Abstract

Artificial or Electronic Noses mimic the gas-detection function of biological olfactory systems. They have been used in a wide variety of applications ranging from testing of fragrances, through analysis of human breath, to the detection of banned drugs and explosives. An artificial nose system is generally composed of a sensing subsystem and a pattern recognition subsystem. More recently, Plasma Emission Spectroscopy (PES) has been proposed as an approach to building low-selectivity sensing subsystems. Due to the complexities of the underlying physical phenomena in PES, the resulting spectra are difficult to interpret. This paper presents an artificial neural network approach to plasma emission interpretation.