Searle, Graham Ellis (2002) Dynamic modelling of electronic nose systems. PhD thesis, University of Warwick.

Preview

PDF WRAP_Theses_Searle_2002.pdf - Submitted Version - Requires a PDF viewer. Download (9Mb) | Preview

Request Changes to record.

Abstract

This thesis details research into the modelling of the dynamic responses of electronic nose systems to odour inputs. Most electronic nose systems contain an array of between 4 and 32 odour sensors, each of which respond in varying degrees to a range of different gaseous stimuli. In almost all electronic nose systems in use today, the steady-state responses of the odour sensors are extracted and passed to one of a variety of pattern recognition systems. The primary aim of this thesis is to investigate the use of information contained within the dynamic portion of the sensor response for odour classification.

System identification techniques using linear time-invariant black box models are applied to both extracted steady state and full dynamic data sets collected from experiments designed to assess the ability of an electronic nose system to discriminate between the strain and growth phases of samples of cyanobacteria (blue-green algae). The results obtained are compared with those obtained elsewhere using the same data, analysed with nonlinear artificial neural networks.

A physical model for the electrochemical mechanisms resulting in the measured responses is translated into a mathematical model. This model consists of a system of coupled nonlinear ordinary differential equations. The model is analysed, and the theoretical structural identifiability of the model is investigated and established.

The parametric model is then fitted to data collected from experiments with simple (single chemical species) odours. An odour discrimination method is developed, based upon the extraction of physically significant parameters from experimental data. This technique is evaluated and compared with the previously explored black box modelling techniques.
The discrimination technique is then extended to the analysis of complex odours, again using the cyanobacteria data sets. Successful classification rates are compared with those obtained earlier in the thesis, and elsewhere with neural networks applied to steady state data.

Item Type:	Thesis (PhD)
Subjects:	Q Science > QP Physiology R Medicine > R Medicine (General)
Library of Congress Subject Headings (LCSH):	Electrochemical sensors, Olfactory sensors, Olfactory receptors
Official Date:	September 2002
Dates:	Date Event September 2002 UNSPECIFIED
Institution:	University of Warwick
Theses Department:	Department of Engineering
Thesis Type:	PhD
Publication Status:	Unpublished
Supervisor(s)/Advisor:	Gardner, J. W. (Julian W.), 1958- ; Chappell, Michael J.
Sponsors:	Engineering and Physical Sciences Research Council
Format of File:	pdf
Extent:	xviii, 214 leaves : illustrations
Language:	eng

Request changes or add full text files to a record

Repository staff actions (login required)

View Item

Downloads