Sensitivity, selectivity and stability are decisive properties of sensors. In chemical gas sensors odor recognition can be severely compromised by poor signal stability, particularly in real life applications where the sensors are exposed to unpredictable sequences of odors under changing external conditions. Although olfactory receptor neurons in the nose face similar stimulus sequences under likewise changing conditions, odor recognition is very stable and odorants can be reliably identified independently from past odor perception. We postulate that appropriate pre-processing of the output signals of chemical sensors substantially contributes to the stability of odor recognition, in spite of marked sensor instabilities. To investigate this hypothesis, we use an adaptive, unsupervised neural network inspired by the glomerular input circuitry of the olfactory bulb. Essentially the model reduces the effect of the sensors' instabilities by utilizing them via an adaptive multicompartment feed-forward inhibition. We collected and analyzed responses of a 4 × 4 gas sensor array to a number of volatile compounds applied over a period of 18 months, whereby every sensor was sampled episodically. The network conferred excellent stability to the compounds' identification and was clearly superior over standard classifiers, even when one of the sensors exhibited random fluctuations or stopped working at all. ; peerReviewed


    Access

    Download


    Export, share and cite



    Title :

    Stable odor recognition by a neuro-adaptive electronic nose.



    Publication date :

    2015-01-01


    Remarks:

    26043043



    Type of media :

    Article (Journal)


    Type of material :

    Electronic Resource


    Language :

    English



    Classification :

    DDC:    629




    Odor Characterization of Automotive Cabins Using an Electronic Nose

    D'Amico, A. / Di Natale, C. / Malvicino, C. et al. | SAE Technical Papers | 2001


    Odor characterization of automotive cabins using an electronic nose

    Amico, A. / Natale, C. / Malvicino, C. et al. | Tema Archive | 2001


    Odor characterization of automotive cabins using an electronic nose

    D Amico,A. / Di Natale,C. / Malvicino,C. et al. | Automotive engineering | 2001


    2001-01-3276 Odor Characterization of Automotive Cabins Using an Electronic Nose

    D Amico, A. / Di Natale, C. / Malvicino, C. et al. | British Library Conference Proceedings | 2001


    Odor source identification by grounding linguistic descriptions in an artificial nose [4385-38]

    Loutfi, A. / Coradeschi, S. / Duckett, T. et al. | British Library Conference Proceedings | 2001