Robust Adaptive Beamforming for Uniform Circular Arrays based on Interference-plus-Noise Covariance Matrix Reconstruction

In practical applications, various model mismatches can significantly degrade the performance of beamformers. In this paper, we propose a robust adaptive beamforming (RAB) algorithm that achieves superior output performance for array models based on the Uniform Circular Array (UCA). The proposed algorithm reconstructs the interference-plus-noise covariance (INC) matrix by estimating the steering vector (SV) of the signal of interest (SOI), interference power, and noise power. Specifically, the eigenvector associated with the SOI is obtained via eigen-decomposition using a subspace algorithm and subsequently employed to estimate the SOI SV. The noise power is estimated by truncated averaging of small eigenvalues obtained through eigen-decomposition. The interference power is then estimated based on the orthogonality property between different signal SVs, facilitating the reconstruction of the INC matrix. Finally, the beamforming weights are calculated using the estimated SOI SV and the reconstructed INC matrix. The proposed algorithm only requires prior knowledge of the UCA geometry and the angular sector that contains the SOI SV. Simulation results are provided to demonstrate that the proposed algorithm effectively mitigates signal self-cancellation at high signal-to-noise ratios (SNR) and achieves significant improvements in output signal-to-interference-plus-noise ratio (SINR) performance under various mismatch conditions.