An Approach Of DOA Estimation With Coprime Array

Array signal processing is a technology that receiving and processing the signal by using arrays of different structure to sample and process the spatial signal.As a major research domain in the array signal processing technology,spatial spectrum estimation mainly research how to use sensor arrays to accurately estimate the signal parameters in order to improve the accuracy of DOA estimate,the angular resolution and the estimation of other parameters.This technology has a wide application and prospect.However,among the classical DOA estimate algorithms,the common uniform linear array(ULA)always requires the number of sensors cannot be less than the number of signal source.This limit makes the computation complexity increase.Thus,how to use the sparse array to sparsely sample and how to improve the performance of the sparse array,or in other word,how to recover more signals by using lesser sensors becomes a hot spot gradually.According to the property of sparse array,this paper analyses the co-array concept of coprime array and compares the array aperture of different co-arrays.As the result,we get the degree of freedom(DOF)of the different co-arrays.We also simulate the co-arrays configurations and the DOA estimate process and contrast the results.At first,we process the co-prime array with high-order cumulant(HOC)to get the difference co-array.The virtual array of the difference co-array highly increases the number of the consecutive virtual array element.It allows the number of source can be estimated could be more than the number of sensors.Summarizing the property and the DOF of the difference co-array,we propose an optimized coprime array and the fourth-order DOA estimation algorithm based on the optimized coprime array.Then,we simulate the corresponding virtual arrays configuration and the DOA of the optimized virtual array.We also compare the results with the ULA circumstance.In this paper,the pseudo-snapshots augmented algorithm is presented.The algorithm can fully adopt second-order statistics so that the corresponding virtual array has a higher DOF and more continuous virtual sensors.The pseudo-snapshots augmented vectorization algorithm can also remedy the lack that the HOC cannot be applied on the Gaussian signal.This paper also summarizes the property of the pseudo-snapshots augmented virtual array.We simulate and compare the different structures of virtual array configurations and the corresponding DOA estimations.The results show that the pseudo-snapshots augmented algorithm can achieve a higher DOF and extend the aperture of the virtual array.