Resrarch Of Spectrum Estimation Methods Based On Quadrature Compressed Sampling

Multiband signals are widely used in communication technology such as Radiofrequency(RF)and Cognitive Radio(CR)technology.The modulated wideband converter(MWC)proposed in 2010,can realize sub-Nyquist sampling of multiband signals without prior locations.In 2015,Quadrature Analog-to-Information Converter(QAIC)was proposed for multi-band signals with cluster structure characteristics.QAIC reduces the switching frequency of mixing sequence by pre-filtering the structural characteristics of multi-band signal clusters,thus reducing the complexity of mixing sequence.Based on QAIC,this paper explores the spectrum estimation methods for multiband signals.First of all,aiming at the problem of spectrum estimation failure caused by I/Q amplitude phase imbalance,this paper proposes a joint estimation method of I/Q imbalance and MWC based on sinusoidal sequence method,which can realize the nonideal joint estimation calibration of the entire QAIC system by inputting sinusoidal sequence.Secondly,aiming at the high complexity of periodic mixing sequence in MWC and QAIC,in this letter,we present a novel Digonal Remainder Matrix based Analog-toInformation Converter(DRM-AIC).DRM-AIC has only one non-zero element in a period in the mixing sequence of each channel,and the mixing sequence of different channels is generated by the delay of a base sequence.The non-uniform delay between sequences is obtained by remainder function and uniform sampling interval.Theoretical analysis and simulation results show that DRM-AIC reduces the complexity of sequence construction without affecting system performance and has better robustness of amplitude errors.Then,aiming at the problem that classical reconstruction algorithms for multiband signals requires the number of carrier frequencies of the original signal in advance which is a very difficult condition to satisfy.In this paper,a Spectral Projection Gradient(SPG)L1,1 algorithm for multiband signals is proposed,which uses the l1,1 norm of the matrix to measure the sparsity of the matrix.The sparsity is determined by assessing the numerical differential iteratively.The algorithm can realize blind reconstruction without carrier frequency and the complexity increases little.Simulation results show that the proposed algorithm has better reconstruction performance than the traditional simultaneous orthogonal matching pursuit(SOMP).Finally,since the theory of cyclic spectrum has good resolution and anti-interference ability,can process the signal under the condition of low SNR,realize the parameter estimation and signal extraction of signal detection,recognition and analysis.In this paper,the cyclic spectrum of the original signal is reconstructed directly from the compressed observations of the QAIC system,and the method of obtaining the carrier frequency and bandwidth of the signal without reconstructing the original signal is analyzed.Through verification of theory and simulation experiments,the diagonal remainder matrix proposed in this paper is applied to the reconstruction of cyclic spectrum,which has better reconstruction performance than the traditional random ±1 sequence.