Research On The Estimation Method Of The Likelihood Product Space Spectrum Of Coprime Matrices

Spatial spectrum estimation attracted the attention of scholars at home and abroad because of its wide applicability,and it is the key research content in the field of array signal processing.With the development of sonar,radar,wireless communication and other fields,the performance requirements of spatial spectrum estimation algorithm are becoming higher and higher.Coprime array can not only expand the array aperture,but also improve the resolution of spatial spectrum estimation,and it is suitable for distributed spatial spectrum estimation.The shortcomings of the traditional method of resolving ambiguity based on coprime matrix,a new method of ambiguity resolution of coprime matrix is proposed by multiplying the likelihood functions of coprime subarrays.Furthermore,a similar likelihood product coprime matrix spatial spectrum estimation method is extended based on the likelihood product ambiguity resolution method.Finally,based on deeply study of existing high-resolution spatial spectrum estimation and its improved algorithm,a new method is proposed a spatial spectrum estimation method of coherent sources based on order statistics and expectation maximization(EM)algorithm is proposed.The performance of this method is simulated under low SNR and low sampling number.The main work of this paper is as follows:1.A new spatial spectrum estimation method for coprime arrays is proposed—likelihood product method,based on the research of the existing defuzzification algorithms.The spatial spectrum estimation is realized by multiplying the spatial spectra of two subarrays.The simulation results are compared with those of the traditional deblurring methods.Experimental results show that the likelihood product deblurring method has higher estimation accuracy,especially in low SNR and less snapshots.2.Music product and PM product are extended to solve ambiguity problem of coprime matrix,on the basis of likelihood product method.The spatial spectral functions of music method and PM method reflect the possibility that the source comes from various spatial angles,so they have similar properties.After multiplying the spatial spectrum of two subarrays,there will be a peak at the real angle,so the spatial spectrum estimation of coprime matrix can be realized.Finally,the simulation results show that the likelihood product method and likelihood product method have higher estimation accuracy than the traditional method.Compared with the likelihood product method,which requires multi-dimensional spectral peak search,the product ambiguity resolution method has lower computational complexity.3.For coherent sources,proposed a new EM spatial spectrum estimation method based on fourth-order cumulants.Simulation results show that this method has good performance in low SNR,good stability and resolution,and can complete high-resolution spatial spectrum of coherent sources estimates.