Extraction of Information of Driving-State from a Gas Sensor Response by Genetic Algorithm

Abstract

In this paper, we tried to extract information of driving-state of gas generating sources which are located in an indoor space(e.g. oil stoves, gas ranges and so on) for recognition of human activities and environmental changes. It is necessary for extracting information of driving-state of gas generating sources that not only a signal of gas concentration in an indoor space but also signals of components in gas concentration which depend on each gas generating source are estimated. In this paper, we proposed methods to find appropriate combination of component signals of gas generating sources using Genetic Algorithm(GA). Then, uniqueness and robustness theorem in estimation by our methods are shown in computing simulations. And we shown it is possible to extract information of driving-state of gas generating sources from a gas sensor response in an actual experiment.