Retea neuronala optoelectronica pe baza de celule sensibilizate cu coloront fara consum electric -AIDSSC
Optoelectronic neural network based on dye-sensitized solar cells for zero electric power consumption- AIDSSC
PN-III-P2-2.1-PED-2021-0624 (728PED/2022)
29/06/2022 - 28/06/2024
Pentru prima data la nivel mondial, proiectul AIDSSC propune sa dezvolte si sa valideze o retea neuronala optoelectronica prototip pe baza de celule solare sensibilizate cu colorant fara consum de energie electrica in conditii de iluminare naturala si artificiala. European Green Deal 2021 propune obiecte ambitioase de mediu care necesita o transformare "verde" a multor sectoare ale societatii, in special referitoare la dezvoltarea sustenabila a mediului urban, conceptul de "oras inteligent" fiind una din cele mai importante provocari. Potentialul adaptabil al inteligentei artificiale (IA) de a contribui la realizarea obiectivelor unei tranzitii "verde" a determinat o dezvoltare continua a conceptului. In acelasi timp, digitalizarea cum ar fi si IA poate determina o crestere a consumului de energie electrica si de resurse, creand riscul efectelor adverse asupra mediului inconjurator. In acest context, celule solare deschid o noua perspectiva a dispozitivelor fara consum de energie electrica. Caracteristicile celulei solare sensibilizate (DSSC) ar putea face din aceste celula candidatul ideal pentru utilizarea lor ca si sisteme de inteligenta artificiala fara consum de energie electrica. Capacitatea de a invata a unui DSSC a fost pus in evidenta doar empiric de Gratzel, inventatorul acestui tip de celula solara, utilizand timpul de iluminare cu un simulator solar ca si rata de invatare. Acest nou concept pentru dispozitivele cu IA si anume, celule solare cu invatare optica, prezinta un real potential pentru a fi folosite ca blocuri de baza pentru dispozitive optoelectronice inteligente activate de lumina naturala si artificiala operand la putere minima si fara consum de energie electrica. In acest context, proiectul AIDSSC isi propune sa demonstreze ca prototipul retelei neuronale optoelectronice pe baza de celule solare sensibilizate cu colorant este un concept validat si credibil din punct de vedere tehnic si economic pentru dispozitive cu inteligenta artificiala.
For the first time worldwide, AIDSSC project proposes to develop and validate an optoelectronic neural network prototype based on dye-sensitized solar cells for zero electric power consumption in the outdoor and indoor lighting conditions. The European Green Deal 2021 proposes ambitious environmental goals which require a green transformation of many sectors of society, in especially a sustainable development of urban areas, "smart city", being a challenge of key importance. The transformative potential of artificial intelligence (AI) to contribute to the achievement of the goals of a green transition have been increasingly and prominently highlighted. At the same time, digital technologies such as AI considerably increase energy and resource consumption and create risks of adverse environmental effects. In this context, the solar cells open a new perspective as the most suitable candidate for zero electric power consumption. The dye-sensitized solar cell (DSSC) has the essential characteristics that could make these cells the ideal candidate for artificial intelligence system with zero electric power consumption. A singular preliminary proof of the DSSCs' learning capabilities was highlighted by Gratzel, the inventor of DSSC, using the exposure time as a cue for learning under 1 sun-simulated conditions. This new perspective for AI devices which are the optically learning solar cells, have the potential to serve as building blocks for intelligent optoelectronics enabling visually interacting machines that operate at minimal power and zero electric power consumption, in both outdoor and indoor lighting conditions. In this context, AIDSSC project aims to demonstrate that the optoelectronic neural network working prototype based on dye-sensitized solar cells is a technically and economically credible concept for the artificial intelligence devices.