Wireless microelectronic nose network for gas classification

Electronic nose systems, containing array of gas sensors, are emerging platforms in numerous disciplines of life for gas classification. Repeated performance of these systems is not guaranteed because of their drift due to temperature and humidity and hence make them inefficient to distinguish gases correctly over time with stand-alone unit. Wireless sensors network is the promising solution to improve the performance of gas identification over single unit. This paper, adopts spike rank order encoding scheme due to its concentration independent feature and proposes wireless microelectronic nose network to get improved performance over single unit. In our scheme, weights are assigned to relative position of spikes in each training gas and these weights play a decisive role about decision of testing gas when received ranks from the nodes are not fully matched with any of the training gas signature. In the network, nodes send their ranks sequence to base station on gas exposure. Base station identifies testing gas as one of the training gas if its reference ranks signature is received from the nodes. If no training gas rank signature is received from any node then the node with maximum ranks matching sequence to any training gas is selected and mismatch pattern is searched in others node. On failure to find a matched sequence, decision is based on the inter-spikes weights. Experimental results validate the improved performance of this approach over single electronic nose system.