Research on parameters estimation of acoustic vector array signals using the compressed sensing theory

In this paper we applied the compressed sensing (CS) theory to signals processing of acoustic vector array, and realized the direction of arrival (DOA) estimation of small number of snapshots data. A new method, called CS, asserts that for sparse or compressible signals, far fewer samples or measurements than traditional methods used can contain all the information of signals. One can recover the original signals accurately from these samples or measurements by using reconstruction algorithms. Herein, we first construct the model of acoustic vector array, and present the corresponding CS algorithm. According to the angle sparse space, the over-complete dictionary can be constructed. The measurement matrix is optimized by the quantum-behaved particle swarm optimization algorithm (QPSO) to decrease the mutual coherence between measurement matrix and over-complete dictionary. An improved orthogonal matching pursuit algorithm (OMP) is used to obtain the estimation of sparse vector. Then from the angle spectrum, the DOA estimation of targets is obtained. By conducting several experiments, we obtained high resolution estimation of targets´ DOA on the condition of low signal-to-noise ration (SNR) and small number of snapshots.