Fuzzy clustering neural networks for real-time odor recognition system

An electronic tongue as a classifier tool for assessing perfume olfactory family and storage time-period

The identification of more than three perfumes is difficult and no analytical tool can completely replace the human olfactory system for fragrance classification. Indeed, no analytical system can mimic the human fragrance perception, being the recognition of perfume aroma patterns by conventional or sensor-based analytical tools a challenging task. For the perfume sector, the possibility of applying fast, cost-effective and green analytical devices for perfume analysis would represent a huge economic revenue. Since the perfume aroma pattern will depend on the composition of the liquid phase and on the diffusion properties of their volatile components, this work aimed to apply a potentiometric electronic tongue, comprising non-specific cross-sensitive lipid polymeric membranes, combined with chemometric techniques, as a novel perfume classifier. The multisensors device allowed establishing perfumes' unique fingerprints, which were successfully used to discriminate men from women perfumes, to identify the perfume aroma family (Citric-Aromatic, Floral, Floral-Fruity, Floral-Oriental, Floral-Woody, Woody-Oriental and Woody-Spicy) and, assessing the perfume storage time-period (≤ 9 months; 9-24 months; and, ≥ 24 months). The established linear discriminant models were based on single-run potentiometric profiles gathered by sub-sets of sensors selected using the simulated annealing algorithm, which enabled achieving correct classification rates of 93-100% (for leave-one-out cross-validation procedure). The satisfactory performance of the electronic tongue demonstrates the versatility of the proposed approach as a practical perfume preliminary classifier sensor device, which industrial application may be foreseen in a near future, contributing to a green-sustained economic growth of the perfume industry.

Hazardous Odor Recognition by CMAC Based Neural Networks

Electronic noses are being developed as systems for the automated detection and classification of odors, vapors, and gases. Artificial neural networks (ANNs) have been used to analyze complex data and to recognize patterns, and have shown promising results in recognition of volatile compounds and odors in electronic nose applications. When an ANN is combined with a sensor array, the number of detectable chemicals is generally greater than the number of unique sensor types. The odor sensing system should be extended to new areas since its standard style where the output pattern from multiple sensors with partially overlapped specificity is recognized by a neural network or multivariate analysis. This paper describes the design, implementation and performance evaluations of the application developed for hazardous odor recognition using Cerebellar Model Articulation Controller (CMAC) based neural networks.