[4Pin1] インタラクティブ(2)

[4Pin1-38] Metal-Oxide-Semiconductor Gas Sensor Array Toward Artificial Olfaction

〇Yuta Saito¹, Takamune Yokoyama¹, Takahisa Tanaka², Ken Uchida² (1. Faculty of Science and Technology, Keio University, 2. Department of Electronics and Electrical Engineering, Keio University)

Keywords:Artificial Olfaction, Machine Learning application

As key components of artificial olfaction, small, stable, low-power, and highly selective gas sensors are demanded. Metal oxide semiconductor (MOS) gas sensors satisfy some of those requirements. However, sensors in this category do not have high selectivity against target molecules. On the other hand, it is considered that human olfaction is realized by a number of various low-selectivity acceptors and processing of acceptor signals in brain. In this work, we prepared a sensor array consisting of three MOS-based gas sensors with different target molecules. Sensor array responses to three kinds of Sake were analyzed using machine-learning techniques. Accuracy of Sake indefication was achieved to be greater than 90%.