Design of an integrated neuromodulable current-mode neuron architecture implemented with Pragmatic's flexible 600nm process

Abstract

Flexible integrated circuits are an emerging technology with a bright future. Using thin film transistors brings new perspectives in various fields thanks to an adaptive form factor and proposes an alternative that is less expensive and more environmentally friendly than silicon in several applications. In this project, the investigation will be about one of them, neuromorphic electronics. Neuromorphic electronics is a wide field where circuits are designed to reproduce behaviors from the human brain. It brings many advantages, like energy efficiency, parallel processing and robustness. The way they work takes advantage of the exponential behavior of transistors in the subthreshold region. The purpose of this project will be to design a neuromodulable current-mode neuron suitable to be implemented in a flexible integrated circuit. In the first instance, the focus will be on learning to know better the components in a flexible IC process, including simulations of useful quantities of the flexible transistors provided by Pragmatic Semiconductors. In the second time, the design of the current-mode neuron will begin following an existing general architecture with adaptation relative to the limitations of flexible technology, such as the design of the Howland current source to replicate the signals in the branches of the circuit. Through the realization of the neuron, this work addresses many limitations of the current technologies used for flexible electronics, leading to choices and different approaches to overcome them.