Accurate And Computationally Efficient Interpolation-based Method For Array Signal Processing

The development of array signal processing leads to an excellent performance of the military reconnaissance and civil communication systems.The spectrum algorithm is a significant research direction in array signal processing.However,it is a contradictory problem about the accuracy and computational efficiencyof spatial spectrum algorithm.Especially for the joint DOD and DOA estimation,the existingspatial spectrum estimation algorithms cannot be applied to the actual engineering environment because of the high computational complexity.In this thesis,the algorithms of DOA estimation and the joint DOD-DOA estimation are studied.An accurate and relatively efficient interpolation-based Method are proposed to solve the large computational problem caused by the computation of sample covariance matrix as well as the eigenvalue decomposition and/orthe spectral peak search in DOA estimation algorithms.The main works of this thesis is summarized as follows: The background,mathematical model,research history and present situation of spatial spectrum algorithm is introduced firstly in this thesis.Secondly,give four classical spatial spectrum estimation algorithms about one dimension and two dimensions respectively.Based on the research of the classical algorithms,a one-dimensional accurate DOA estimation algorithm and an accurate and computationally efficient interpolation-based method for joint DOA and DOA estimation are proposed.One dimensional accurate DOA estimation algorithms proposedfirstly,which using the idea of algorithm,Rapid Accurate Frequency Estimation of Multiple Resolved Exponentialsin Noise,and aims to obtain the accurate target direction in linear uniform array system.In this algorithm,the spectrum peak search can be avoided,in result of it,the performance of the algorithm is improved significantly.However,in the domain of DOA estimation,the received data often have multiple rapids,so that we can't simply use the linear addition of different beats to estimate the parameters.The algorithm proposed in this thesis can effectively solve the problem of DOA estimation.Finally,given the simulation results to show that the algorithm is more accurate compared with other DOA algorithms.Then,a new method for joint DOA and DOD estimation is proposed.At beginning,itis derived from the algorithm,one dimensional frequency estimation interpolation-based algorithm.Then applying this method to MIMO radar situation,the DOA and DOD can be estimated accurately and efficiently.Further,we more focus on the theoretical analysis which can prove the unbiased property of the proposed algorithm and give the theoretical deviation.And the corresponding simulation is tested at the end.Considering the data-missing situation,we propose a new algorithm,whichmakes full use of the iterative property of the interpolation algorithm.The interpolation algorithm and Lagrange algorithm are combined at each iteration to improve the accuracy and efficiency of the algorithm.Finally,the whole work of our thesis is summarized.Both the advantages and the deficiencies of our work is contained in the thesis.By the way,the future research direction is also represented briefly.