ABSTRACT
This chapter introduces a hardware friendly algorithm for gas identification. It describes the sensitivity difference of any two sensors in the array as an input feature and a subset of the features is extracted by evaluating the capability of each pair of sensor to split the gases into two branches. Lack of selectivity and low repeatability of the gas sensors are the major challenges in all gas identification problems. An artificial olfactory system, referred to an electronic nose, is a multi-sensor platform used for gas classification. An artificial olfactory system, referred to an electronic nose, attempts to mimic the human olfactory system to sense and discriminate between different odors. All possible pairs of sensors are explored at every decision node of the tree to find the best split of gases data during the learning stage. The sensor pair corresponding to the maximum split value is selected for classification of test vectors.
