The how and why of electronic noses
IEEE Spectrum
Adaptive Blind Signal and Image Processing: Learning Algorithms and Applications
Adaptive Blind Signal and Image Processing: Learning Algorithms and Applications
Fast and robust fixed-point algorithms for independent component analysis
IEEE Transactions on Neural Networks
| Hi-index | 0.00 |
This paper addresses the separation of ion activities in ion-selective field transistor (ISFETs) arrays. These solid-state electrochemical sensors are designed to be sensitive to one kind of ion. However, their response is a non-linear function based on a mix of several ion activities found in the solution. Hence, blind source separation (BSS) techniques can be applied to recover the original main ion activity from the mixed response. Although the non-linear recovery problem could be considered difficult at first glance, our work shows how the use of prior knowledge in BSS based on a local ISFET model allows the transformation of the original problem into a linear BSS model through a simple pre-calibration step. In this context, the overall processing system is capable of recovering not only the original main ion activity but also the interfering ion activity, which has classically been treated as noise. Several preliminary experiments based on a linear ICA algorithm look promising. This work is in progress, as a part of the SEWING EU project (contract IST-2000- 28084).