CO). A Partial Least Squares (PLS) regression model is then employed to determine the approximate concentration of gas detected in ppm after the corrected datasheet undergoes baseline manipulation. This Dissertation will detail how the proposed signal processing approaches increase the reproducibility of the sensor output responses by enhancing the regression model stability while maintaining an accuracy suitable for breath analysis applications.