The implementation of a fuzzy neural network with an array of tin oxide based gas sensors for both quantitative and qualitative gas sensing is demonstrated. The architecture of the system is presented with some references to the general theory of fuzzy sets and fuzzy calculus. Experimental results are presented in the case of gas identification between CO, ethanol and methane and in the case of CO detection in different levels of relative humidity. Finally the effect of network parameters to the functionality of the system is discussed, especially in the case of functions evaluating the fuzzy AND and OR operations.
(EN)
