Implementation Of UWB Spectrum Estimation System Based On Compressed Sampling

With the rapid development of wireless communication technologies,wireless spectrum resources are becoming scarcer,and their efficient management and utilization are great significance for military and civilian use.Which closely related to it is spectrum monitoring,traditional Nyquist sampling requires a sampling rate at lesat twice the signal bandwide in order to perform distortion-free sampling.As wireless communication technology develops to higher frequencies year by year,the excessively high sampling rate makes the analog-to-digital converter(ADC)a bottleneck for new receivers.Compressive Sensing(CS)technology is becoming more and more mature,which making the recevier based on sub-Nyquist sampling possible.This new sampling techonology not only reduces the receiver's requirements for the ADC,but also compresses the signal during the sampling process and reducing the cost of system data storage.Therefore,people urgently want to apply this new theory to production.For example,the nonuniform sampling(NUS)system and the Modulated Wideband Converter(MWC)system.Threse are typical representatives of early compressive sensing applications.This paper implements a real-time wideband spectrum sensing system based on the fast compressed power spectrum estimation algorithm.The time complexity of algorithm is linear with the number of sub-Nyquist samples and the downsampling factor.This paper has done the following works:1)Realization and time complexity analysis of fast compressed power spectrum estimation algorithm;2)Simulation of the hardware circuit and the design of the printed circuit board,as well as component welding and debugging,compare the data of the real circuit test with the simulation data,to ensure that the hardware circuit meets the design requirements;3)Based on the designed hardware,a multi-coset sampling architecture with eight channels of parallel sampling is realized,and the data was transmissed from the Filed-Programmable Gate Array(FPGA)to the computer,so that the sub-Nyquist samples collected by the ADC can be sent to the computer for real-time power spectrum reconstruction.4)Algorithm verification and system performance analysis,in-depth discussion of the problems and solutions in the system.Receiver Operating Characteristic(ROC)curves are drawn in different signal-to-noise ratio scenarios,and the performance is compared with the existing frequency domain power spectrum reconstruction methods.Finally,this paper realizes that only eight 80 MHz ADCs are used to realize real-time spectrum sensing in the bandwidth of 0~1GHz.Compared with the existing research,the system realized in this paper has high real-time performance,low time complexity,accurate spectrum estimation and with excellent performance at low signal-to-noise ratio.