OSCAR: Open-Source AI-Driven PA-DPD with Computing-In-Memory Acceleration for Energy-Efficient 6G FR3 Wideband Transmitters

Bidragets beskrivning

Mobile networks consume hundreds of TWh of electric energy annually, and most of which is consumed by base stations. A key reason is the transmitter, where the power amplifier wastes energy and the digital correction that keeps its signal clean grows more complex as 6G uses wider bandwidths. The OSCAR project develops an energy-efficient transmitter for 6G networks by designing the power amplifier, an AI algorithm that corrects its distortion, and a dedicated computing-in-memory chip that runs this AI with very little energy, all together rather than separately. Aalto University works with Yonsei University and GIST in Korea, combining expertise in radio circuits, AI and chip design. The project will build prototype chips and release its AI training framework, circuit models and measured datasets as open source for researchers and industry. The results can help make future wireless networks more sustainable and lower the energy and cooling costs of network operators.
Visa mer

Startår

2026

Slutår

2029

Beviljade finansiering

Kwantae Kim Orcid -palvelun logo
499 998 €

Finansiär

Finlands Akademi

Typ av finansiering

Kansainvälinen yhteisrahoitteinen ohjelma

Beslutfattare

Suomen akatemian muu päättäjä
16.11.2026

Övriga uppgifter

Finansieringsbeslutets nummer

381424

Vetenskapsområden

El-, automations- och telekommunikationsteknik, elektronik

Forskningsområden

Sähkötekniikka ja elektroniikka

Identifierade teman

5G, 6G, wireless networks, wireless communication