Embedded Electronic Nose and Supporting Software Tool for its Parameter Optimization

This paper details the work done towards a low cost, small size, portable embedded electronic nose (e-nose) and its application for analysis of different VOC mixtures. The sensor array is composed of commercially available metal oxide semiconductor sensors by Figaro. The Embedded E-nose consists of an ADuC831 and has an RS 232 interface to Desktop PC for higher level data collection and NN training. An E-nose Software Package (ESP) Integrated Development Environment (IDE) is implemented for the design and development of Embedded E-nose. The IDE facilitates the parameter optimization of data acquisition, signal pre-processing, feature extraction and Multilayer Perceptron Neural Network (MLP NN) training. The optimized parameters are then downloaded onto Embedded E-nose by means of docking the device to the ESP tool via PC serial port. The Embedded E-nose is used for realization of both qualitative as well as quantitative analysis and their results are discussed.