Adaptive Beamformer based Left-Right Ambiguity Resolution Using Twin array

Abstract

A linear array with omnidirectional sensors cannot discriminate azimuth and its complementary angle due to symmetry along the array axis. Twin or triplet arrays are used as preferred solutions to the bearing ambiguity problem, also known as Left-Right (LR) ambiguity. In this paper, we use an adaptive beamformer based on Minimum Variance Distortionless Response (MVDR) beamformer along with cardioid beamforming to produce a null electronically steered to the complementary angle of the desired direction of a twin array. In the adaptive scheme, beamforming is first performed along the azimuth plane to obtain two sets of beams. The beam signal from each array is then combined linearly with weights determined to minimize the output power at a complementary angle subject to maintaining unity response in the desired direction. The minimization principle tends to develop a null at the complementary direction, while the unity response constraint ensures an undistorted signal from the desired direction. We demonstrate the performance of MVDR based beamformer on a twin array in the correlated noise modeled by a first-order auto-regressive process. The performance is characterized based on two parameters, namely detection gain (DG) based on the deflection criterion and rejection ratio (RR). At a higher spatially correlated noise, adaptive beamformer has higher DG and RR.