DocumentCode
1065986
Title
Odor Recorder Capable of Wide Dynamic Recordable Range Based on Higher Order Sensing and Signal Extraction Technique for Small Signal
Author
Somboon, Pakpum ; Wyszynski, Bartosz ; Nakamoto, Takamichi
Author_Institution
Grad. Sch. of Sci. & Eng., Tokyo Inst. of Technol., Tokyo
Volume
9
Issue
2
fYear
2009
Firstpage
93
Lastpage
102
Abstract
The odor recording method for extending dynamic recordable range was proposed. In this method, the enriched information obtained from higher order sensing based on preconcentrator with variable temperature was utilized so that the contribution of component with small ratio to sensor response could be separately extracted from that of the main component. Then, the signal extraction technique for small signal was developed using Savitzky-Golay filter for qualitative and quantitative detection of component with small ratio, and then this extracted information was used in recording process. It was found that the target odors including component with small ratio down to 1% were successfully recorded using this proposed method. This method is useful to extend the dynamic recordable range especially in the case that the component with high contribution in impression has low contribution to sensor responses.
Keywords
chemical variables measurement; electronic noses; feature extraction; filters; signal processing; smoothing circuits; Savitzky-Golay filter; higher order sensing; odor recorder; preconcentrator; qualitative detection; quantitative detection; signal extraction technique; wide dynamic recordable range; Data mining; Displays; Dynamic range; Information filtering; Information filters; Olfactory; Prototypes; Sensor phenomena and characterization; Signal processing; Temperature sensors; Detection of low-concentration component; Savitzky– Golay filter; higher order sensing; odor recorder; preconcentrator;
fLanguage
English
Journal_Title
Sensors Journal, IEEE
Publisher
ieee
ISSN
1530-437X
Type
jour
DOI
10.1109/JSEN.2008.2011072
Filename
4749394
Link To Document