Direct-RF 64 GS/s Wideband Approximate DFT Analysis Filterbanks for Spectrum AI-Perception

Situational awareness across a wide band of radio spectrum is crucial for wireless systems. Sensing, followed by artificial intelligence (AI)-based perception, leads to machine intelligence that has an awareness on what is happening in the radio frequency (RF) environment. RF-AI perception requires signal acquisition across a wideband followed by digitization and channelization so that multiple AI engines can perceive in realtime. We demonstrate ongoing work on RF-AI spectrum perception across 0–32 GHz in a single mixer-free direct-digital receiver. Using 32 GHz across in-phase (I) and quadrature (Q) channels, we showed the suitability of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathcal{O}(N)$</tex> approximate DFT (ADFT) analysis filterbanks for low-complexity spectrum monitoring up to 64 GHz in intermediate frequency (IF), and covering FR1, FR3 and FR2 bands up to 32 GHz in a direct-digital receiver incorporating Intel Altera Stratix-10 AX chiplet technology.