Combination of weighted ℓ2,1 minimization with unitary transformation for DOA estimation

•The weighted ℓ2,1ℓ2,1 minimization is combined with the unitary transformation.•This combination improves the resolution and accuracy of DOA estimation.•More credible weights can be obtained after performing the unitary transformation.•The jointly sparse constraint can be further enhanced.•The implied spatial smoothing improves the decorrelation ability.

Using the centro-symmetry property of uniform linear array (ULA), we propose an algorithm that combines the weighted ℓ2,1ℓ2,1 minimization with the unitary transformation to improve the performance of DOA estimation. Exploiting the result of the unitary transformation, more credible weights can be obtained and the jointly sparse constraint can be further enhanced. Moreover, the unitary transformation incorporates the forward–backward spatial smoothing, which improves the performance of the weighted ℓ2,1ℓ2,1 minimization for correlated sources. Simulations demonstrate that the proposed method can achieve better performance in terms of resolution and estimation accuracy.