Comparison between MLP and RBF network in improving CHEMFET sensor selectivity
This work presents the classification of potassium ion concentration in the presence of interfering ammonium ions from Chemical Field-Effect Transistor (CHEMFET) sensors involving neural network post-processing stage. Data collection for the purpose of supervised learning training data is obtained from sample solutions prepared by keeping the main ion concentration constant while the activity of the interfering ions based on the fixed interference method. The measurement setup includes a readout interface circuit that ensures constant-current constant-voltage across the drain-source for isothermal point operation. The training algorithm is back-propagation with generalized delta rule on a multilayer feed-forward network. Activation function based on the MOSFET drain current equation in the linear region is attempted in the hidden layer. Using function fitting approach, the network aims to find the potassium ion concentration despite the presence of interfering ion, without having to estimate device and chemically related parameters that would otherwise require further experiments.