Intelligent Wireless E-Nose for Power Savvy Distributed Chemical Sensing

In this work, we present the preliminary results for a wireless electronic nose platform embedding local sensor fusion component for the analysis of gas mixture in the framework of indoor pollution monitoring. This approach allow for significant reduction of power consumption by exploiting sensor censorship algorithms i.e. avoiding the transmission of uninformative contents to data sink. At the same time local situation assessment capabilities will allow the mote to perform adaptive, local reaction strategies e.g. duty cycle modifications, actuation etc. Performance are encouraging both on discrimination and estimation problem, we believe that a significant performance enhancement can be obtained by using a tapped delay neural network architecture.