Two-Dimensional DOA Estimation Algorithm for Sparse Arrays under Sensor Failures

To address the problem of the destruction of virtual array continuity and the degradation of degree of freedom due to missing data in two-dimensional sparse array under the conditions of sensor failures, a two-dimensional DOA estimation algorithm is proposed. Firstly, the virtual array is constructed based on the two-dimensional difference coarray, and then the covariance matrix data is recovered in the form of matrix completion by using decoupled atomic norm minimization to realize the virtual array interpolation. Finally, the SS-MUSIC algorithm is used for the two-dimensional DOA estimation of multiple sources. The proposed method compensates for the effects caused by the failure of physical sensors. It recovers the complete aperture characteristics of the original virtual array and maintains the degree of freedom of the virtual array, which ensures a higher-precision two-dimensional DOA estimation performance. Simulation results demonstrate that under the same number of physical elements and sensor failure, the proposed algorithm can effectively estimate more sources compared with the existing methods, and exhibits higher robustness under the conditions of a small number of snapshots and low signal-to-noise ratio. This approach maximally retains and utilizes the degree of freedom advantage of sparse arrays in the two-dimensional DOA estimation.