Pattern Classification Techniques for Lung Cancer Diagnosis by an Electronic Nose - Computational Intelligence in Healthcare: Advanced Methodologies - page 387

MOSFET sensors. Metal-Oxide-Silicon Field-Effect-Transistor (MOSFET)
odor sensing devices are based on the principle that VOCs in contact with a
catalytic metal can produce a reaction in the metal that can diffuse through
the gate of a MOSFET to change the electrical properties of the device. MOS-
FET sensors have been investigated for numerous applications, but, to date, few
have been used in commercial electronic nose systems because of a dearth of
sensor variants. The advantage of MOSFETs is that they can be made with IC
fabrication processes; the disadvantage is that the catalyzed reaction products
(e.g., hydrogen) must penetrate the catalytic metal layer. Moreover they suffer
of baseline drift and their sensitivity is lower (order of ppm) then other odor
sensors.

Optical-fiber sensors. In optical-fiber sensors, the active material contains
chemically active fluorescent dyes immobilized in an organic polymer matrix,
which polarity is altered as VOCs interact with it. The reaction is a shift of
the fluorescent emission spectrum. The advantage of these sensors is their low
cost, but the complexity of the instrumentation control system increases the
production cost. They have a limited lifetime due to photobleaching, but the
exploitation of an array of fiber can provide a wide ranging sensitivity.

5 Signal Conditioning and Preprocessing

The second fundamental module of an odor-sensing instrument consists in the
signal conditioning and preprocessing phase, that consists in converting the ac-
quired sensor response in the most appropriate form for the further multivariate
pattern analysis module, minimizing the effects of noise and of all those effects
(e.g. baseline drift, humidity) that negatively affect the quality of the signal.
This task can be distinguished in three main steps:

1. Conversion of the odor sensors response into an electrical signal by means
of interface circuits
2. Enhancement of signal information through analog conditioning (e.g. filter-
ing), as well as sampling and analog to digital conversion
3. Digital processing of the sampled signal to make it suitable for further mul-
tivariate pattern analysis

5.1 Interface Circuits

The first stage of any electronic instrumentation is the conversion of sensors
response (e.g., resistance change) into an electrical signal; this operation is made
by means of interface circuits and varies according to the specific technology on
which sensors are based. In chemoresistors sensors, for example, interface circuits
are relatively simple since they only involve measuring resistance changes; on the
contrary, electronics instrumentation for piezoelectric sensors are more complex,
as they involve AC signals for high frequency. In [4] an extensive overview of
interface circuits theory is provided.