A Chronic Diseases Detection Integrated Interface With Anti-Drift Normalization and On-Chip Classification Schemes

An on-chip classifying biochemical multi-sensor integrated interface is presented for diseases diagnosis applications, including diabetes and liver diseases. Against critical problems of sensor drift over time and long settling time from analyte concentration variations, two proposed circuit-level schemes of anti-drift normalization and settling-period detection are designed by utilizing two different detectors of min-max and level-crossing. For disease classification without computation burden in edge-computing devices, an on-chip analog neural network circuit is included together, and the whole readout integrated circuit (ROIC) is fabricated in a 0.18-<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mu $ </tex-math></inline-formula>m CMOS process. A system-level prototype with an in-house organic electrochemical transistor (OECT) sensor array was implemented and experimentally verified to provide biochemical sensing of four electrolyte ions and glucose. It achieved on-chip classification model F1 scores of 0.79 and 0.72 on diabetes and liver disease respectively.