Projection-based robust adaptive beamforming with quadratic constraint

• We give a new objective function of optimization problem.
• Performance of basic beamformer versus SNR is analyzed with random matrix theory.
• Performance of basic beamformer improves at low SNR with a quadratic constraint.
• Proposed method is robust against arbitrary steering vector mismatch.

A new robust adaptive beamforming technique is proposed in this study to address performance degradation of adaptive beamforming methods in the presence of steering vector mismatch. Actual steering vector of desired signal is estimated by solving a convex optimization problem with the objective constructed by minimizing the sum of estimated desired steering vector projections onto noise eigenvectors. The beamformer performs well at high signal-to-noise ratio (SNR) with the orthogonality between presumed desired steering vector and mismatch vector as a single constraint. Feasibility and necessity of adding an additional quadratic constraint are verified through detailed performance analysis with random matrix theory, improving the performance at low SNR. The parameter determination approach is provided to allow the proposed beamformer to function properly in practical situations. Both the theoretical analysis and simulation results demonstrate the proposed method is robust against any steering vector mismatch.