A Study On Robust Array Processing Algorithms For Direction Of Arrival Estimation

Array signal processing is one of the fundamental problems in signal processing,it deals with the received signals obtained by means of an array of sensors located at different points in space in the field of interest to detect signals and extract useful information about them.Array signal processing has been commonly applied in radar,sonar,communication, seismology,radio astronomy,tomography and other areas.This paper deals with the robustness of the direction of arrival estimation algorithms,and mainly discusses the robust DOA estimation algorithms for distributed signals,wideband signals and cyclostationary signals,as well as the situation in non-Gaussian noise environment.In brief this dissertation involves following innovatory researchs:In chapter 2,directions of arrival estimation algorithms for distributed signals are investigated,mainly for non-constant modulus distributed signals.Firstly the array signal model for distributed signal is analysised,then an AMLNM(abbr.of Approximate Maximum Likelihood algorithm for Non-constant Modulus distributed signals)named algorithm for non-constant modulus distributed signals is proposed to improve the algorithm robustness in the non-constant distributed signals situation.Chapter 3 discusses the DOA estimating algorithms for wideband signals,mainly analyzing the focusing technology in CSM algorithm.Since consistent focusing need not pre-estimating the direction of arrival,this chapter deduces two new consistent focusing matrixes named CRSS and CMTLS for wideband signals based on the ideal of AMI(abbr. of Array Manifold Interpolation)algorithm.The algorithm robustness is enhanced as well as the realtime using the proposed focusing matrixes.Chapter 4 analyzes the DOA estimating algorithm for cyclostationary signals,mainly discussing the robustness to cycle frequency error.Firstly an array signal model namedατmodel is constituted,then the effect to BFατand MVατalgorithms caused by cycle frequency error is analyzed,based on this analyzing two robust beamformers are deduced lastly.Comparing with BFατand MVατalgorithms,the two proposed beamformers are more robust to the cycle frequency error.Chapter 5 discusses the direction of arrival estimation in SαS(abbr.of Symmetric alpha Stable)noise environment,mainly investigating the method to simplify the computation algorithms.Firstly an approximate maximum likelihood algorithm is proposed under the assumption of Cauchy distributed noise.In this proposed algorithm,the multi-dimensional optimization problem is reduced to a one-dimensional optimization problem,thus quite simplifies the algorithm at relatively smaller loss in performance.