Normal mode separation based on compressive sensing with a horizontal array

According to the issue that traditional beamforming has low resolution and warping modal filtering is not applicable to arbitrary complex source signal,a normal mode separation method based on compressive sensing with a horizontal array is proposed.Taking advantage of compressive sensing,which has high resolution ability in azimuth estimation,normal modes can be separated in beam domain.That is to estimate the azimuth spectral distribution by compressive sensing,separate each normal mode on the plane of frequency and azimuth,and finally recover the waveform by inverse Fourier transformation.A pseudo-random as well as impulse sound source signal with bandwidth of 20-200 Hz is used to simulate sound field signal received by a horizontal linear array,whose aperture is 1 km and element space is 10 m.And then apply the method proposed above to separate normal modes and compute the correlation coefficients with the theoretical ones.The correlation coefficient is between 0.97-1.0,which verify the compressive sensing normal mode separation method is applicable to any form of sound source signal.Experiment data of air-gun signals received by the seafloor-deployed 28 element horizontal line array at the North Yellow Sea in 2011,accompanied with synthetic aperture method to get a 1 km array,is used for compressive sensing normal mode separation.The correlation coefficient of the first 5 normal modes separated with warping filtering is between 0.82-0.93.Simulation and experiment verify the effectiveness of the proposed method.