BAYOEN: BAYesian inference with organic electrochemical transistors for neuromorphic sensors
running time: 01.10.2021 - 31.03.2025
funding amount: 193.575,60€
funded by: BMBF
Summary:
The aim of the joint project "BAYOEN" is to use innovative organic components to efficiently implement Bayes' law in hardware to demonstrate the real-time and on-site classification of biophysiological data (e.g. ECG). The special properties of organic components are used to develop so-called dynamic Müller-C elements, which can work with fewer components and less power than comparable Si-based circuits.
As a first step, the technology of organic electrochemical transistors with an adjustable memory function is to be developed in BAYOEN. In particular, the focus is on the uniformity and reproducibility of the transistors. At the same time, the French partner (Universite Paris Saclay, UPS) is developing a component model and designing both dynamic and static Müller C elements. Furthermore, UPS will use artificial, software-based neurons (neuron detectors) to generate stochastic and uncorrelated bit patterns from EKG data, which serve as input signals for the Müller elements. The classification of the ECG data will be done both in software and finally with the help of the hardware-based Müller-C elements.
Partner: C2N/Université Paris-Saclay, Frankreich
Contact:
© Kai Schmidt/IAP
Dr. Hans Kleemann
ODS group (Organic Devices and Structures)
Send encrypted email via the SecureMail portal (for TUD external users only).