Spectrum sensing and beamforming operations in modern cognitive radios and battlespace are contending in a congested spectrum where jamming is intentional. The discrete Fourier transform (DFT) is a cornerstone operation, in this regard, which requires high computation. In this work, we are introducing analog approximate DFT (ADFT) which is a multiplierless operation. ADFT-based maximally-decimated uniform polyphase filterbanks are realized to evaluate the performance of the filterbanks against different signals of interest. The approximations introduced by the ADFT are mitigated by using a power-correction scheme tested for different signals. Finally, the analog ADFT is proposed to further reduce the burden of RF front-ends.
