The Study Of The DOA Estimation Based On Sparse Signal Representation

The direction-of-arrival estimation (DOA) is one of the most significant branchesof array signal processing research, and plays an important role in many applicationsincluding radar, sonar, wireless communication and seismic sensing. An important goalfor source localization methods is to be able to locate closely spaced sources in presenceof considerable noise. The DOA estimation method based on sparse signalrepresentation reconstructs the sparse signal with the sparse representation of the signalreceived, then the direction can be estimated. This kind of DOA estimate algorithm isdifferent from these subspace methods, which dose not require the number of sources inadvance, and the assumption that the signals are independent need not to be made,what’t more, this algorithm is not sensitive to the array error, and has thesuperresolution. This paper represents a method that localizes the sources with thesparse signal representation method, and the work can be divided into the followingsections:1. Introduce two classical DOA estimation methods in detail, including MUSICand Capon algorithm, and their performances are analyzed.2. Introduce the basic theories of sparse signal representation, the sparse signalmodel is made,1SVDand FOCUSS algorithm are elaborated in detail to illustratethe application of sparse signal in DOA estimation field.3. Solve the signal reconstruction with reiterative DOA estimation algorithm whenthe sparse signal representation is used to estimate DOA, and this paper extends thisalgorithm from narrowband signal model to wideband signal model, what’s more, theperformances of the algorithm are analyzed with two signal models.4. Construct the sparse signal representation DOA estimation array signalmathematics model when the array error exists, and improve the reiterative DOAestimation algorithm to make it be more robustious to the array error. Besides, theanalysis of the influence of array error to robust reiterative DOA estimation algorithm isdone, then study the performance of this algorithm. 5. Consider the main drawback of reiterative DOA estimation algorithm that it’scomputation is large when the snapshots are large, combine reiterative algorithmwith1SVD, the signal space vectors can be got from the singular valuedecomposition(SVD) of received signals, then deal with the signal space vectors withreiterative algorithm, and the computation complexity will be reduced.