Beam-domain deconvolution beamforming algorithm based on compressive sensing

Abstract

Deconvolution beamforming is often used to improve the azimuth resolution in target detection. However, most traditional deconvolution beamforming algorithms require the array directivity function to be shift-invariant, which are suitable for specific arrays such as linear arrays and circular arrays. A beam-domain deconvolution beamforming method suitable for arbitrary array is proposed based on compressive sensing. First, conventional beamforming is used to obtain several complex beam outputs, then the Sparse Bayesian Learning (SBL) reconstruction algorithm is applied to the beam-domain model to achieve deconvolution of complex beam outputs. The deconvolution process leads to more accurate estimation for the Direction Of Arrival (DOA) for target signals. The proposed method can also effectively reduce the computational complexity of the algorithm by controlling the number of output beams from the conventional beamforming and is applicable for both non coherent and coherent signals. The simulation and experiment results show that the proposed algorithm has azimuth resolution performance comparable to the traditional element-domain SBL beamforming algorithm, and is superior to the conventional beamforming and Minimum Variance Distortionless Response (MVDR) algorithms. When applied to short and dense arrays, the computation complexity of the proposed algorithm is significantly lower than that of the traditional element-domain SBL beamforming algorithm.