Title :
Feature-level signal processing for odor sensor arrays
Author :
Roppel, T. ; Dunman, K. ; Padgett, M. ; Wilson, D. ; Lindblad, T.
Author_Institution :
Dept. of Electr. Eng., Auburn Univ., AL, USA
Abstract :
A recurrent back-propagation neural algorithm is trained to classify nine odors. The algorithm is capable of correctly identifying the odors regardless of the time sequence of presentation. The classification is performed in near-real time and is based upon the transient response of an array of 15 tin-oxide gas sensors
Keywords :
array signal processing; backpropagation; chemical variables measurement; electric sensing devices; recurrent neural nets; transient response; SnO2; back-propagation; feature-level signal processing; near-real time; odor sensor arrays; odors classification; recurrent neural network algorithm; tin-oxide gas sensors; transient response; Array signal processing; Clustering algorithms; Gas detectors; Neural networks; Petroleum; Sensor arrays; Signal processing algorithms; Software algorithms; Steady-state; Transient response;
Conference_Titel :
Industrial Electronics, Control and Instrumentation, 1997. IECON 97. 23rd International Conference on
Conference_Location :
New Orleans, LA
Print_ISBN :
0-7803-3932-0
DOI :
10.1109/IECON.1997.671050
