Information-theoretic optimization of chemical sensors

A gas-sensor optimization scheme for odor discrimination is introduced in this paper. We formulate the odor class separability in terms of a fundamental tool in information theory, namely the Kullback–Leibler distance (KL-distance), which gives a quantitative measure of the mutual difference between two probability distributions. We argue that maximizing this measure over a controllable operating parameter of a sensing element promotes robust odor discrimination. We demonstrate on a sample dataset that tuning the operating temperature of a metal oxide sensor based on the suggested criterion not only yields a substantial improvement in classification performance but also informs about those operating temperatures that cause a total confusion in the odor discrimination.