A study on environmental recognitions by odor sensors considered on the concept of human senses

Abstract

Many kinds of chemical sensors have been developed to detect various air-pollutants. Generally sensor systems are composed of plural chemical sensors and a computer, and the systems can derive kinds of pollutants, densities of the pollutants, and various risks for environments and humans. Regarding the research of sensor systems, there are literatures about a sensor agent, which regards a sensor as an agent, and a utility mobile robot which attaches odor sensors. In these systems, it is necessary that observed results output by the sensors are exactly analyzed with information processing; especially, it is important to recognize the occurrence of events such as outbreaks of odor, however, this kind of recognitions are still not enough and it does not mean that the systems can recognize all the environmental events. In this research, odor sensors are treated and generally odor sensors are designed modeling functions of human nose. Hence, as well as odor sensors, we aim improving the accuracy of environmental recognitions using an odor sensor system attaching an information processing function according to human senses. In addition, we show the usefulness of the proposed system comparing with a conventional technique.