E-Nose Pattern Recognition and Drift Compensation Methods

Abstract

With new developments and upcoming technologies, new sensing techniques are becoming available. Unfortunately, none of these techniques provides output interpreted the way human perception works. An inability to improve the effectiveness of these technologies limits their use in dedicated applications and increases their complexity. The growing adoption of this technology makes it critical to create a system capable of handling e-nose challenging issues such as noise, drift, imbalanced data, dynamic environment, and high uncertainties. Without appropriate pattern recognition methods that allow inferences to be derived based on patterns observed within these data sets, it will not be possible to improve the performance of current e-nose systems. In this chapter, e-nose drift issue is introduced and the available drift counteraction methods is discussed.