Seminal
The magazine and atlas of EU-funded researchA new publication from the web and AI team at Scott Charnley

Atlas / France

European Research Council grant · Running

RF-ImagingNet RF-ImagingNet: Spintronic RF Neural Front-Ends for Embedded Microwave Sensing

Coordinator
Centre National De La Recherche Scientifique CNRS, France
Runs
1 January 2027 to 30 June 2028
Status
Running
EU funding
€150k
Total cost
—
Partners
1 organisations in 1 countries
Type
European Research Council grant
Grant agreement
101335077

What the project is doing

In the project's own words, from its summary on CORDIS.

Active microwave sensing probes biological tissue, walls or soil with safe, easy-to produce non-ionising waves. It could enable cheap, portable scanners for medical and field inspection applications with fast on-site diagnosis. But current systems rely on wideband digitisation and heavy digital processing, making them too bulky, slow, and power-hungry to be embedded. RF-ImagingNet shifts the sensing-computing boundary by performing the first inference step directly in the radio-frequency (RF) domain during acquisition, so that only a low-rate decision output is digitised.

By processing frequency-dependent RF signals before sampling, the approach reduces data movement and relaxes ADC and compute requirements. Building on recent results showing that spintronic devices can act as programmable frequency-selective RF synapses, we will integrate an RF neural layer with a compact active sensor (2 transmitters, 2 receivers, 1-9 GHz) and demonstrate direct target detection and classification from microwave measurements on hidden objects, without first reconstructing a full image.

Objectives are sub-10 ms end-to-end latencyand near-digital decision accuracy (<2% degradation) on the selected tasks. Energy benefits will be assessed at the architecture level by quantifying the avoided cost of wideband digitisation, data movement and digital inference. The RF neural front-end can operate either as a stand-alone decision sensor or as a preprocessor that triggers full imaging only when needed, reducing duty cycle and limiting raw data storage and transmission.

The Proof of Concept will deliver a reproducible demonstrator, quantified benchmarks, and an exploitation package including foreground IP assessment and early adopter feedback from industrial stakeholders.

Topics

Who is involved

OrganisationCityCountryRole
Centre National De La Recherche Scientifique CNRSParisFranceCoordinator

Source: CORDIS, European Commission. Data of 5 Nov 2026, reused under the Commission's reuse policy. View on CORDIS · Grant DOI