FPGA-based Odor Classification System Using Principal Component Analysis

Odor classification has been increasingly applied in various applications. The types of odors can be identified by analyzing the output pattern obtained from the odor sensors with different sensitivities to the targeted odor. Therefore, the use of reconfigurable hardware offers advantage for the flexibility to work with a number of sensors with various characteristics. In this study, the FPGA was employed in the development of an odor classification system using principal component analysis (PCA). The soft processor system was developed so that the odor classification based on PCA can be performed standalone. A number of experiments in the closed system were conducted to evaluate the validity of the proposed odor classification system and its capability to identify the kinds of odors. The effectiveness to classify the odor samples was confirmed.