Application an Odour Recognition using Nearest Neighbour Algorithm

Abstract

This project was undertaken with the aim of developing control techniques and data processing methods to allow a relatively simple electronic nose developed for robotics applications to classify a wide range of odours. The electronic nose used tin oxide gas sensors as the main sensing elements. By modulating the sensor temperature inside the sensors%2C additional information was gathered which helps to identify unknown odours. A simple K-nearest neighbor algorithm was implemented as the basis of a pattern recognition system for recognizing odours presented to the system. Further%2C we considered ways of compressing the stored data and techniques for finding the best match among the trained data using Principal Component Analysis.
Abstract in Bahasa Indonesia :

Dalam tulisan ini dibahas sebuah proyek sistem pengenalan aroma menggunakan sensor hidung elektronik. Berbagai aroma diklasifikasi menggunakan sensor hidung elektronik jenis sensor gas oksida timah. Empat buah sensor digunakan untuk mengakusisi aroma yang selanjutnya diklasifikasi dengan algoritma nearest neighbour. Sebagai metoda ekstraksi fitur digunakan principal component analysis %28PCA%29 yang memungkinkan representasi data lebih kompak. electronic+nose%2C+tin+oxide+gas+sensor%2C+K-nearest+neighbor%2C+Principle+Component+Analysis.