RFSoc based high speed frequency spectrum analyzer

Abstract

This project focuses on the characterization of wideband signals and the identification of un- known transmissions in dynamic wireless environments. As modern Software Defined Radios (SDRs) advance, traditional systems with bandwidths around 56 MHz are now being replaced by RFSoCs capable of processing up to 2 GHz of bandwidth, with sampling rates as high as 4 GS/s for 16-bit representations. This surge in data throughput necessitates the development of efficient and high-speed filter banks or channelizers to handle real-time wideband signal analysis. To establish a strong foundation, preliminary work was conducted on multi-channel and FIR filtering techniques, offering insights into effective parallel signal decomposition. Building on this, the project employs an oversampled polyphase filter bank (OPFB) to extract key features such as bandwidth, center frequency, and spectral occupancy. The design is targeted for implementation on the Xilinx Zynq UltraScale+ RFSoC platform, enabling scalable and real-time spectrum monitoring and unknown signal identification in high-throughput RF systems.