Electronic nose for odor identification

Abstract

Electronic Nose is developed to function like a real nose in daily lives where its basic principles come from the human olfactory (sensing) system. It is used as automated detection and classification of odours, vapours and gases through several complex technologies. The objective of this project is to identify two categories of smell which consists of seven fruits and seven spices.

The whole system basically can be categorized into four main components, which are sensing system, data acquisition, classification and user interface. The sensing system is constructed based on an array of seven Figaro gas sensors. With the aid of LabVIEW, data acquisition program is developed, in capturing and collecting results from the signals generated by the sensors. The Probabilistic Neural Network has been adopted for the classification of various odours and a graphical user interface in displaying the identification of the odour. Both the classification process and the graphical user interface are developed by MATLAB.

There are great potential in developing Electronic nose for quality control in food and beverage industry, toxic gas detection and environment monitoring. The overall accuracy of the system is about 65 to 70 percents.