Design Of DSP Circuit And Signal Processing For The Broadband Spectrum Analysis Module

Spectrum analysis for broadband signal is focused in electronic measurement recently.Especially, in the field of non-cooperative signal detection, wireless communicationmeasurement, high-speed electronic equipment detection, the signal testing platformcharacterised in such as broadband signal sampling, deep memory capture, real-timespectrum analysis and high data transfer rate, have a wide range of demands.Based on a weighted overlap-add structure, which sequentially channellize band andreal-time channelized FFT on the four IF wideband signal, this thesis proposed a fastbroadband spectrum analysis scheme with variable frequency resolution and analysisbandwidth. In combination with plentiful logical computing resources in FPGA, adedicated signal processing DSP chip can show a powerful capability in real-timesignal processing. The thesis therfore introduces the broadband spectrum analysismodule circuit and resource allocation of the DSP system design.The DSP subsystem mainly is charged to dealwith the real-time channelize FFTprocessing and processing flow control. First of all, considering the structure of thechannelized data, it reassmbled the channelized data and extracted new multi-channeldata ordering into a sequent channel number. A fast windowing algorithm is operatedon the time-domain data before the multi-band signal FFT processing. Finally, as thenone-blind zone channelization, real-time spectrum of the input broadband signal canbe obtained by interception and joint of sub-band. The number of channels covered byprocess data is designed for tuning frequency resolution and analysis bandwidth. Inaddition, a windowing generation upon the fast short-table polling is proposed in thethesis, and it is well suited for the case of process speed and storage resources limited.For the experiments results, the variable frequency resolution and bandwidth can bevalidated exactly using CCS emulating environment, that is the maximum analysisbandwidth of400MHz and minimum frequency resolution of1kHz. Other results showthe good efficiency of the real-time channelized FFT processing algorithms.