Signal Processing Architectures for Chemical Sensing Microsystems

Abstract

This paper reviews several levels of signal processing and associated architectures for chemical sensing microsystems that use either arrays of optical or of physical sensors. Many chemical sensors, because of their interaction and vulnerability to the environment, have been eliminated from inclusion in sensing systems that require high precision and accuracy. This discussion evaluates parametric vs. nonparametric techniques and linear vs. nonlinear signal processing approaches for addressing chemical classification problems using imperfect sensing technologies. Hardware implementations of signal processing and biologically inspired signal processing are also reviewed. Future research into the development of more accurate chemical classification systems demands the customization of current approaches, so that underlying principles of chemical sensors and associated interfering influences do not overburden the computational space, thereby allowing higher accuracy rates.