Application of data-based dynamic modeling techniques to temperature-programmed microhotplate gas sensors

Conference · Tue Dec 31 04:00:00 EST 1996

OSTI ID:441253

Kunt, T A; McAvoy, T J; Ratton, L ^[1]

---

1. Univ. of Maryland, College Park, MD (United States); and others

The olfactory systems of animals and humans are exceptional at detecting and classifying odors even when exposed to only trace amounts of an odorant. Recordings from single olfactory receptors have shown that the receptors are broadly tuned, and that they respond to a number of different odorants. Thus, nature seems to use numerous, non-specific sensors and a good signal processing/pattern recognition algorithm to achieve excellent odor detection and classification. Many researchers have already coupled multiple sensors with available artificial intelligence tools such as artificial neural networks (ANNs) to detect and classify gas components in mixtures. In this paper, the authors demonstrate the application of a somewhat different approach for development of an artificial nose. They examine the information content available from one conductometric microsensor when it is operated in a variety of different modes. Mode variation is accomplished by changing the sensor`s temperature to enhance its detection/classification capabilities. Data-based dynamic modeling techniques are used for attaining predictive models of sensor response.

• OSTI ID: 441253
• Report Number(s): CONF-960782--
• Country of Publication: United States
• Language: English

Related Subjects

44 INSTRUMENTATION
INCLUDING NUCLEAR AND PARTICLE DETECTORS
DATA ANALYSIS
DESIGN
GASES
MEASURING INSTRUMENTS
MONITORING
NOSE
PERFORMANCE