Spin-Torque Based Radio-Frequency Signal Classification Front-End.

Many classification applications rely on real-time processing and detection of RF signals at high frequencies. RF signal sampling requires sophisticated hardware, i.e., broadband analog front-ends and high- speed analog-to-digital converters according to the well-known Shannon-Nyquist theorem. Such devices either are expensive or suffer from limited detection bandwidths and sampling rates. Here, we proposed a novel spintronic- based classification front- end for real-time analysis and classification of RF signals. In comparison to the conventional CMOS-based systems, the proposed system can increase the classification speed dramatically while consuming an order of magnitude less power.