Abstract
Differential beamforming, with its compact size and high directivity factor (DF), offers significant advantages for addressing various acoustic problems. While white noise gain (WNG) serves as a fundamental performance criterion for the robustness of differential microphone arrays (DMAs), simultaneously achieving both high WNG and frequency-invariant DF remains a significant challenge. This study presents a new method for the design of nonuniform linear robust DMAs. Our method effectively eliminates high-frequency extra nulls in the mainlobe of robust DMAs, thus preserving robustness and ensuring both frequency-invariant DF and consistent mainlobe width. The proposed approach achieves a WNG comparable to the maximum WNG method while mitigating its shortcomings in high-frequency, frequency-dependent variations of mainlobe width and DF, demonstrating its potential to enhance acoustic performance in complex environments.