Eigenvectors, diagonal loading and white noise gain constraints for robust adaptive beamforming

In this paper, the use of Initial diagonal loading in addition to a white noise gain constraint (WNGC) is investigated for robust adaptive beamforming. The WNGC controls both the amount of self-ing loss and the resolving capability. The fundamental problem with WNGC is the setting of the maximum WNG must vary to achieve satisfactory self-ing protection for different levels of array errors. Since array errors are usually not known precisely, the WNGC is set overly conservative to avoid self-ing losses causing degraded resolution performance. Initial diagonal loading is a means of preventing self-ing losses on weak signals and allows the WNGC setting to be much more aggressive, thus improving resolution. The philosophy behind the approach is to use initial diagonal loading to protect weak signals and the WNGC to preserve other stronger signals for which a small amount of self-ing loss is tolerated. In addition, the performance of this robust ABF approach using WNGC with initial diagonal loading is compared to robust ABF methods that model steering vector uncertainties.