Array antenna failure robust compensation via tensor Hankelization in MIMO radar for DOA estimation

Abstract

Utilizing low-rank tensor completion (LRTC) provides great success in the recovery of missing entries by exploiting the multidimensional structure. However, the assumption of most LRTC methods that the observed samples are randomly distributed is violated in multiple-input multiple-output (MIMO) radar under antenna failure, in which some fibers or slices are missing in the third-order measurement tensor. An antenna failure compensation method based on Hankel low-rank tensor recovery is proposed for robust direction-of-arrival (DOA) estimation in MIMO radar. To reduce computational burdens and noise sensitivity, we first use higher-order singular value decomposition (HOSVD) to lower the dimensions of the observations. Then, each frontal slice of the reduced-dimensional tensor is mapped into twofold Hankel structure to better reveal the relationship among different fibers. We further explore the Hankelization operation in a tensor manner to constitute high-dimensional structured Hankel tensor by concatenating the subtensors to capture the slice correlations. Because of the stronger low-rank property and the lack of entirely missing fibers or slices, the LRTC method can be applied to this structured Hankel tensor to recover structurally incomplete tensor. Simulation experiments verify that the proposed method achieves the robust compensation of faulty antennas in MIMO radar, ensuring remarkable DOA estimation performance.