Study On Sparse Iterative Approach For Array-based Direction Of Arrival Estimation

Direction of arrival estimation is the core of array signal processing,and widely used in sonar,radar application.In the sonar application,long linear array leads to the failure of the conventional high-space algorithm of DOA when the signal is coherent or snapshot is deficient.And spares reconstruction methods provide a new perspective to find the direction of arrival in array processing field,and perform well in the scenarios of snapshot deficiency or coherent signal.In this paper,we study the performance of sparse iterative method in the DOA of array processing field and verify the validity by process of the experimental data.1.Build the DOA estimation model based on sparse reconstruction theory,including far-field model and near-field model.Then analyzes and compares L1-SVD method,OMP method and SPICE method.The results show that sparse methods perform better in low snapshots scenario and signal coherent scenario compared with conventional methods.But the class of SPICE methods have poor performance in the near field when snapshot is deficient.2.The performance of sparse iterative method is relative to the initialization.The simulation result show initializing with MVDR method outperform with periodogram method.Furthermore,the class of SPICE method still working when MVDR failed,although initialized with MVDR method.3.Apply the sparse iterative method in the real sea trial.And the results show the sparse method performs well in the real data compared with MVDR method,which has lower sidelobe,especially in the scenarios of snapshot deficiency or coherent signal.