Our researcher Fábio Passos was interviewed by "90 Segundos de Ciência", on the usage of genetic algorithms and machine learning to automatically design electronic circuits.
Fábio Passos participates in the SYSTEMIC-RF project, an initiative that aims to use genetic algorithms and machine learning for the design and development of optimized electronic circuits.
The focus of the SYSTEMIC-RF project, led by Nuno Horta, is to develop advanced design techniques, where the circuit sizing and layout are treated in a fully integrated and automated fashion by using machine learning techniques and evolutionary algorithms. The different parts of the proposed methodology will be integrated into an EDA framework (www.aidasoft.com) in order to ease the usage by circuit designers.
“All genetic algorithms are evolutionary algorithms that, based on some input parameters, manage to evolve towards finding the most optimal circuit. Our idea is to apply these algorithms to the design of electronic circuits and automatically arrive at a circuit with an optimized performance”, he says. For example, these algorithms could be used to develop an amplifier capable of maximizing the gain and minimizing the power it consumes. Based on these parameters, genetic algorithms can find the most optimal design for this amplifier in a faster and more efficient way than a human.
“Recently we tested these algorithms and managed to manufacture a fully functional circuit that we sent to the factory, and which worked without problems”, he reveals.
The ultimate goal of SYSTEMIC-RF is to have a tool capable of producing factory-ready circuits that only need a designer to control the input parameters of the algorithms. “In this way, we will be able to make the electronic circuit design process much more efficient”, he reinforces.
Note: SYSTEMIC-RF – Automated synthesis methodology for reliable RF integrated circuits project is funded under a Marie Curie grant (grant agreement No. 892431) through funds from the European Horizon 2020 programme.
About Fábio Passos:
Fábio Passos received his Ph.D. degree from Universidad de Sevilla, Seville, Spain, in 2018, while conducting his work at Instituto de Microelectrónica de Sevilla (IMSE-CNM), Seville, Spain. He has performed research stays in several academic institutions such as IMEC (2016), Instituto de Telecomunicações (2017) and University of Barcelona (2017). In 2018 he was an intern at Analog Devices (as part of a technology transfer collaboration between IMSE and ADI) developing automated design methodologies for mm-Wave systems (77GHz transceiver) in an industrial and challenging environment. He is now a post-doctoral researcher at Instituto de Telecomunicações, Lisbon, Portugal. Dr. Passos has authored or co-authored over 30 publications, including books, book chapters, international journals and conferences papers. Dr. Passos was the SMACD 2016 EDA Competition Winner and received the Best Paper Award in SMACD 2018. He was the Competition Chair of SMACD 2019. His current research interests include the design and modeling of RF passive devices using analytical and machine-learning techniques and automated design methodologies for RF and mm-Wave circuits.
More on this project: https://www.it.pt/Projects/Index/4775
Follow to the podcast's URL: