KFU. Card publication. An Approach to Synthesis of the Neuromorphic Functional Models for Analog Components and Blocks

KAZAN
FEDERAL UNIVERSITY

- RUS
- 中文
- ES

AN APPROACH TO SYNTHESIS OF THE NEUROMORPHIC FUNCTIONAL MODELS FOR ANALOG COMPONENTS AND BLOCKS

Form of presentation

Articles in international journals and collections

Year of publication

2022

��

��

Authors, employees of KFU

Mosin Sergey Gennadevich, author

Bibliographic description in the original language

Mosin S., An Approach to Synthesis of the Neuromorphic Functional Models for Analog Components and Blocks//Lecture Notes in Computational Science and Engineering. - 2022. - Vol.141, Is.. - P.335-346.

Annotation

Numerical simulation of analog circuits and functional blocks (FB) is an important design stage of analog and mixed-signal integrated circuits as well as the state-of-the-art embedded systems. The application of adequate mathematical models for components and FB defines the quality of simulation and influences the time and cost characteristics of the up-to-date microelectronic devices development process. An approach to the automated synthesis of functional models for analog components and blocks using machine learning methods—neuromorphic functional models (NFM)—is proposed in the paper. The approach implements the possibility to use either analytically defined dependencies (model-based) or dependencies obtained during natural experimental measurements (data-driven) as the raw data for the NFM synthesis. The design flow including the description of the mathematical models for an analog circuit (MMC) applying the NFMs and further numerical simulation in accordance with the assigned type of the circuit analysis is presented. The results of experimental research for a model of the semiconductor diode D1N4934 and circuits of the voltage rectifiers on its base are showed. The obtained results demonstrate the high precision of the synthesized NFM and the high quality of simulation. The comparison of obtained results with results of simulation in the Cadence CAD tools based on a structural model of the diode is performed. The simulation errors consist of less than 1% of the input signals' amplitude.

Keywords

Neuromorphic models, functional blocks, numerical simulation

The name of the journal

Lecture Notes in Computational Science and Engineering

URL

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85138788526&doi=10.1007%2f978-3-030-87809-2_26&partnerID=40&md5=62d2b230d7cc35f3cb67607243ad39ab

Please use this ID to quote from or refer to the card

https://repository.kpfu.ru/eng/?p_id=272028&p_lang=2

Full metadata record

Field DC	Value	Language
dc.contributor.author	Mosin Sergey Gennadevich	ru_RU
dc.date.accessioned	2022-01-01T00:00:00Z	ru_RU
dc.date.available	2022-01-01T00:00:00Z	ru_RU
dc.date.issued	2022	ru_RU
dc.identifier.citation	Mosin S., An Approach to Synthesis of the Neuromorphic Functional Models for Analog Components and Blocks//Lecture Notes in Computational Science and Engineering. - 2022. - Vol.141, Is.. - P.335-346.	ru_RU
dc.identifier.uri	https://repository.kpfu.ru/eng/?p_id=272028&p_lang=2	ru_RU
dc.description.abstract	Lecture Notes in Computational Science and Engineering	ru_RU
dc.description.abstract	Numerical simulation of analog circuits and functional blocks (FB) is an important design stage of analog and mixed-signal integrated circuits as well as the state-of-the-art embedded systems. The application of adequate mathematical models for components and FB defines the quality of simulation and influences the time and cost characteristics of the up-to-date microelectronic devices development process. An approach to the automated synthesis of functional models for analog components and blocks using machine learning methods—neuromorphic functional models (NFM)—is proposed in the paper. The approach implements the possibility to use either analytically defined dependencies (model-based) or dependencies obtained during natural experimental measurements (data-driven) as the raw data for the NFM synthesis. The design flow including the description of the mathematical models for an analog circuit (MMC) applying the NFMs and further numerical simulation in accordance with the assigned type of the circuit analysis is presented. The results of experimental research for a model of the semiconductor diode D1N4934 and circuits of the voltage rectifiers on its base are showed. The obtained results demonstrate the high precision of the synthesized NFM and the high quality of simulation. The comparison of obtained results with results of simulation in the Cadence CAD tools based on a structural model of the diode is performed. The simulation errors consist of less than 1% of the input signals' amplitude.	ru_RU
dc.language.iso	ru	ru_RU
dc.subject	Neuromorphic models	ru_RU
dc.subject	functional blocks	ru_RU
dc.subject	numerical simulation	ru_RU
dc.title	An Approach to Synthesis of the Neuromorphic Functional Models for Analog Components and Blocks	ru_RU
dc.type	Articles in international journals and collections	ru_RU