Research Of Wideband Spectrum Sensing Based On Coprime Sampling

Coprime spectrum analysis is an emerging spectrum analysis method that can break through the Nyquist sampling theorem.This method can reconstruct the spectrum of signal by processing undersampled samples of two coprime sampling path.For the spectrum analysis of wide band signal with high frequency,coprime spectrum analysis structure is able to bypass the procedure of recovering the original signal to obtain the spectrum analysis result.But the original coprime spectrum theory can not be directly employed in wide band spectrum sensing,that is because there are four flaws in it: 1)The undersampled samples collected by coprime sampling are not sufficiently been utilized,and this will cause longer time of observation to acquire the ideal result of spectrum sensing;2)The spectrum resolution is strictly limited by the processing structure of coprime samples,and the spectrum resolution can be only adjusted by changing the sampling rate of two Analog-to-Digital Converters;3)The coprime spectrum analysis method based on DFT filter banks will produce many spurious peaks when there are some specific stimulus components;4)The coprime theory lacks the theoretical analysis of the frequency estimation accuracy which the coprime samples could provide.Contraposing these deficiencies,this thesis proposes corresponding solutions and theoretical supplement as: 1)Process an appropriate extension and overlap strategy on the sample block of coprime samples,and consider the utilization of the autocorrelation information between coprime samples in the same coprime sampling path;2)Employ the extension and overlap strategy of sample blocks on coprime MUSIC method,coprime spectrum analysis method based on autocorrelation function and phase difference method to realize the aim of adjusting spectrum resolution with not altering the sampling rate;3)Introduce a modified analyzer by synthesizing the original and the complementary analyzer for removing the spurious peaks;4)Deduce the CRB(Cram?er-Rao Bounds)of frequency estimation which is yielded by coprime samples of one single path,and the CRB is verified by numerical simulations;5)Apply the coprime spectrum analysis method based on DFT filter banks and the coprime MUSIC method on wide band spectrum sensing.Generally speaking,by means of these improvement,the effectiveness of coprime spectrum analysis is demonstrated at first;And then more spectrum analysis methods are introduced to cooperate with coprime sampling;In the end,the modified process structure of coprime samples is employed in coprime spectrum analysis to achieve the goal of enhancing the resolution and decrease the sensing time of wide spectrum sensing.