Sensor Registration and Vapor Classification with an Artificial Nose

Abstract

The artificial nose developed by Dr. David Walt et al. is constructed by attaching hundreds of microsphere sensors to a piece of etched imaging fiber. When multiple types of sensors are attached to the same fiber they self-organize themselves, thus the types of the sensors at each location are unknown. Vapor pulses of highly concentrated analytes cause each sensor type to generate a unique fluorescence pattern over time. Based on sensor fluorescence, we present methods for identifying the type of each sensor on a piece of imaging fiber. Then we show how these sensors can be used to classify unknown vapors.