Introduction:

Advances in the understanding of olfaction are leading to artificial electronic noses based on an array of sensors that bind airborne molecules with only modest specificity. Detection is based on unique patterns of responses generated by the sensor array in the presence of an odor.
 

Research studies:

• A high-tech company is working on this project with funds from the Defense Advanced Projects Research Agency (DARPA)
• Stephen R. Johnson from the Pennsylvania State University, Department of Chemistry is working on an Artificial Nose project.

His research focuses on the development of routines to identify the patterns generated by analyte presentation to an array of optical sensors.
The sensor array consist of a collection of fiber optics coated with Nile Red dye immobilized in various polymer matricies. The responses from the sensors consist of the change in fluorescence with time resulting from the presentation of an analyte vapor.
Methods developed thus far include the use of Learning Vector Quantization (LVQ) neural networks and Fuzzy ARTMAP neural networks for the classification of the presented analytes. The inputs to these neural networks were a collection of descriptors derived from the sensor responses. The best subset of descriptors was chosen using a genetic algorithm (GA) with either LVQ or ARTMAP as the fitness evaluator.
 
• A team at the Tufts School of Medicine apply the same method. They simulate the olfactory sensors by employing the special fluorescent dye, Nile Red, embedded in polymers and coupled to optical fibers in a neural network complex that will identify specific vapors.

Their strategy is to use a miniaturized array of fibers containing the Nile Red
dye/polymer matrix. The dye exhibits large wavelength shifts in its strong fluorescent response to various vapors. The dye is photochemically stable and can be immobilized in polymers. Furthermore, the response time is in tens to hundreds of milliseconds.
 
• A research group at the Department of Mechanical and Materials Engineering in the University of Western Australia developed a hyper-sensitive artificial "nose" to smell the minute traces of gas leaking from explosive charges.