Research And Implementation On Spectrum Sensing Of Wideband Signals Based On Sub-Nyquist Sampling

With the rapid development of radio technology and the continuous expansion of wireless spectrum,wireless spectrum resources are becoming more and more scarce.In order to improve spectrum utilization,monitoring and efficient management of broadband spectrum are important in the fields of radar and cognitive radios(CR).The traditional sampling method that relies on the Nyquist theorem is stretched when the Analog to Digital Converter(ADC)cannot meet the high-rate sampling.Therefore,the Compressed Sensing(CS)theory came into being.The maturity of technology makes digital receivers based on sub-Nyquist sampling possible.Unlike traditional post-sampling compression,this technology not only reduces the requirements for ADCs,but also reduces the burden of the system on data storage and transmission.This thesis focuses on the design and implementation of a wideband spectrum sensing system based on sub-Nyquist sampling.By analyzing the sub-Nyquist sampling technology and signal reconstruction algorithm based on CS theory,the multi-coset sampling architecture is studied and we proposed an improved sampling architecture and a frequency domain windowed fast power spectrum reconstruction algorithm,and give the overall design plan of the broadband spectrum sensing system,and complete the system implementation based on the FPGA(Field Programmable Gate Array,FPGA)platform.The main research work of this thesis includes:1.Through the analysis and comparison of suitable signal model,sensitivity,sampling matrix,reconstruction calculation amount,technical difficulty,and hardware direction for ADC bandwidth requirements,we introduced three typical sub-Nyquist sampling techniques: Random Demodulator Sampling(RDS),Modulated Wideband Converter(MWC),and Multi-Coset Sampling(MCS).Secondly,we performe simulation experiments on the reconstruction performance of compressed data for traditional greedy algorithms,convex relaxation algorithms,and sparse Bayes The Sri Lankan algorithm.2.To solve the problem that the traditional MCS structure is not flexible enough or the clock design requirements are too high,the sampling structure is optimized,and a design scheme for searching the optimal sampling mode is proposed;the relationship between the data after MCS sampling and the original spectrum is analyzed,and a Frequency domain windowed fast power spectrum reconstruction algorithm based on compressed sampling is proposed.3.Based on the improved MCS architecture and power spectrum estimation algorithm,we design and implement the broadband spectrum sensing system.Through system requirements and circuit analysis,the power supply module,sampling module,signal processing module,DSP(Digital Signal Processor,DSP)module and the realization of its interface circuit were completed,and the spectrum sensing system of the six-channel 160 MHz ADC to DC-1GHz bandwidth was completed.The performance test of the system hardware platform is completed under the existing laboratory conditions.