Hardware electronic circuit modeling neuron with diode based non-linearity
The purpose of this study is hardware implementation of the nonlinear electronic generator of neuron-like activity developed previously using nGSPICE software. Based on the previously developed circuits, we simplified the design by reducing the number of components and rearranging nonlinear elements. Various approximations of the nonlinear function were considered by means of different number n of diodes in the circuit, and the dependence of the output signal on n was investigated. Quasi-linear and nonlinear generation regimes were studied. Also, a simple partial bifurcation analysis was performed in order to study the mechanisms of oscillation birth in the circuit. As a result, taking into account the non-ideality of physical components, we came to the conclusion that our circuit developed previously in the ngSPICE simulator can be easily constructed in the hardware.
Authors
- В. И. Пономаренко (ORCID: https://orcid.org/0000-0002-1579-6465)
- Ilya V. Sysoev (ORCID: https://orcid.org/0000-0002-9203-5504)
- Lev V. Takaishvili (ORCID: https://orcid.org/0009-0006-7860-7883)
Institutions
- Saratov State University (RU)
- Kotelnikov Institute of Radioengineering and Electronics of the Russian Academy of Sciences (RU)
Publication Details
- Published
- 2024-09-19
- DOI
- https://doi.org/10.1109/dcna63495.2024.10718580
- Primary Topic
- Neural Networks and Applications
- Type
- article
- Field-Weighted Citation Impact
- 0.00