Toward Unsupervised Smart Chemical Sensor Arrays

This chapter discusses the main aspects underlying the application of blind source separation (BSS) methods to the problem of quantitative analysis using smart sensor arrays (SSAs). It focuses on sensor arrays composed of potentiometric electrodes. The chapter provides an overview of current results concerning the application of BSS methods to quantitative analysis using potentiometric sensor arrays. It discusses the some important practical aspects related to the use of BSS methods for performing chemical analyses. Potentiometric sensors are devices that measure the variation in electrical potential induced by variation in the concentration of a given chemical species. Most chemical SSAs are based on supervised signal processing methods, which require calibrating samples to adjust the parameters of the adopted signal processing method. The application of supervised methods in both quantitative and qualitative analysis has been proving quite successful in tasks such as odor and taste automatic recognition systems.