Neural-Network Fault Diagnosis for Electrode Structures in Bio-fluidic Microsystems

Lab-on-chip devices are of great interest for analysis in fields including biochemistry, biomedical engineering and bioelectronics. Within these systems, highlevels of reliability and robustness are crucial and normally complemented by requirements for extremely low probabilities of false positives or negatives being generated. Optimizing the design of these devices and investigating new methods for validating functionality and integrity of the readings are therefore required. This paper proposes a new fault diagnosis approach using Artificial Neural Network(ANN) for detecting degradation in electrodes that interface to fluidic or biological systems and form the basis of numerous actuation and sensing mechanisms in the biofluidics area. In this approach, the ANN is constructed and trained with a subset of experimental impedance data which was extracted at different degradation levels. New sets of data are used to test the network and the results show that the ANN has the ability to provide an early warning for degradation within the electrode structure.