On the feasibility to apply a neural network processor for analyzing a gas response of a multisensor microarray

We describe an effort to implement a hardware neural network processor to carry out pattern recognition of signals generated by a multisensor microarray of electronic-nose type. The multisensor microarray is designed on a SnO2 thin film segmented by co-planar electrodes according to KAMINA (KArlsruhe Micro NAse) electronic-nose architecture. The response of this microarray to reducing gases mixed with synthetic air is processed by principal component analysis technique realized in conventional personal computer and hardware neural microprocessor NeuroMatrix NM6403. It is shown that the neural network processor is able to perform successfully the gas-recognition algorithms at a real time scale. The results open a way to fully mimicking a biology-inspired approach to analyze gas mixtures by hybrid chips consisting of a sensor array and a processing hardware based on neural networks.