Semi-Adaptive Beamforming for Co-Prime Circular Microphone Arrays

This paper proposes a semi-adaptive beamforming (SAB) algorithm for the co-prime circular microphone array (CPCMA), which takes advantage of subarray-based adaptive beamforming to achieve an optimised overall beampattern. The SAB approach calculates the bearings of grating lobes or the largest side lobes in the first sub-array before setting nulls at these directions in the beampattern of the second sub-array, aiming at removing the largest amplifications of undesired sources in the first beampattern. Compared with state-of-the-art co-prime array adaptive beamforming methods, the SAB considers the beamformer weights for each sub-array separately before combining them and generating the overall microphone array output, which fully utilises the co-prime property. Simulations indicate that the SAB improves the beampattern and array gain (AG) at low frequencies, which are dominant components of the speech energy, whilst maintaining equivalent results to the conventional CPCMA at high frequencies, leading to overall better performances in terms of beampattern and AG.