Brain inspired isfet array: edge-computing for point-of-care diagnostics
File(s)
Author(s)
Tripathi, Prateek
Type
Thesis
Abstract
Recent years have witnessed growth in the development of point-of-care for real-time medical diagnostics. The COVID-19 pandemic highlighted the need for technologies that can provide rapid and accurate diagnosis of infectious diseases without requiring specialised labs. While lateral flow tests supported mass testing during the pandemic, they suffer from low accuracy and do not allow multiplexing of several diseases, which becomes critical as any pandemic progresses. While most techniques rely on optical methods, electrochemical sensing enables miniaturisation, scalability and robustness. We qualify the integration of electrochemical sensing with novel AI algorithms as ‘sensor learning’, leveraging on-chip methodologies to automatically calibrate the sensors and extract accurate diagnostic information. This work presents a spatial correlation between the non-ideal effects to facilitate inter-pixel processing using neuromorphic ISFETs. We begin with neuromorphic ISFET arrays using spike domain encoding and spatial device compensation. This is followed by a completely autonomous cluster topology of neuron-based pixels based on a multiple-channel Integrate and Fire architecture for temporal integration and spatial averaging. The designs have been implemented in TSMC 180nm and TSMC 65nm CMOS technology. We have also created a novel winner-take-all (WTA) architecture for background inhibition in ISFET neurons that can form Clustered WTA and Distributed WTA architecture while at the same time performing drift compensation using temporal and spatial averaging. We have also implemented an in-pixel detection mechanism that uses the calcium conductance channel to allow the circuit to adapt the spike frequency and help us reduce the activity from the neuron before and after the amplification event. Further, we have established a state-of-the-art with our first models for Lab-on-chip platforms that have been trained to identify infectious diseases and cancer biomarkers using tinyML. In addition, this thesis also presents a framework that can accelerate our testing response to future pandemics using AI at the edge.
Version
Open Access
Date Issued
2024-11-16
Date Awarded
01/03/2025
License URL
Advisor
Georgiou, Pantelis
Moser, Nicolas
Sponsor
Imperial College London
Publisher Department
Department of Electrical and Electronic Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)